Re: [sword-devel] SWORD 1.7.0 final stretch
On 9 July 2013 15:03, Greg Hellings wrote: > > > > On Tue, Jul 9, 2013 at 6:33 AM, Dmitrijs Ledkovs > wrote: >> >> On 9 July 2013 03:55, Jonathan Morgan wrote: >> > Hi Greg, >> > >> > Python bindings were packaged by Ubuntu, and I believe 1.6.2 (current >> > latest) is still available for Ubuntu. >> > >> >> And subsequently dropped, due to failure to compile with gcc-4.7: >> >> sword (1.6.2+dfsg-4) unstable; urgency=low >> >> [ Matthias Klose ] >> * Fix build failure in installed header file with GCC 4.7. >> Addresses: #667387. >> >> [ Dmitrijs Ledkovs ] >> * Remove python-sword swig bindings. Loads of build-time warnings (4.6) >> and errors (4.7) and the resulting module was reported not to >> work. Should be readded if/when it can be build with gcc 4.7 without >> errors. (Closes: #667387) >> >> -- Dmitrijs Ledkovs Tue, 17 Apr 2012 21:39:11 +0100 > > > I'd appreciate if you can give them another go. I spent quite some time in > the latter part of last year cleaning them up. On Fedora with gcc 4.7 now I > only get a single warning about unused code when I build them. If you're > still seeing errors and warnings, then I'd like to know what your settings > are and what your errors are. > > I'm keen to have the Swig bindings in a place where they can be included in > distros, so if you're still seeing problems, I'd like to work with you and > any other packagers to get them resolved. > At the time it was failing, with just simple gcc-4.7 -O2. Please note that debian & ubuntu is up at gcc-4.8 by default now. I haven't tried bindings in a while. (not since then) perl bindings never compiled for me on debian, as of yet. So no clue about those. Regards, Dmitrijs. > --Greg > >> >> >> >> >> > Jon >> > >> > >> > On Tue, Jul 9, 2013 at 12:55 AM, Greg Hellings >> > wrote: >> >> >> >> So I was trying to solve both the "Perl won't install" and the "Python >> >> installs to the wrong place" issues. I committed a fix to the Python >> >> install >> >> (it was previously only installing to /usr/local regardless of the >> >> value of >> >> CMAKE_INSTALL_PREFIX) but this fix appears to have magically caused >> >> Perl to >> >> begin installing on my machine. >> >> >> >> Peter, can you try and do a new install from the latest versions of SVN >> >> HEAD? >> >> >> >> Perl is still only installing to /usr/local regardless of where I set >> >> CMAKE_INSTALL_PREFIX, but I imagine that fix is low priority at the >> >> moment >> >> as the bindings are not presently packaged by any distributions I am >> >> aware >> >> of. >> >> >> >> I would invite distribution maintainers to begin packaging the Python >> >> and >> >> Perl bindings starting with 1.7.0. I have done lots to clean them up, >> >> suppress warnings, and resolve build failures. Packaging the Python >> >> bindings >> >> would permit packaging of BPBible as well, and also help with keeping >> >> the >> >> various build options I've created for the bindings clean and tested. >> >> >> >> --Greg >> >> >> >> >> >> On Sun, Jul 7, 2013 at 3:20 PM, Peter von Kaehne >> >> wrote: >> >>> >> >>> sword-tools/modules/crossreferences/xreffix.pl without arguments will >> >>> test loading Perl Sword bindings >> >>> >> >>> and in Python: >> >>> >> >>> Simply run from commandline Python and do "import Sword" and see it >> >>> fail: >> >>> >> >>> >>> import Sword >> >>> Traceback (most recent call last): >> >>> File "", line 1, in >> >>> File "Sword.py", line 26, in >> >>> _Sword = swig_import_helper() >> >>> File "Sword.py", line 18, in swig_import_helper >> >>> import _Sword >> >>> ImportError: /usr/local/lib/python2.7/dist-packages/_Sword.so: >> >>> undefined >> >>> symbol: u_strFromUTF8_48 >> >>> >>> >> >>> >> >>> >> >>> Gesendet: Sonntag, 07. Juli 2013 um 19:06 Uhr >> >>> Von: "
Re: [sword-devel] SWORD 1.7.0 final stretch
On 9 July 2013 03:55, Jonathan Morgan wrote: > Hi Greg, > > Python bindings were packaged by Ubuntu, and I believe 1.6.2 (current > latest) is still available for Ubuntu. > And subsequently dropped, due to failure to compile with gcc-4.7: sword (1.6.2+dfsg-4) unstable; urgency=low [ Matthias Klose ] * Fix build failure in installed header file with GCC 4.7. Addresses: #667387. [ Dmitrijs Ledkovs ] * Remove python-sword swig bindings. Loads of build-time warnings (4.6) and errors (4.7) and the resulting module was reported not to work. Should be readded if/when it can be build with gcc 4.7 without errors. (Closes: #667387) -- Dmitrijs Ledkovs Tue, 17 Apr 2012 21:39:11 +0100 > Jon > > > On Tue, Jul 9, 2013 at 12:55 AM, Greg Hellings > wrote: >> >> So I was trying to solve both the "Perl won't install" and the "Python >> installs to the wrong place" issues. I committed a fix to the Python install >> (it was previously only installing to /usr/local regardless of the value of >> CMAKE_INSTALL_PREFIX) but this fix appears to have magically caused Perl to >> begin installing on my machine. >> >> Peter, can you try and do a new install from the latest versions of SVN >> HEAD? >> >> Perl is still only installing to /usr/local regardless of where I set >> CMAKE_INSTALL_PREFIX, but I imagine that fix is low priority at the moment >> as the bindings are not presently packaged by any distributions I am aware >> of. >> >> I would invite distribution maintainers to begin packaging the Python and >> Perl bindings starting with 1.7.0. I have done lots to clean them up, >> suppress warnings, and resolve build failures. Packaging the Python bindings >> would permit packaging of BPBible as well, and also help with keeping the >> various build options I've created for the bindings clean and tested. >> >> --Greg >> >> >> On Sun, Jul 7, 2013 at 3:20 PM, Peter von Kaehne wrote: >>> >>> sword-tools/modules/crossreferences/xreffix.pl without arguments will >>> test loading Perl Sword bindings >>> >>> and in Python: >>> >>> Simply run from commandline Python and do "import Sword" and see it fail: >>> >>> >>> import Sword >>> Traceback (most recent call last): >>> File "", line 1, in >>> File "Sword.py", line 26, in >>> _Sword = swig_import_helper() >>> File "Sword.py", line 18, in swig_import_helper >>> import _Sword >>> ImportError: /usr/local/lib/python2.7/dist-packages/_Sword.so: undefined >>> symbol: u_strFromUTF8_48 >>> >>> >>> >>> >>> Gesendet: Sonntag, 07. Juli 2013 um 19:06 Uhr >>> Von: "Greg Hellings" >>> An: "SWORD Developers' Collaboration Forum" >>> Betreff: Re: [sword-devel] SWORD 1.7.0 final stretch >>> >>> There should be a sample of each in the swig directory tree somewhere. I >>> can check in an hour our so, when I get home. >>> >>> - Greg >>> >>> On Jul 7, 2013 11:42 AM, "Troy A. Griffitts" >>> wrote: >>>> >>>> OK guys, >>>> >>>> I have just updated the autotools stuff and the README files to build >>>> the swig bindings for both Perl and Python. Anyone know how to use SWORD >>>> with Perl or Python and can let me know if it actually works? >>>> >>>> Troy >>>> >>>> >>>> >>>> On 07/06/2013 03:04 PM, Peter von Kaehne wrote: >>>> >>>> On Sat, 2013-07-06 at 04:00 +0200, Troy A. Griffitts wrote: >>>> >>>> Peter, still wondering if we have a problem we need to fix for you. Do >>>> our Perl and Python bindings work for other people with a similar >>>> configuration as you? >>>> >>>> I have not heard from anyone that the Perl bindings are working nor have >>>> I heard from anyone with any clever ideas why they are not working. >>>> >>>> From a module making point of view the bindings are a must. But I do not >>>> care if they are working in in 1.7 or in a subsequent iteration as long >>>> as it is soon enough again. >>>> >>>> Peter >>>> >>>> >>>> >>>> ___ >>>> sword-devel mailing list: sword-devel@crosswire.org >>>> http://www.crosswire.org/m
Re: [sword-devel] Trends in mobile platforms
On 21 January 2013 21:44, Matěj Cepl wrote: > On 21/01/13 18:44, Dmitrijs Ledkovs wrote: >> But when not-docked, i.e. touch screen mode. The supported SDK/UI are >> Webapps or Qt 5 / QML with specific ubuntu phone touch UI components. > > And just to add one more reasons why I believe Ubuntu Phone is DOA ... > notice Qt 5. Mark Shuttleworth was sounding in the Ubuntu Phone PR video > that these are just different UIs for the same OS. Except it is not true > ... all those Gtk applications which currently run on plain Ubuntu need > to be completely rewritten to the new framework (and yes, those new apps > could run both on desktop and phone, but complete rewrite of apps is > required). > I like Gtk+ a lot, but hardly any existing GTK apps are touch friendly. So they need to be overhauled anyway. Re factoring & rewriting is usually quicker. iPhone is Objective-C, while android is approx. java based. Yet, many popular applications are available on both with near matching features (i recently switched from iphone to android so had a fresh switch-over comparison). I'm not competent enough to judge similarities & differences between Objective-C (only have grass-root NSObject knoweledge) / Java (nil) / Qt (nil). And how wide spread Qt is. W.r.t. DOA well... time will tell =) Regards, Dmitrijs. ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Trends in mobile platforms
On 21 January 2013 16:44, Chris Burrell wrote: > > I thought the whole point of the Ubuntu phone OS is that it ran a > full-blown OS... In which won't every front-end theoritically run on it. > When docked, yes you get full ubuntu unity desktop. But when not-docked, i.e. touch screen mode. The supported SDK/UI are Webapps or Qt 5 / QML with specific ubuntu phone touch UI components. For further details on Ubuntu Phone SDK see: http://developer.ubuntu.com/get-started/gomobile/ Regards, Dmitrijs. ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Greek l10n on launhcpad
On 14 January 2013 12:24, Klearchos-Angelos Gkountras wrote: > I want to contribute on Greek ( my native language) into xiphos and I > can't do that , I don't want to download the .po file and but via launchpad > > My LP account is https://launchpad.net/~jemadux > At one point in time the xiphos translations import/export did work in launchpad. But currently it is failing. Therefore unfortunately it's not currently possible to translate xiphos via launchpad. > How to contribute ? Well, you can try using launchpad staging instance and figure out why the translations import is failing. Regards, Dmitrijs. ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] New public git mirror of Sword SVN trunk and why
On 18 December 2012 13:28, Matěj Cepl wrote: > On 18/12/12 01:10, Greg Hellings wrote: >> were mentioned (but Linux packagers will absolute forbid that); once > > That's just plainly not true if that should read "all Linux packagers" > ... see for example > https://fedoraproject.org/wiki/Packaging:NamingGuidelines#Snapshot_packages > (I don't know Debian policy enough anymore, so I cannot point you to the > URL). The world is just not so neat anymore, so large distribution > couldn't survive insisting on having only releases with nice tags. > Debian policy does support for snapshot packaging, where & when appropriate http://www.debian.org/doc/debian-policy/ch-binary.html#s3.2.1 But stable & longer-term supported release are more appropriate for Debian as a rule of thumb. > Yes, many packagers (me included) hugely prefer having released > tarballs, but it does not mean there is no way at all. > But I package stable releases + selective patches. Which goes both ways. I had to disable binding for wheezy release, due to FTBFS with new gcc & no patches available to resolve the issue. Now wheezy is frozen, and new features are not accepted. Oh well, bindings didn't make it into previous release & will not make it into the current one either. > Also, yes, it is hugely preferable to have as few as possible patches in > the distribution packages, but it is not like the end of the world if > there are some. Especially patch which resolves FTBFS on the particular > platform is quite common. > From time to time, I do have to apply debian-specific integration patches, which is fine. On the other hand, multiple releases of sword have not been accepting debian patches to check for status/error codes from api calls > All that saying, yes, where is the list of blockers of the new release? > I would like to see such a list as well. Regards, Dmitrijs. ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] GCC 4.7 compatibility
On 6 August 2012 13:57, Greg Hellings wrote: > It's become more and more difficult to do any work against the SVN > HEAD because it fails to compile on distributions that use gcc 4.7. Is > there any chance someone could land the 4.7 compat patch that is > currently available from those various distros? It is also available > in the gcc47 branch from https://github.com/greg-hellings/sword. > Please, please, please can we do this, so I don't have to try and > cross-merge between different branches to get things working in trunk? > > --Greg patched in debian, fedora and ubuntu. (as far as I know). for gcc4.7 only in debian we have the attached patch. I am not sure if it depends on the patches we have for gcc4.6 and just compiler warnings. Regards, Dmitrijs. fix-ftbfs-gcc4.7.diff Description: Binary data ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
[sword-devel] Removal of ichthux from Ubuntu
Dear all, A request to remove ichthux meta-packages from Ubuntu has been filed. If you have any objections, please reply on: http://pad.lv/1017907 There is no upstream activity, the meta-packages are old and buggy, they are un-installable in the Ubuntu Archive. -- Regards, Dmitrijs. ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Wiki and spam
Hello, I am afraid I cannot answer any of those questions, as I only read Bible in Russian and the names are different. Regards, Dmitrijs. On 04/06/12 17:45, DM Smith wrote: > In getting the wiki back up, I didn't put any anti-spam prevention into > place. > > As a result there were hundreds of new user accounts created in the last > two weeks. I've deleted nearly everyone of them. If I deleted yours, > accept my apology and create it again. > > Additionally, I cleaned up all their spam. > > Previously, we had a simple captcha used on account creation that > presented a simple math problem. I've changed that to a series of simple > tests of biblical knowlege: > Who is the savior of your soul? > Who died for your sins? > Who is Jesus' mother? > Who wrote the Gospel According to Mark? > Who wrote the Gospel According to Matthew? > Who wrote the Gospel According to Luke? > Who wrote the Gospel According to John? > > A random question is picked from those that I supplied. > > If this proves too onerous, let me know. I'll change them to also > provide the answer. > > If you have other ones that have obvious answers that you'd like to add, > let me know. > > In Him, > DM > > > ___ > sword-devel mailing list: sword-devel@crosswire.org > http://www.crosswire.org/mailman/listinfo/sword-devel > Instructions to unsubscribe/change your settings at above page -- Regards, Dmitrijs. signature.asc Description: OpenPGP digital signature ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Test message at 19:34
On 30/05/12 19:34, Peter von Kaehne wrote: > test > roger, out. -- Regards, Dmitrijs. signature.asc Description: OpenPGP digital signature ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] FTBFS on Fedora/x86_64 [Was: Re: Versification scheme for a new Bible translation]
On 30/04/12 22:50, Matěj Cepl wrote:
> On 30.4.2012 19:44, Greg Hellings wrote:
>> They worked for me a few weeks ago, but I only ever build from SVN.
>
> on this more or less Fedora 17, I get this error when running make on
> trunk:
>
> What I am doing wrong?
>
You are using gcc-4.7. See attached patch. Bindings are broken. I do not
have a fix.
--
Regards,
Dmitrijs.
Partially fix ftbfs with gcc-4.7
Index: sword-1.6.2+dfsg/include/multimapwdef.h
===
--- sword-1.6.2+dfsg.orig/include/multimapwdef.h 2004-05-04 21:01:39.0 +
+++ sword-1.6.2+dfsg/include/multimapwdef.h 2012-04-16 20:07:17.256834065 +
@@ -19,14 +19,14 @@
}
T& operator[](const Key& k) {
- if (find(k) == this->end()) {
- insert(value_type(k, T()));
+ if (this->find(k) == this->end()) {
+ this->insert(value_type(k, T()));
}
- return (*(find(k))).second;
+ return (*(this->find(k))).second;
}
bool has(const Key& k, const T &val) const {
- typename std::multimap::const_iterator start = lower_bound(k);
- typename std::multimap::const_iterator end = upper_bound(k);
+ typename std::multimap::const_iterator start = this->lower_bound(k);
+ typename std::multimap::const_iterator end = this->upper_bound(k);
for (; start!=end; start++) {
if (start->second == val)
return true;
signature.asc
Description: OpenPGP digital signature
___
sword-devel mailing list: sword-devel@crosswire.org
http://www.crosswire.org/mailman/listinfo/sword-devel
Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Ubuntu 12.04 ....
On 23 April 2012 15:35, Greg Hellings wrote: > From the CMakeLists.txt file: > > SET(CMAKE_CXX_FLAGS_RELEASE "-Wall -O2 -fexceptions") > SET(CMAKE_CXX_FLAGS_DEBUG "-Wall -Werror -O2 -ggdb -fexceptions") > > It is the -Werror switch, only enabled in Debug builds, that causes > the deprecation warnings to be show stoppers. Last I checked Debian > demands to build with "-Wall -Werror" and debugging flags (Dmitriijs > has thrown The Debian Book at me several times for circumventing that > build policy in my CMakeLists.txt files; apparently Debian knows > better than its application and library developers what flags its > applications and libraries need when building) and will strip out > debugging post-build. So the problem isn't actually BibleTime's but > Debian's. > Hmmm let's just be clean on this = Yes, debian by default builds with debug symbols but we then strip them into separate package. -Wall -Werror is a good measure against newer gcc. Because a warning *now* may turn into an error in gcc+1. (you may have noticed I sometimes yank -Werror out, just to get it build against gcc-snapshot/gcc+1) Regards, Dmitrijs. > Also, because us developers (almost) always build Debug, we don't > build against SWORD SVN and haven't vetted against it as well as we > would like so we don't encourage building BT against SVN. I only ever > build against SVN and haven't noticed a problem, but we still don't > encourage it. > > --Greg > > On Mon, Apr 23, 2012 at 8:46 AM, Troy A. Griffitts > wrote: >> Not disputing the long term desire to not build against deprecated API >> methods (that's why we deprecate), and I certainly understand the policy to >> not release Bibletime until it is not using any deprecated methods of the >> most recent SWORD library, but I can't see how it is useful to cause >> Bibletime to fail building for an end user once it is released. >> >> If >> Bibletime x.y.z is released which enforces this policy against >> SWORD a.b.c and we subsequently release >> SWORD a.b.d which is binary compatible with a.b but which adds deprecation >> notices to additional method, then Bibletime x.y is violating the policy and >> won't build for any user who simply wants to upgrade their SWORD engine. >> >> Could the check be added as a developer build option, so you can still >> enforce it inhouse, but it wouldn't be forced on the end users? >> >> Just a thought. >> >> Troy >> >> >> >> >> On 04/23/2012 03:14 PM, Greg Hellings wrote: >>> >>> On Mon, Apr 23, 2012 at 8:07 AM, Troy A. Griffitts >>> wrote: >>>> >>>> Not disputing any of this, but I'm curious why BT can't compile against >>>> SVN. >>>> I don't believe we've intentionally made any API interface changes in a >>>> while, and we try our hardest to be backward compatible. If you let me >>>> know, I can have a look at what we broke. >>>> >>> >>> BibleTime will not compile against a deprecated API and there is one >>> place where you've deprecated a method in favor of its new camel-cased >>> version. We maintain a git branch with the necessary updates, but we >>> don't want to release that branch as it would break compat with the >>> current released SWORD (which is what we package on Windows and Mac). >>> >>> The refusal to build against deprecated APIs is a policy choice >>> selected by Jaak which the rest of the developers agreed was >>> desirable. >>> >>> --Greg >>> >>>> Troy >>>> >>>> >>>> On 04/23/2012 02:49 PM, Greg Hellings wrote: >>>>> >>>>> >>>>> Xiphos 3.1.5 requires SWORD SVN. In the Crosswire Ubuntu PPA current >>>>> SVN of SWORD is named libsword9. >>>>> >>>>> BibleTime 2.9.1 cannot be built against SWORD SVN but only against >>>>> SWORD 1.6.2. In Ubuntu's current release this version of SWORD is >>>>> named libsword8. >>>>> >>>>> There is a branch in BibleTime's git repository which allows it to be >>>>> built against SWORD's SVN, but because of API changes since then our >>>>> current trunk cannot be built. BibleTime will not release a version >>>>> which is incompatible with the current SWORD release and that is why >>>>> we requested that Troy cut a new release of SWORD. Once SWORD 1.7 &
Re: [sword-devel] Ubuntu 12.04 ....
On 23 April 2012 15:14, Barry Drake wrote: > On 23/04/12 13:49, Greg Hellings wrote: >> >> If you want to get more to the bottom of it, try providing Dmitrijs >> with the information he specified, but it sounds like you are drawing >> the official Ubuntu package of BibleTime and the PPA package of Xiphos >> to get the conflict you indicated. > > > You are right of course. Back when 12.04 was in Alpha, the Xiphos that was > then in the 12.04 repo was broken. I'd temporarily taken Xiphos from the > official ppa, and although I'd removed it (I thought) I was still taking > Xiphos from the official ppa and BibleTime from the launchpad ppa. I ought > to have checked a bit more thoroughly. I've cleared the apt cache now and > I'm certain there are no references to the crosswire ppa. I now have both > BibleTime and Xiphos installed on my netbook which I have just upgraded to > 12.04 and also on my desktop. Sorry to ask a stupid question and > thanks for your patience. > > The information Dmitrijs asked me for did give me the information I needed > to sort the problem. > I'm glad you got sorted it out = Mixing development repositories, can lead to all sort of weird stuff. At one point I had an acient sword build installed into /usr/local and boy did that cause trouble!!! Regards, Dmitrijs. > > God bless, Barry. > > -- > From Barry Drake (The Revd) Health and Healing advisor to the East > Midlands Synod of the United Reformed Church. See > http://www.urc5.org.uk/index for information about the synod, and > http://www.urc5.org.uk/?q=node/703 for the Synod Healing pages. > > Replies - b.dr...@ntlworld.com > > > ___ > sword-devel mailing list: sword-devel@crosswire.org > http://www.crosswire.org/mailman/listinfo/sword-devel > Instructions to unsubscribe/change your settings at above page ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Ubuntu 12.04 ....
On 23 April 2012 13:49, Greg Hellings wrote: > Xiphos 3.1.5 requires SWORD SVN. In the Crosswire Ubuntu PPA current > SVN of SWORD is named libsword9. > Xiphos 3.1.5 does not require Sword SVN, only recommends it. (I did the build script for xiphos) libsword9 is not svn of sword. It is still 1.6.2. The soname version bump was required due to ABI brackage in 1.6.2 errorsly named libsword8. (the ABI incompatability was not noticed early enough basically.) > BibleTime 2.9.1 cannot be built against SWORD SVN but only against > SWORD 1.6.2. In Ubuntu's current release this version of SWORD is > named libsword8. > Note that ubuntu archive has 2.8.1-2build1. It built fine against sword 1.6.2/libsword9 > There is a branch in BibleTime's git repository which allows it to be > built against SWORD's SVN, but because of API changes since then our > current trunk cannot be built. BibleTime will not release a version > which is incompatible with the current SWORD release and that is why > we requested that Troy cut a new release of SWORD. Once SWORD 1.7 > (libsword9 in Ubuntu/Debian parlance) is released the next version of > BibleTime released after that will be compatible with the new API and > will therefore require libsword9. > libsword9 is correct soname in Debian for sword 1.6.2. Pkg-crosswire-devel packaging team has not yet decided on the soname for sword 1.7 yet. And it doesn't matter for now, since sword 1.7 is not released yet. I am inclined to switch to libsword-1.7.0.so now that it is allowed by the debian policy. Since we didn't have a stable abi release, since I got involved in packaging sword for debian. I do not track Bibletime's git, so I have no comments about what it does / doesn't build against. > If you want to get more to the bottom of it, try providing Dmitrijs > with the information he specified, but it sounds like you are drawing > the official Ubuntu package of BibleTime and the PPA package of Xiphos > to get the conflict you indicated. > For precise release: - sword & xiphos have higher version in Ubuntu Archive - bibletime does not I will need to re-upload bibletime 2.9.1 into the PPA or Ubuntu Archive to rebuild against libsword9. Regards, Dmitrijs > --Greg > > On Mon, Apr 23, 2012 at 6:40 AM, Barry Drake wrote: >> On 23/04/12 12:11, Dmitrijs Ledkovs wrote: >>> >>> do not mix PPAs with official archive. >>> >> >> I understand that. What I'm doing is reporting that there is a problem in >> the packages held in Ubuntu 12.04 repo. I'm asking whoever is responsible >> for submissions to Ubuntu to get things up to date as soon as possible. I >> know the official packages are OK but 12.04 is released on Thursday of this >> week so I'm doing some final testing! The new user of 12.04 is NOT going to >> look for the official ppa but will simply use Ubuntu Software Centre - and >> get the conflict. >> >> God bless, Barry. >> >> -- >> Barry Drake is a member of the the Ubuntu Advertising team. >> http://ubuntuadverts.org/ >> >> >> >> ___ >> sword-devel mailing list: sword-devel@crosswire.org >> http://www.crosswire.org/mailman/listinfo/sword-devel >> Instructions to unsubscribe/change your settings at above page > > ___ > sword-devel mailing list: sword-devel@crosswire.org > http://www.crosswire.org/mailman/listinfo/sword-devel > Instructions to unsubscribe/change your settings at above page ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Ubuntu 12.04 ....
On 23 April 2012 12:40, Barry Drake wrote: > On 23/04/12 12:11, Dmitrijs Ledkovs wrote: >> >> do not mix PPAs with official archive. >> > > I understand that. What I'm doing is reporting that there is a problem in > the packages held in Ubuntu 12.04 repo. I'm asking whoever is responsible > for submissions to Ubuntu to get things up to date as soon as possible. I > know the official packages are OK but 12.04 is released on Thursday of this > week so I'm doing some final testing! The new user of 12.04 is NOT going to > look for the official ppa but will simply use Ubuntu Software Centre - and > get the conflict. > You did not provide the information I have asked you for. Unfortunatly I will not be able to assist you further to resolve dependency problem on your machine. The Ubuntu 12.04 archive is correct: https://launchpad.net/ubuntu/precise/amd64/bibletime/2.8.1-2build1 it depends on libsword9. currently PPA is broken, because of the ABI break in the 1.6.2, hence libsword8 is borked in the PPA. You may want to use ppa-purge to test the packages, as they are in the Ubuntu 12.04 archive. This mailing list is not appropriate place for third-party / distribution specific bugs: * bugs in Debian: bugs.debian.org or report-bug * bugs in Ubuntu: launchpad.net/ubuntu or ubuntu-bug * bugs in PPA: pkg-crosswire-de...@lists.alioth.debian.org or launchpad maintainers & respective distribution QA teams monitor above. I only occasionally follow sword-devel. Developers of sword-devel do not upload packages to debian/ubuntu/ppa Regards, Dmitrijs. ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Ubuntu 12.04 ....
do not mix PPAs with official archive. What do $ apt-cache policy bibletime $ apt-cache policy xiphos $ apt-cache policy libsword-dev Say? On 23 April 2012 09:57, Barry Drake wrote: > On 23/04/12 02:46, Karl Kleinpaste wrote: >> >> I wasn't aware, but you'll have to be more specific: What sort of >> dependency conflicts? What does apt-get complain about? > > > Specifically, the packages offered under 12.04 use libsword8 for BibleTime > and libsword9 for Xiphos. If BibleTime is installed first, apt-get gives: > ~$ sudo apt-get install xiphos > Reading package lists... Done > Building dependency tree > Reading state information... Done > The following extra packages will be installed: > libgsf-1-114 libgsf-1-common libsword-common libsword9 xiphos-data > Suggested packages: > sword-text > The following packages will be REMOVED > bibletime libsword8 > The following NEW packages will be installed > libgsf-1-114 libgsf-1-common libsword-common libsword9 xiphos xiphos-data > 0 upgraded, 6 newly installed, 2 to remove and 0 not upgraded. > Need to get 0 B/5,108 kB of archives. > After this operation, 5,543 kB of additional disk space will be used. > Do you want to continue [Y/n]? > > Continuing (obviously) kills BibleTime. If the process is done installing > Xiphos first, apt-get refuses to replace the later libsword with the earlier > one and fails to install. It is clearly a packaging problem - we just need > a packaged version of a later build of BibleTime. > > > God bless, Barry. > > -- > From Barry Drake (The Revd) Health and Healing advisor to the East > Midlands Synod of the United Reformed Church. See > http://www.urc5.org.uk/index for information about the synod, and > http://www.urc5.org.uk/?q=node/703 for the Synod Healing pages. > > Replies - b.dr...@ntlworld.com > > > ___ > sword-devel mailing list: sword-devel@crosswire.org > http://www.crosswire.org/mailman/listinfo/sword-devel > Instructions to unsubscribe/change your settings at above page ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] CMakeLists.txt SWORD_GLOBAL_CONF_DIR
THANK YOU! I had a similar bug in Xiphos, which was not picking up the Global_Conf_Dir in the Xiphos module manager. I will see if I can patch / fix this. Regards, Dmitrijs. On 27 March 2012 11:09, Dominic Tubach wrote: > Hi > > when looking for the reason why the sword locales were not loaded on my > ArchLinux system, I had a look at the CMakeLists.txt. There I found the clause > IF(NOT SWORD_GLOBAL_CONF_DIR STREQUAL "") [l. 104 in version 1.6.2]. This > means if SWORD_GLOBAL_CONF_DIR is not defined at all "/sword.conf" is used for > GLOBCONFPATH. I'm not sure if this is the intended behaviour. I would suggest > to use the following clause: IF(DEFINED SWORD_GLOBAL_CONF_DIR). So if > SWORD_GLOBAL_CONF_DIR is not defined, GLOBCONFPATH will not be defined and the > default specified in the sources will be used. > > (Note: SWORD_GLOBAL_CONF_DIR is not defined in the ArchLinux PKGBUILD for > sword; a moment ago I've created a bug report.) > > Regards > Dominic > ___ > sword-devel mailing list: sword-devel@crosswire.org > http://www.crosswire.org/mailman/listinfo/sword-devel > Instructions to unsubscribe/change your settings at above page -- With regards, Dmitrijs Ledkovs ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] HTML5 front-end
Is Internet, instead if local storage, becomes required? From phone, sorry for top-post. Regards, Dmitrij. On 26 Mar 2012, at 13:24, John Austen wrote: > I just returned from a conference where there was a request from the > Institute for Bible Translation (IBT) to create a web interface to access > their new SWORD repository. The idea is to provide something with similar > capabilities to their SWORD program: "Holy Bible", but instead using their > server to provide cross-browser HTML/Javascript using AJAX. > > So I'm looking into building a PHP extension which will serve as a wrapper > around xulsword's existing dynamic library, which was designed for use with > Javascript. It should be a fairly easy adaptation. If this works out, then > this will be a nice server side SWORD implementation. SWORD modules would > remain on the server, and large local storage would not be needed. > > John > > On 02/29/2012 07:38 PM, Matej Cepl wrote: >> With the advent of Boot2Gecko (and ChromeOS and iPad) it is getting >> increasingly interesting to have HTML5 version of Sword client. I know it is >> very difficult (it would probably require porting of a lot of C/C++ code to >> Javascript, also I have no idea how to store in IndexedDB or some other >> HTML5 storage something so huge as Bible text), but I am interested if >> anybody thought about something like that. >> >> In meanwhile, is there a mobile version of http://www.crosswire.org/study/ ? >> >> Just curious, >> >> Blessings, >> >> Matěj >> >> ___ >> sword-devel mailing list: sword-devel@crosswire.org >> http://www.crosswire.org/mailman/listinfo/sword-devel >> Instructions to unsubscribe/change your settings at above page > > ___ > sword-devel mailing list: sword-devel@crosswire.org > http://www.crosswire.org/mailman/listinfo/sword-devel > Instructions to unsubscribe/change your settings at above page ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] [PATCH] CMake improved buildflags
Yes, please test anywhere you can. What targets do you mean? Please test and provide the options passed to cmake and environmental variables set as well as cmake version. I don't have access to Mac/Windows. So yes, please test as much as possible =) On 22 March 2012 23:57, Greg Hellings wrote: > I'll have to look more closely at this patch when I get back to a > Linux machine. Most of it I will probably apply but it looks like you > failed to account for the fact that there are two separate targets > that need to be set. This should also improve build on Windows and Mac > targets as they require the suffix rules you've included. > > --Greg > > > On Thu, Mar 22, 2012 at 5:58 PM, Dmitrijs Ledkovs > wrote: >> Dear all, >> >> this patch improves cmake buildflags handling >> >> * default flags provided for various configurations >> * environmental flags are appended to the end >> >> this allows to arbitrary experiment with CXXFLAGS / CFLAGS / LDFLAGS >> >> Reviews and comments are welcome. >> >> Please consider applying this to trunk or tell me why not. >> >> -- >> With regards, >> >> Dmitrijs Ledkovs >> >> ___ >> sword-devel mailing list: sword-devel@crosswire.org >> http://www.crosswire.org/mailman/listinfo/sword-devel >> Instructions to unsubscribe/change your settings at above page > > ___ > sword-devel mailing list: sword-devel@crosswire.org > http://www.crosswire.org/mailman/listinfo/sword-devel > Instructions to unsubscribe/change your settings at above page -- With regards, Dmitrijs Ledkovs ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] [PATCH] Unused code
Troy, thank you for taking the time to look at them again. On 22 March 2012 23:07, Troy A. Griffitts wrote: > Dmitrijs, > > Thank you for the patches. I will review your patches again, but this one > specifically I would comment that I'm not in favor of removing commented out > code just to make distro checks pass. There is a reason the code is still > there, though commented out. I didn't write the URL class, but would > venture to thing that the intent was to eventually support '#' mark, and > code has been started but not completed. > - reduces readability of the code - increases complexity - why comment stuff out if there is version control system for this? to e.g. resurrect > Same with the isCommentary mark in osis2mod. I would guess that the > developers have plans to use that eventually to do special processing if the > text is a commentary text rather than a Bible text. > ok. How about stripping this code and replace it with: #TODO add special processing for features X, Y, Z if text is commentary that would be more useful and might prompt someone to write some code. Dead, unused code looks untidy. > I'm excited about the other patches (besides the checking return values > patch which we disagreed on quite a number of months ago and I'll have to > dig up my reasons then, but rather than a theoretical reason, I believe the > patch at the time actually change the logic, which I was not happy with the > new results). Can you please remind me? I do not have all the archives of sword-devel. Maybe we should wrap reads & write into a reusable functions which does sanity checks and provide useful error messages back to the application?! I thought it's good practise to check for errors on IO code. Is there any reason to not check for it? How would you want to check for it? > And again, I believe these are all to make a pedantic distro > code checker happy, yes? > Nothing to do with distro checks. Simply general code quality & pragmatic codebase maintenance. -- With regards, Dmitrijs Ledkovs ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
[sword-devel] [PATCH] Unused code
Dear all,
this patch simply removes some unused code.
Reviews and comments are welcome.
Please consider applying this to trunk or tell me why not.
--
With regards,
Dmitrijs Ledkovs
=== modified file 'src/utilfuns/url.cpp'
--- src/utilfuns/url.cpp 2011-10-25 17:35:20 +
+++ src/utilfuns/url.cpp 2012-03-22 22:39:21 +
@@ -133,7 +133,6 @@
//3.Get the hostname part. This is the part from pos up to the first slash
bool checkPath = true;
bool checkParams = true;
- bool checkAnchor = true;
end = strchr(urlPtr, '/');
if (!end) {
@@ -145,7 +144,6 @@
end = strchr(urlPtr, '#');
}
if (!end) {
- checkAnchor = false;
end = urlPtr+strlen(urlPtr);
}
@@ -160,7 +158,6 @@
end = strchr(urlPtr, '#');
}
if (!end) {
- checkAnchor = false;
end = urlPtr+strlen(urlPtr);
}
@@ -174,15 +171,6 @@
SWBuf paramName;
SWBuf paramValue;
- if (checkAnchor) checkAnchor = false;
-/*
- end = strchr(urlPtr, '#');
- if (!end) {
- checkAnchor = false;
- end = urlPtr+strlen(urlPtr);
- }
-*/
- //end = (start && strchr(start, '?')) ? strchr(start, '?')+1 :0;
end = urlPtr;
while (end) {
paramName = "";
=== modified file 'utilities/osis2mod.cpp'
--- utilities/osis2mod.cpp 2011-11-13 00:33:32 +
+++ utilities/osis2mod.cpp 2012-03-22 22:39:21 +
@@ -1422,7 +1422,6 @@
const char* osisDoc= argv[2];
int append = 0;
SWBuf compType = "";
- bool isCommentary = false;
int iType = 4;
int entrySize = 0;
SWBuf cipherKey= "";
@@ -1470,9 +1469,6 @@
}
usage(*argv, "-s requires one of <2|4>");
}
- else if (!strcmp(argv[i], "-C")) {
- isCommentary = true;
- }
else if (!strcmp(argv[i], "-d")) {
if (i+1 < argc) debug |= atoi(argv[++i]);
else usage(*argv, "-d requires ");
@@ -1480,8 +1476,6 @@
else usage(*argv, (((SWBuf)"Unknown argument: ")+ argv[i]).c_str());
}
- if (isCommentary) isCommentary = true; // avoid unused warning for now
-
if (compType == "ZIP") {
#ifndef EXCLUDEZLIB
compressor = new ZipCompress();
___
sword-devel mailing list: sword-devel@crosswire.org
http://www.crosswire.org/mailman/listinfo/sword-devel
Instructions to unsubscribe/change your settings at above page
[sword-devel] [PATCH] CMake FindICU
Dear all,
this patch improves FindICU for CMake configuration
* --cppflags on my machine returns '-D_REENTRANT -I/usr/include'
This was not handled correctly.
Reviews and comments are welcome.
Please consider applying this to trunk or tell me why not.
--
With regards,
Dmitrijs Ledkovs
=== modified file 'CMakeLists.txt'
--- CMakeLists.txt 2012-03-22 21:39:06 +
+++ CMakeLists.txt 2012-03-22 22:25:00 +
@@ -171,6 +171,7 @@
ENDIF(WITH_CLUCENE)
IF(WITH_ICU)
INCLUDE_DIRECTORIES(${ICU_INCLUDE_DIRS})
+ ADD_DEFINITIONS(${ICU_DEFINITIONS})
SET(SWORD_LINK_LIBRARIES ${SWORD_LINK_LIBRARIES} ${ICU_LIBRARIES} ${ICU_I18N_LIBRARIES})
ENDIF(WITH_ICU)
IF(REGEX_FOUND AND CROSS_COMPILE_MINGW32)
=== modified file 'cmake/FindICU.cmake'
--- cmake/FindICU.cmake 2011-07-11 01:30:52 +
+++ cmake/FindICU.cmake 2012-03-22 22:25:00 +
@@ -5,6 +5,7 @@
# ICU_INCLUDE_DIRS - Directory to include to get ICU headers
# Note: always include ICU headers as, e.g.,
# unicode/utypes.h
+# ICU_DEFINITIONS- Definitions
# ICU_LIBRARIES - Libraries to link against for the common ICU
# ICU_I18N_LIBRARIES - Libraries to link against for ICU internationaliation
# (note: in addition to ICU_LIBRARIES)
@@ -26,14 +27,15 @@
# Get include directories
execute_process(COMMAND "${ICU_CONFIG_BIN}" ${ICU_CONFIG_OPTS} "--cppflags"
-OUTPUT_VARIABLE ICU_INCLUDE_DIR
+OUTPUT_VARIABLE ICU_CPPFLAGS
OUTPUT_STRIP_TRAILING_WHITESPACE
)
# Sanitize output
+ string(REGEX MATCHALL "-I[^ ]*\ |-I[^ ]*$" ICU_INCLUDE_DIR ${ICU_CPPFLAGS})
+ string(REGEX MATCHALL "-D[^ ]*\ |-D[^ ]*$" ICU_DEFINITIONS ${ICU_CPPFLAGS})
string(REPLACE "-I" "" ICU_INCLUDE_DIR ${ICU_INCLUDE_DIR})
- string(REGEX REPLACE "^[ ]+" "" ICU_INCLUDE_DIR ${ICU_INCLUDE_DIR})
- # Try to get the Libraries we need
+ # Try to get the Libraries we need
execute_process(COMMAND "${ICU_CONFIG_BIN}" ${ICU_CONFIG_OPTS} "--ldflags"
OUTPUT_VARIABLE ICU_LIBRARY_RAW
OUTPUT_STRIP_TRAILING_WHITESPACE
___
sword-devel mailing list: sword-devel@crosswire.org
http://www.crosswire.org/mailman/listinfo/sword-devel
Instructions to unsubscribe/change your settings at above page
[sword-devel] [PATCH] CMake improved buildflags
Dear all,
this patch improves cmake buildflags handling
* default flags provided for various configurations
* environmental flags are appended to the end
this allows to arbitrary experiment with CXXFLAGS / CFLAGS / LDFLAGS
Reviews and comments are welcome.
Please consider applying this to trunk or tell me why not.
--
With regards,
Dmitrijs Ledkovs
=== modified file 'CMakeLists.txt'
--- CMakeLists.txt 2011-11-02 14:53:20 +
+++ CMakeLists.txt 2012-03-22 21:39:06 +
@@ -90,19 +90,6 @@
# Some random user settings
#
-IF(SWORD_ENABLE_PROFILE STREQUAL "Yes")
- SET_TARGET_PROPERTIES(sword
- PROPERTIES COMPILE_FLAGS "-pg"
- )
-ENDIF(SWORD_ENABLE_PROFILE STREQUAL "Yes")
-
-IF(SWORD_ENABLE_PROFILEFN STREQUAL "Yes")
- SET_TARGET_PROPERTIES(sword
- PROPERTIES COMPILE_FLAGS "-g -finstrument-functions"
- )
- TARGET_LINK_LIBRARIES(libsword fnccheck)
-ENDIF(SWORD_ENABLE_PROFILEFN STREQUAL "Yes")
-
IF(NOT SWORD_GLOBAL_CONF_DIR STREQUAL "")
ADD_DEFINITIONS(-DGLOBCONFPATH="${SWORD_GLOBAL_CONF_DIR}/sword.conf")
ENDIF(NOT SWORD_GLOBAL_CONF_DIR STREQUAL "")
@@ -125,33 +112,41 @@
ENDIF(BUILDING_SHARED)
###
-# Some options are only needed if we're going to be building a debug option into the library
-# These are generally only for developer building and testing
-#
-# Debug testing
-IF(CMAKE_BUILD_TYPE STREQUAL "Debug")
- SET(SWORD_CFLAGS "-g3 -Wall -Werror -O0")
-ELSE(CMAKE_BUILD_TYPE STREQUAL "Debug")
- SET(SWORD_CFLAGS "-O3")
-ENDIF(CMAKE_BUILD_TYPE STREQUAL "Debug")
+# The buildflags depend on the CMAKE_BUILD_TYPE supplied at the command line.
+# For a full lists of different flags see http://cmake.org/Wiki/CMake_Useful_Variables
+#
+# Cmake has sane defaults for None aka '', DEBUG, RELEASE, RELWITHDEBINFO for known compilers.
+#
+# You can override these variables on the command-line or here.
+# We provide our own defaults below
+#
+# Note the below two initialisations done for us by cmake:
+#
+#SET(CMAKE_C_FLAGS "${CMAKE_C_FLAGS}" or "$ENV{CFLAGS}" or "" )
+#SET(CMAKE_CXX_FLAGS"${CMAKE_CXX_FLAGS}" or "$ENV{CXXFLAGS}" or "" )
IF(SWORD_ENABLE_WARNINGS STREQUAL "Yes")
- SET(SWORD_CFLAGS "${SWORD_CFLAGS} -Werror")
+ SET(CMAKE_C_FLAGS "-Werror ${CMAKE_C_FLAGS}")
+ SET(CMAKE_CXX_FLAGS "-Werror ${CMAKE_CXX_FLAGS}")
ENDIF(SWORD_ENABLE_WARNINGS STREQUAL "Yes")
-IF(BUILDING_SHARED)
- SET_TARGET_PROPERTIES(sword
- PROPERTIES COMPILE_FLAGS ${SWORD_CFLAGS}
- )
-ENDIF(BUILDING_SHARED)
-
-IF(BUILDING_STATIC)
- SET_TARGET_PROPERTIES(sword_static
- PROPERTIES COMPILE_FLAGS ${SWORD_CFLAGS}
- )
-ENDIF(BUILDING_STATIC)
-
-MESSAGE(STATUS "Setting CFlags to ${SWORD_CFLAGS}")
+IF(SWORD_ENABLE_PROFILE STREQUAL "Yes")
+ SET(CMAKE_C_FLAGS "-pg ${CMAKE_C_FLAGS}")
+ SET(CMAKE_CXX_FLAGS "-pg ${CMAKE_CXX_FLAGS}")
+ENDIF(SWORD_ENABLE_PROFILE STREQUAL "Yes")
+
+IF(SWORD_ENABLE_PROFILEFN STREQUAL "Yes")
+ SET(CMAKE_C_FLAGS "-g -finstrument-functions ${CMAKE_C_FLAGS}")
+ SET(CMAKE_CXX_FLAGS "-g -finstrument-functions ${CMAKE_CXX_FLAGS}")
+ TARGET_LINK_LIBRARIES(libsword fnccheck)
+ENDIF(SWORD_ENABLE_PROFILEFN STREQUAL "Yes")
+
+SET(CMAKE_C_FLAGS_DEBUG"-g3 -Wall -O0 ${CMAKE_C_FLAGS}")
+SET(CMAKE_C_FLAGS_RELEASE "-O3 ${CMAKE_C_FLAGS}")
+SET(CMAKE_C_FLAGS_RELWITHDEBINFO "-O3 -g ${CMAKE_C_FLAGS}")
+SET(CMAKE_CXX_FLAGS_DEBUG "-g3 -Wall -O0 ${CMAKE_CXX_FLAGS}")
+SET(CMAKE_CXX_FLAGS_RELEASE"-O3 ${CMAKE_CXX_FLAGS}")
+SET(CMAKE_CXX_FLAGS_RELWITHDEBINFO "-O3 -g ${CMAKE_CXX_FLAGS}")
##
# Setting libraries and includes
___
sword-devel mailing list: sword-devel@crosswire.org
http://www.crosswire.org/mailman/listinfo/sword-devel
Instructions to unsubscribe/change your settings at above page
[sword-devel] [PATCH] Checking the return value of read/write functions
Dear all,
Here is the updated patch against trunk which checks the return values
of various IO operations.
Reviews, comments are welcome.
Please consider applying this to trunk or tell me why not.
--
With regards,
Dmitrijs Ledkovs
=== modified file 'src/utilfuns/zlib/untgz.c'
--- src/utilfuns/zlib/untgz.c 2001-03-23 09:00:15 +
+++ src/utilfuns/zlib/untgz.c 2012-03-22 19:32:07 +
@@ -13,6 +13,8 @@
#include
#ifdef unix
# include
+# include
+# include
#else
# include
# include
@@ -277,7 +279,7 @@
intremaining = 0;
FILE *outfile = NULL;
char fname[BLOCKSIZE];
- time_t tartime;
+ time_t tartime = 0;
while (1) {
len = gzread(in, &buffer, BLOCKSIZE);
=== modified file 'tests/testblocks.cpp'
--- tests/testblocks.cpp 2009-02-07 08:42:00 +
+++ tests/testblocks.cpp 2012-03-22 19:32:07 +
@@ -29,14 +29,18 @@
void addEntry(EntriesBlock *eb) {
string input;
string body;
- char line[1024];
+ char line[1024], *c;
std::cout << "\nEnter new Entry's text. '.' on an empty line to finish:\n";
do {
std::cout << "> ";
- fgets(line, 1000, stdin);
- input = line;
- if (input.compare("."))
- body.append(input);
+ c = fgets(line, 1000, stdin);
+ if (c == NULL) {
+ std::cerr << "ERROR: fgets failed during addEntry\n";
+ } else {
+ input = line;
+ if (input.compare("."))
+body.append(input);
+ }
}
while (input.compare("."));
std::cout << "Adding new entry. Index is: " << eb->addEntry(body.c_str()) << "\n\n";
@@ -72,29 +76,33 @@
EntriesBlock *eb = new EntriesBlock();
string input;
- char line[1024];
+ char line[1024], *c;
std::cout << "Initial entry count should be 0: " << eb->getCount() << "\n";
do {
std::cout << "[" << eb->getCount() << "] > ";
- fgets(line, 1000, stdin);
- input = line;
- if (input.length() > 0) {
- switch (input[0]) {
-case 'a': addEntry(eb); break;
-case 'p': printEntry(eb, atoi(input.c_str()+1)); break;
-case 'r': removeEntry(eb, atoi(input.c_str()+1)); break;
-case 's': printSize(eb); break;
-case 'q': break;
-case '?':
-default:
- std::cout << "\n a - add a new entry\n";
- std::cout << " p - print entry\n";
- std::cout << " r - remove entry\n";
- std::cout << " s - print size of raw data\n";
- std::cout << " q - quit\n\n";
- break;
+ c = fgets(line, 1000, stdin);
+ if (c == NULL) {
+ std::cerr << "ERROR: fgets failed in main\n";
+ } else {
+ input = line;
+ if (input.length() > 0) {
+switch (input[0]) {
+ case 'a': addEntry(eb); break;
+ case 'p': printEntry(eb, atoi(input.c_str()+1)); break;
+ case 'r': removeEntry(eb, atoi(input.c_str()+1)); break;
+ case 's': printSize(eb); break;
+ case 'q': break;
+ case '?':
+ default:
+ std::cout << "\n a - add a new entry\n";
+ std::cout << " p - print entry\n";
+ std::cout << " r - remove entry\n";
+ std::cout << " s - print size of raw data\n";
+ std::cout << " q - quit\n\n";
+ break;
+}
}
}
}
=== modified file 'utilities/cipherraw.cpp'
--- utilities/cipherraw.cpp 2011-07-23 17:39:20 +
+++ utilities/cipherraw.cpp 2012-03-22 19:32:07 +
@@ -48,7 +48,7 @@
SWCipher *zobj;
VerseKey key;
RawVerse *rawdrv;
- int ofd[2], oxfd[2];
+ int ofd[2], oxfd[2], w;
long tmpoff = 0, offset, loffset = 0, lzoffset = 0;
unsigned short size, lsize = 0, lzsize;
char *tmpbuf;
@@ -74,10 +74,18 @@
delete [] tmpbuf;
printf("\n");
- write(oxfd[0], &lzoffset, 4);
- write(oxfd[0], &lzsize, 2);
- write(oxfd[1], &lzoffset, 4);
- write(oxfd[1], &lzsize, 2);
+ w = write(oxfd[0], &lzoffset, 4);
+ if (w < 0)
+ perror("ERROR: write failed in main");
+ w = write(oxfd[0], &lzsize, 2);
+ if (w < 0)
+ perror("ERROR: write failed in main");
+ w = write(oxfd[1], &lzoffset, 4);
+ if (w < 0)
+ perror("ERROR: write failed in main");
+ w = write(oxfd[1], &lzsize, 2);
+ if (w < 0)
+ perror("ERROR: write failed in main");
key.AutoNormalize(0);
key.Headings(1);
@@ -89,8 +97,13 @@
printf("using previous offset,size %d\n", size);
offset = lseek(oxfd[key.Testament() - 1], 0, SEEK_CUR);
printf("%ld %ld %d \n", offset, lzoffset, lzsize);
- write(oxfd[key.Testament() - 1], &lzoffset, 4);
- write(oxfd[key.Testament() - 1], &lzsize, 2);
+ w = write(ox
Re: [sword-devel] Xiphos 3.1.5 is released
Can you run xml2po manually? It seems to be incomplete cause the error
is comming from xml2po itself of not being able to import its python
module.
>From my iphone.
Regards,
Dmitrij.
On 20 Feb 2012, at 10:23, "Steven P. Ulrick" wrote:
> On 02/19/2012 08:48 PM, Karl Kleinpaste wrote:
>> After entirely too long a delay, we have finally released Xiphos 3.1.5.
>>
>> Windows users can get the installer immediately:
>> http://sourceforge.net/projects/gnomesword/files/Xiphos/3.1.5/xiphos-3.1.5-win32.exe/download
>> Or get the source tarball:
>> http://sourceforge.net/projects/gnomesword/files/Xiphos/3.1.5/xiphos-3.1.5.tar.gz/download
>> Or get it from SVN:
>> svn co
>> https://gnomesword.svn.sourceforge.net/svnroot/gnomesword/branches/webkit
>>
>> Building against latest Sword SVN is straightforward.
>> Building against not-the-latest requires that you add
>> -DNO_SWORD_SET_RENDER_NOTE_NUMBERS=1.
>>
>> Comments and bug reports welcome.
>
> Hello, Karl
> When I attempt to build Xiphos from SVN, I get the following error if I use
> the included "waf":
> [ 11/130] po: po/ar.po -> build/default/po/ar.mo
> Traceback (most recent call last):
> File "/usr/bin/xml2po", line 191, in
>main(sys.argv[1:])
> File "/usr/bin/xml2po", line 88, in main
>from xml2po import Main
> ImportError: No module named xml2po
> Traceback (most recent call last):
> File "/usr/bin/xml2po", line 191, in
>main(sys.argv[1:])
> File "/usr/bin/xml2po", line 88, in main
>from xml2po import Main
> ImportError: No module named xml2po
> Waf: Leaving directory `/home/steve/webkit/build'
> Build failed:
> -> task failed (err #1):
>{task: xml2po fr.po,xiphos.xml -> xiphos.xml}
> -> task failed (err #1):
>{task: xml2po fa.po,xiphos.xml -> xiphos.xml}
> steve@localhost webkit$
>
>
>
> "/usr/bin/xml2po" is installed:
>
> steve@localhost ~$ rpm -q --whatprovides /usr/bin/xml2po
> gnome-doc-utils-0.20.6-1.fc16.noarch
> steve@localhost ~$
>
> If you need any further information, let me know.
>
> Steven P. Ulrick
>
>
> ___
> sword-devel mailing list: sword-devel@crosswire.org
> http://www.crosswire.org/mailman/listinfo/sword-devel
> Instructions to unsubscribe/change your settings at above page
___
sword-devel mailing list: sword-devel@crosswire.org
http://www.crosswire.org/mailman/listinfo/sword-devel
Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] BibleTime problem under Precise ....
On 21 December 2011 11:46, Barry Drake wrote: > Hi there ... I have a slight problem with BibleTime 2.8.4 running under > Ubuntu Precise Pangolin. I can't use the Ubuntu bug tracker as I got the > package from Crosswire as the one supplied from the Ubuntu repository has > not yet been updated. It's best to use pkg-crosswire-devel mailing list, it's a mailing list of maintainers of Debian, Ubuntu & PPA packages. It should be possible to report bugs against PPA, I thought launchpad started to support that Need to check. Which of the 4 crashes from these http://bugs.debian.org/cgi-bin/pkgreport.cgi?pkg=bibletime;dist=unstable is most similar to yours? We are currently trying to figure out what is going on there. -- With regards, Dmitrijs Ledkovs ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] BPBible 0.5 released
On 21 December 2011 10:53, Jonathan Morgan wrote: > Hi David, > > On Tue, Dec 20, 2011 at 12:29 AM, David Haslam > wrote: >> >> Just a reminder that BPBible is not the only front-end using XulRunner. >> See http://code.google.com/p/xulsword/ >> >> The developer of xulsword is already aware of the expressed concerns. > > > The two are actually quite different. BPBible is embedding XULRunner for > use in HTML rendering, but otherwise using wxWidgets for UI. XULSword is > actually written in XUL as a XULRunner app. > As far as I know XULRunner apps are still supported in Debian/Ubuntu. Embedding XULRunner is not. I think my previous email about it came across very brisk. I apologise for that. I wanted to point out that xulrunner embedding engine has been aggressively dropped from Ubuntu/Debian archives (oneiric and up) Such that as a user it is not easy to run the app; as a developer not easy to package as a deb. It is dropped from Ubuntu/Debian due to: * mozilla dropping xulrunner embedding * decreasing release cycle to 3-4 months * decreasing support timeframe to 6-8 months In long term it will be harder and harder to get xulrunner embedding up and running. -- With regards, Dmitrijs Ledkovs ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] BPBible 0.5 released
On 19 December 2011 16:30, Jonathan Morgan wrote: > Hi Greg, > > On Mon, Dec 19, 2011 at 9:53 PM, Greg Hellings > wrote: >> >> Jonathan, >> >> Just so I keep track in my brain, in versions <0.5 you had wxHTML as >> your rendering engine, yes? And what are you upgrading to now? > > > You are correct that the old rendering engine was wxHTML. > The new one is XULRunner (at the moment, XR 1.9.2 - approx FF 3.6, though I > hope that newer versions will be able to be used without too much trouble). > Please do not upload it to debian nor ubuntu then. XULRunner embedding is deprecated and not supported any more in Ubuntu and is phased out in debian. -- With regards, Dmitrijs Ledkovs ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] [Pkg-crosswire-devel] Packagers Take Note
On 9 December 2011 20:06, Greg Hellings wrote: > On Fri, Dec 9, 2011 at 1:49 PM, Dmitrijs Ledkovs > wrote: >> On 2 November 2011 15:06, Greg Hellings wrote: >>> I don't know all the appropriate emails for SWORD packagers, but >>> figured this is as good a place as any to put out this notice. >>> >>> SWORD SVN contains a few commits which should be interesting from the >>> point of view of packagers for Linux distributions. While the current >>> SVN is not directly compatible with the latest release of SWORD >>> (1.6.2), interested parties should take note of commits 2661, 2662 and >>> 2665 when packaging. These commits introduce compatibility with >>> CLucene version 2.x while maintaining backwards compatibility with >>> version 0.9x. Additionally, commit 2665 fixes a major bug in the >>> CMake build chain. Versions of the library without commit 2665 which >>> are built with CMake will detect CLucene during configure but will not >>> build against it, leaving the library with limited search >>> capabilities. >>> >> >> Although the patches are good, they do no apply to 1.6.2 cleanly. >> >> They depend on other commits (changes to cmake, addition of util functions) >> On top of that they introduce more gcc-4.6 warnings which make it >> failt to build from source. >> I've looked at it for a little, but it's too much work for now for me. > > r2665 should apply cleanly, I thought. Or, at least, proper building > against CLucene 0.9x only relies on the line in that commit > ADD_DEFINITIONS(-DUSELUCENE). > > You won't gain CLucene2 compatibility with it. > > Also - you should fix Debian, because it builds great outside of > Debian's whatever-in-the-world business. ;) > Please elaborate. I did not understand. Is there a bug in packaging or did you make a joke a did not get? (I'm still not a native speaker ;-) ) >> >> Clean patch series against 1.6.2 would be highly appreciated (I know >> I'm asking too much, but hey it's christmas soon ;-) ) >> Or for example 1.6.3 release ;-) > > A 1.6.3 release would be very difficult as current SVN head contains > many things which are not bug fixes and which break compatibility and > really should constitute a 1.7. Troy isn't against someone doing a > 1.6.3 release, but that person would need to sift through the > 1.6.2..HEAD commits and cherry pick the bug-fix-only ones. > ok. > --Greg > >> >>> Anyone who used CMake in the past or who packages for a system with >>> CLucene 2 is encouraged to take a look at these three commits and >>> backport them to your patch system until the next release of the SWORD >>> library which ought to include these commits. >>> >>> Thanks, >>> Greg >>> >>> ___ >>> Pkg-crosswire-devel mailing list >>> pkg-crosswire-de...@lists.alioth.debian.org >>> http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-crosswire-devel ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] [Pkg-crosswire-devel] Packagers Take Note
On 2 November 2011 15:06, Greg Hellings wrote: > I don't know all the appropriate emails for SWORD packagers, but > figured this is as good a place as any to put out this notice. > > SWORD SVN contains a few commits which should be interesting from the > point of view of packagers for Linux distributions. While the current > SVN is not directly compatible with the latest release of SWORD > (1.6.2), interested parties should take note of commits 2661, 2662 and > 2665 when packaging. These commits introduce compatibility with > CLucene version 2.x while maintaining backwards compatibility with > version 0.9x. Additionally, commit 2665 fixes a major bug in the > CMake build chain. Versions of the library without commit 2665 which > are built with CMake will detect CLucene during configure but will not > build against it, leaving the library with limited search > capabilities. > Although the patches are good, they do no apply to 1.6.2 cleanly. They depend on other commits (changes to cmake, addition of util functions) On top of that they introduce more gcc-4.6 warnings which make it failt to build from source. I've looked at it for a little, but it's too much work for now for me. Clean patch series against 1.6.2 would be highly appreciated (I know I'm asking too much, but hey it's christmas soon ;-) ) Or for example 1.6.3 release ;-) > Anyone who used CMake in the past or who packages for a system with > CLucene 2 is encouraged to take a look at these three commits and > backport them to your patch system until the next release of the SWORD > library which ought to include these commits. > > Thanks, > Greg > > ___ > Pkg-crosswire-devel mailing list > pkg-crosswire-de...@lists.alioth.debian.org > http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-crosswire-devel ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Python Bindings
On 24 November 2011 05:41, Greg Hellings wrote: > $ apt-cache showpkg swig > Package: swigVersions: 2.0.4+really1.3.40-3ubuntu4 > While 2.0.4 was uploaded sometime in the past into the archive, it was decided to go back to 1.3.40 (downgrade). Regards, Dmitrijs. ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Discussion on sword inclusion on openSUSE mailing list
On 17 September 2011 12:56, Benjamin Misja wrote: > Nice reminder! I did know about that "law" some time ago. Still, is there > anything that can be done to make things clearer, like standard texts for > packagers etc.? Example wording: "sword - a shared library and utilities to read and process e-books, maps, commentaries and dictionaries in various formats including vpl, etc..." while it is designed to fit biblical texts, nothing prevents from encoding any other type of literary texts using these formats. Ps. similar discussion were raised when I proposed sword+xiphos/bibletime for default inclusion into Edubuntu (look in edubuntu-devel mailing list archives 1-3 years old, can't be bothered to find a link now). It was relatively calm, but straight away the answer was "will not include by default, too biased or very sensitive topic". At least it is available directly from the archive ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] incorrect characters in Latvian module
Heya, Which module has errors? I can sort them out or help with getting it to txt and back =) kura modulaa tu atradi kludas? Es varu tos izmainit. paldies. On 5 September 2011 14:03, Martins Donga wrote: > Hi I am using annBible on my android phone and have problems with Latvian > characters. > Probably I can correct this issue if the text would be in txt or similar > file format. > I am very interested to see this problem fixed. > hope to hear from you soon. > Best regards > Martins Donga > > -- Forwarded message -- > From: And Bible Help > Date: 2011/9/5 > Subject: Re: And Bible - Latvian letters? > To: Martins Donga > > > Hi Martins, > I am not directly involved with Bible module creation or maintenance. The > Bible module creation is done by Crosswire Sword and it would be best if you > posted to the sword-devel mailing list that they use. They would be very > happy to hear from you. > Best regards > Martin > I am using AndBilble on my Android phone and have problem with Latvian > Character > On 5 September 2011 01:17, Martins Donga wrote: >> >> Hi Martins, >> How can I help to fix this problem? I saw error report. >> Probably I could fix those incorrect characters if that would be txt or >> similar format. >> I really wanna see Latvian text legible. :) >> best regards >> Martins >> >> 2011/8/29 And Bible Help >>> >>> Hi Martins & Gerda, >>> Thanks for this error report. >>> I have investigated this problem and it appears to be a problem with the >>> Latvian module rather than with And Bible so I have raised an error report >>> against the module. >>> Best regards >>> Martin >>> >>> On 23 August 2011 17:09, Martins Donga wrote: Thank you very much for AND BIBLE!!! I use it every day. Could you please change the encoding of Latvian New Testament? The letters in Latvian look wrong. This is how it looks : Matthew 1:1 Jçzus Kristus, Dâvida dçla, Âbrahama dçla, cilts grâmata. This is how it should be: Jēzus Kristus, Dāvida dēla, Ābrahāma dēla, cilts grāmata. Thx & God bless your team :-) Martins & Gerda Donga >> >> >> > > > ___ > sword-devel mailing list: sword-devel@crosswire.org > http://www.crosswire.org/mailman/listinfo/sword-devel > Instructions to unsubscribe/change your settings at above page > ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
[sword-devel] gcc-4.6 warnings
Dear all, gcc-4.6 has additional warning enabled by default, most notably -Wunused-but-set-variable here are two patches: one for sword itself, and the other one for the zlib (untgz code) they can be applied against trunk with $ patch -p1 < *.patch Regards, Dmitrijs. gcc-4.6-fix-sword.patch Description: Binary data gcc-4.6-fix-zlib.patch Description: Binary data ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Russian russynodal (prize for maximum use of negation?)
On 10 July 2011 10:56, Chris Little wrote: > On 7/10/2011 12:56 AM, Jonathan Marsden wrote: >> >> On 07/09/2011 10:07 PM, Chris Little wrote: >> >>> Doesn't And Bible not support av11n? If it doesn't, I would not >> >> >> That's a *lot* of "doesn't" and "not" -- rewriting this with fewer >> negations, for ease of parsing, would be good. I think you just won a >> prize for maximum use of negation? :) >> >> Jonathan > > Negation in English can be used to make speech more indirect and, as a > consequence, is a politeness strategy. Stated more directly/less politely, I > might have said, "I know full well that And Bible doesn't support av11n, so > I would not" (Of course, even that retains politeness via use of the > subjunctive.) It seems quite apropos to me, in any case, to use as much > negation as possible when the subject is Russian, where multiple negation > has quite distinct semantics from English. > Multiple negation is a great tool to express yourself in Russian and everyone uses it, unfortunately it doesn't translate to English as expected in many cases. ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
[sword-devel] Bits from Debian Crosswire Packaging Team
Hello everyone! Here are some updates from Debian Crosswire Packaging team. We have three PPA's. Stable, Developer Preview and Daily Builds. All three of these are now build automatically by launchpad using recipes. All recipes can be viewed here: https://code.launchpad.net/~pkgcrosswire/+recipes = Developer preview PPA = * Preview the latest available Work-in-progress packaging * Uses released versions of upstream software (possibly patched) * Builds for currently supported Ubuntu Releases * Build are automatic Example is sword-stable recipe here: https://code.launchpad.net/~pkgcrosswire/+recipe/sword-stable You can clearly see which branches are used (just our packaging branch lp:sword/debian aka lp:~pkgcrosswire/sword/main) as well as the build results for each Ubuntu Release and architecture. When we are packaging software we only need to do `$ bzr push` and launchpad will build updated packages for all four currently supported Ubuntu releases. No need to manually change package versions, recreate debian source packages and uploading them to launchpad. Huray =) = Daily Builds = * Automatically does daily builds of upstream svn trunk * Uses latest available, "official" packaging * Builds for all currently supported Ubuntu Releases * Builds are not tested and not supported Example is sword-daily recipe here https://code.launchpad.net/~pkgcrosswire/+recipe/sword-daily This recipe is based on lp:sword - a bzr mirror of sword svn trunk (pulled multiples times a day) and it is merged with our packaging. A "bridge" branch is used to set correct ancestry and it will eventually go away when launchpad will support nest-part command. On this page you can see latest build results for Ubuntu releases and supported architectures. = Stable PPA = * Manually transitioned packages from Developer-Preview * Well Tested * Use "official" packaging * Builds for all currently supported Ubuntu Releases * Limited support available on best effort basis This stays as it is, except that it is much easier for us to update now. We can simply perform a manual build request from the stable recipe page web-page. = What pages should you look at? = Sword Code Import: https://code.launchpad.net/~sword-devel/sword/trunk Daily Builds (latest): https://code.launchpad.net/~pkgcrosswire/+recipe/sword-daily Developer Preview (latest): https://code.launchpad.net/~pkgcrosswire/+recipe/sword-stable Xiphos Code Import: https://code.launchpad.net/~xiphos-devel/xiphos/trunk Daily Builds (latest): https://code.launchpad.net/~pkgcrosswire/+recipe/xiphos-daily Developer Preview (latest): https://code.launchpad.net/~pkgcrosswire/+recipe/xiphos-stable BibleTime Code Import: https://code.launchpad.net/~bibletime/bibletime/gitorious Daily Builds (latest): https://code.launchpad.net/~pkgcrosswire/+recipe/bibletime-daily (NOT ENABLED YET) Developer Preview (latest): https://code.launchpad.net/~pkgcrosswire/+recipe/bibletime-stable PPA build results: Daily (all): https://launchpad.net/~pkgcrosswire/+archive/daily/+packages Developer-Preview (all): https://launchpad.net/~pkgcrosswire/+archive/developer-testing/+packages Stable (all): https://launchpad.net/~pkgcrosswire/+archive/ppa/+packages How to add these PPA: Daily (all): $ sudo add-apt-repository ppa:pkgcrosswire/daily Developer-Preview (all): $ sudo add-apt-repository ppa:pkgcrosswire/developer-testing Stable (all): $ sudo add-apt-repository ppa:pkgcrosswire/ppa = About Debian Pkgcrosswire Packaging Team = We are packaging Bible study related packages based on SWORD API and Crosswire Bible Society repository of Bible texts. We are official maintainers of these packages in Debian and Ubuntu. We provide regular backports updates to Ubuntu via PPA as well as offering work-in-progress packages. Our packages adhere to Debian Packaging Policy. All our builds provide debug (-dbg) packages as well to help with troubleshooting ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
[sword-devel] CMake & Install paths
= ICU = With autoconf icu / transliterations files were installed into: /usr/lib/sword/1.6.1_icu_4.2.1 With CMake they end up in: /usr/lib/1.6.1_icu_4.2.1 = locales.d, mods.d & modules = With CMake I'm missing: /etc/sword.conf /usr/share/sword/mods.d/globals.conf /usr/share/sword/locales.d/*.conf It's not hard to install these by-hand, but I would prefer them to be installed automagically. Thanks. ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Xiphos 3.1.4 source tarball is released
On 22 October 2010 17:55, David Haslam wrote: > > It's not just a matter of strict dependencies. > > Xiphos for Windows also comes packed with all the Sword Win32 utilities. > > As there have been some bugs fixed in these between 1.6.1 and 1.6.2 - it > would be frustrating for Xiphos users to not have these as part of the > Xiphos 3.1.4 installation. > > I think it is wise for the Xiphos team to await 1.6.2 to be ready in place > before rolling out the download. > > David > Good thinking! If you have a windows machine and spare time please attempt to create a release for windows. I will gladly try to help with any issues that you might encounter. > > -- > View this message in context: > http://sword-dev.350566.n4.nabble.com/Xiphos-3-1-4-source-tarball-is-released-tp2997831p3007333.html > Sent from the SWORD Dev mailing list archive at Nabble.com. > > ___ > sword-devel mailing list: sword-devel@crosswire.org > http://www.crosswire.org/mailman/listinfo/sword-devel > Instructions to unsubscribe/change your settings at above page > ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] using a dvcs for module creation
On 16 September 2010 15:44, Peter von Kaehne wrote: > Thanks to Matthew who has got me onto the idea to use a DVCS for module > creation. This is a very useful idea particularly if the underlying > texts are messy and need gradual adjustment at a variety of places. > > So, I fell for git, largely because the name appealed to me and it works > so far quite well. > > My question: If I want to push intermittently point releases - like "NT > completed", "with book introductions", "with footnotes" etc to a central > server of what do I need to have for that? I.e. what does the remote > server need installed? I want to set up my home server as a backup etc > for that. > Remote server will need ssh & git daemon running + set-up a bare repository on the server: e.g. $ git init --bare tools.git Then on your machine you need to add remote, fetch it and then push the repo. > And how can I use git to mess about with code at home but then commit to > svn - e.g. the sword-tools section on CrossWire? You will have to create a git-svn clone & fidle with git-svn cache and meta data. It takes a couple of tries to get the import right. After that you can use your git repo to fiddle around and commit, but there will be no roundback since you will have to rebase each time you commit to svn and hence do force updates. I recommend to install bzr & bzr-svn. Then it's a matter of $ bzr branch svn://example.com/trunk/subdir $ bzr push sftp://just.sftp.on.the.host/without/any/daemons (findling) $ bzr push svn://just.push.to/svn/without/troubles/and/rebases (pushing/pulling from svn) > At the moment I have > two directories - one is sword-tools under svn where I check out from > CrossWire and commit from too and another is a git based one. Whenever I > think i have something worth pushing to CrossWire, I copy from one to > the other and do a svn commit. I am sure this is not the most clever way... > git-svn automates this, but it's pretty much what it does. Bzr is a bit smarter when talking to svn and it remembers what went were instead of doing blind rebases like git. If you google up git-svn guides from KDE, GCC, Webkit you will get more hang of it. They have everything covered. > Thanks > > Peter > > ___ > sword-devel mailing list: sword-devel@crosswire.org > http://www.crosswire.org/mailman/listinfo/sword-devel > Instructions to unsubscribe/change your settings at above page > ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Sword 1.6.2 NOW!
On 14 September 2010 19:46, Troy A. Griffitts wrote: > Thanks for the email link Greg. Yeah, the submitted changes in question > from project A were filter updates, and we do specifically state in that > email that filter updates are allowed in a stable branch. It was kindof an > odd situation. The updates merely added css classes to a few of the html > tags which are outputted from the filters, none of the projects involved > thought adding css classes would break anyone. > > One of the things in head currently which isn't mentioned in that email is > binding improvements and also your additional cmake make system. I'm not > sure how I feel about binding improvements being included in a stable > branch, but I don't see an issue including an additional make system. Any > thoughts? > cmake stuff is not finished yet =/ there are still quite a bit of stuff that needs to be tested. Xiphos is not on a verge of a new release yet. /me changing hats Debian & Ubuntu do not particularly see any urgency in a sword release and rather wait for cmake to stabilise (at least I want CPack which can create tarballs) > Troy > > On 9/14/2010 6:49 PM, Greg Hellings wrote: >> >> On Tue, Sep 14, 2010 at 12:43 PM, Troy A. Griffitts >> wrote: >>> >>> Understood going forward. >>> >>> There were a few factors which made this a slightly non-standard >>> situation. >>> First, this wasn't exactly a bug fix. It was a workaround for a bug in >>> a >>> version of libcurl. I was hoping libcurl would be patched. And no, I >>> personally didn't report the issue to the curl team, which I should have. >>> Secondly, one of our frontend projects submitted a small update which >>> changed something another of our frontends depended on. The second >>> project >>> had updated their code to still work with the new change, but they hadn't >>> released yet. If they had released then everyone would be happy with us >>> releasing SVN as is. As it stands right now 1 of the 2 projects will >>> need >>> to patch SVN for their frontend to work. So delaying was a hopeful but >>> unfruitful exercise. It was a choice we made to with the best >>> information >>> we had at the time. >> >> Not knowing the nature of the changes, etc, I don't mean to provide >> this as a comment on that, but I'd just like to bring back up this >> email: >> >> http://www.crosswire.org/pipermail/sword-devel/2009-June/032108.html >> >> and see if that's still the plan? If you're actually changing how >> things are working inside (in the sense of enhancing for new modules >> and content like the NASB and not just for fixing bugs), then maybe it >> is time to branch and allow for bug fixing/feature branches to develop >> separately until 1.7 is made? >> >> --Greg >> >> ___ >> sword-devel mailing list: sword-devel@crosswire.org >> http://www.crosswire.org/mailman/listinfo/sword-devel >> Instructions to unsubscribe/change your settings at above page > > > ___ > sword-devel mailing list: sword-devel@crosswire.org > http://www.crosswire.org/mailman/listinfo/sword-devel > Instructions to unsubscribe/change your settings at above page > ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Sword 1.6.2 NOW!
I know for sure that Fedora (a few F11+ i think) Debian (squeeze+) and Ubuntu (well Karmic+ if you consider PPA or Lucid+ from main archive) are all patched. This is way to small for a bugfix release (I'd rather see more personally). This issue was not present a the time sword 1.6.1 was released, curl was the culprit. Just this bugfix is not worth a branch, a release, nothing. It was widely discussed on this mailing list, the patch is readily available from svn, most distributors have applyed it, and it doesn't affect statickly linked builds which are used (as far as I know) on Windows, Mac, Android, iOS (all of these together probably outweigh desktop distribution based Linux). This particular fix has been deployed in one of the OBS project for OpenSUSE users. I cannot comment about any source based distros as I don't follow their development. ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Managing Bible translation
On 6 September 2010 07:09, Teus Benschop wrote: >> > I am thinking of ways to manage Bible translation well in >> > git/bzr/svn or >> > whatever other VCS or DVCS. > There is xml2po tool designed in Gnome which extracts translatable strings from XML and generates "pot" template and "po" translation files. This can then be stored in a bzr branch and imported into launchpad.net. That way both bzr can be used to translate offline using familiar tools e.g. poedit. As well translations can be done online using Launchpad.net website. xml2po might need changes to process OSIS better. ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Android SWORD
On 2 September 2010 18:47, Manfred Bergmann wrote: > > Am 02.09.2010 um 18:18 schrieb Dmitrijs Ledkovs: > >> On 2 September 2010 17:28, Matthew Talbert wrote: >>> On Thu, Sep 2, 2010 at 5:31 AM, Manfred Bergmann >>> wrote: >>>> >>>> I don't know but isn't it contra productive and actually a lot more work >>>> to maintain the sources on a variety on source control platforms? >>>> We end up having SWORD source trees on Github, Bitbucket and Launchpad and >>>> eventually on any other system that will come up. >>>> For all of those we need people to maintain the sources otherwise we will >>>> have a mess of different sources version and implementations. >>>> In my opinion, the main and official sources of SWORD have to be >>>> maintained by responsible people from Crosswire. >>>> And it has to be clear that Crosswire will have and is in control of the >>>> official sources. I don't actually care which source control system is >>>> used for that, though I think that Subversion is the greatest common >>>> divisor because all other DVCSs can bi-directional access Subversion. >>>> >>> >>> As far as I know, only Launchpad is being used (other than svn). It is >>> quite helpful, because it allows people to publish branches of code >>> and work on them together to finish a feature before it is committed >>> to sword svn. If there is a proliferation of source control systems >>> being used, I'm unaware of it. >>> >>> Matthew >>> >> >> Plus on Launchpad it is clearly stated where the upstream svn >> repository is. The bzr imports are run every 6 hours for many sword >> projects (jsword, sword, xiphos, bibletime). And bzr is fully >> compatible with svn, i.e. you can seamlessly commit to svn from bzr if >> you have access. > > That is totally fine. > The use case you are outlining is comprehensible. Btw: MacSword is mirrored > on Launchpad as well. > Who is maintaining the Launchpad repos? I mean, one of the branches probably > is used to push back to SWORD svn? > > > Manfred I and Jonathan Marsden try to keep up with them from ~pkg-crosswire-devel Debian/Ubuntu packaging team. The mirrored bzr branches are read-only. If you have commit access to svn you can $ bzr push svn://svn.example.com/trunk. And then it will ask for login/password. Anyone can create, commit and push bzr branches to launchad =) ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Android SWORD
On 2 September 2010 18:47, Manfred Bergmann wrote: > > Am 02.09.2010 um 18:18 schrieb Dmitrijs Ledkovs: > >> On 2 September 2010 17:28, Matthew Talbert wrote: >>> On Thu, Sep 2, 2010 at 5:31 AM, Manfred Bergmann >>> wrote: >>>> >>>> I don't know but isn't it contra productive and actually a lot more work >>>> to maintain the sources on a variety on source control platforms? >>>> We end up having SWORD source trees on Github, Bitbucket and Launchpad and >>>> eventually on any other system that will come up. >>>> For all of those we need people to maintain the sources otherwise we will >>>> have a mess of different sources version and implementations. >>>> In my opinion, the main and official sources of SWORD have to be >>>> maintained by responsible people from Crosswire. >>>> And it has to be clear that Crosswire will have and is in control of the >>>> official sources. I don't actually care which source control system is >>>> used for that, though I think that Subversion is the greatest common >>>> divisor because all other DVCSs can bi-directional access Subversion. >>>> >>> >>> As far as I know, only Launchpad is being used (other than svn). It is >>> quite helpful, because it allows people to publish branches of code >>> and work on them together to finish a feature before it is committed >>> to sword svn. If there is a proliferation of source control systems >>> being used, I'm unaware of it. >>> >>> Matthew >>> >> >> Plus on Launchpad it is clearly stated where the upstream svn >> repository is. The bzr imports are run every 6 hours for many sword >> projects (jsword, sword, xiphos, bibletime). And bzr is fully >> compatible with svn, i.e. you can seamlessly commit to svn from bzr if >> you have access. > > That is totally fine. > The use case you are outlining is comprehensible. Btw: MacSword is mirrored > on Launchpad as well. > Who is maintaining the Launchpad repos? I mean, one of the branches probably > is used to push back to SWORD svn? > > > Manfred > Does it make sence to create crosswire meta projects for all of the sword related projects? ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Android SWORD
On 2 September 2010 17:28, Matthew Talbert wrote: > On Thu, Sep 2, 2010 at 5:31 AM, Manfred Bergmann > wrote: >> >> I don't know but isn't it contra productive and actually a lot more work to >> maintain the sources on a variety on source control platforms? >> We end up having SWORD source trees on Github, Bitbucket and Launchpad and >> eventually on any other system that will come up. >> For all of those we need people to maintain the sources otherwise we will >> have a mess of different sources version and implementations. >> In my opinion, the main and official sources of SWORD have to be maintained >> by responsible people from Crosswire. >> And it has to be clear that Crosswire will have and is in control of the >> official sources. I don't actually care which source control system is used >> for that, though I think that Subversion is the greatest common divisor >> because all other DVCSs can bi-directional access Subversion. >> > > As far as I know, only Launchpad is being used (other than svn). It is > quite helpful, because it allows people to publish branches of code > and work on them together to finish a feature before it is committed > to sword svn. If there is a proliferation of source control systems > being used, I'm unaware of it. > > Matthew > Plus on Launchpad it is clearly stated where the upstream svn repository is. The bzr imports are run every 6 hours for many sword projects (jsword, sword, xiphos, bibletime). And bzr is fully compatible with svn, i.e. you can seamlessly commit to svn from bzr if you have access. ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] PocketSword Mercurial repo
On 26 August 2010 10:29, Nic Carter wrote: > > Hi all. > > I was asked where PocketSword is now located, source-wise, and so I'm going > to post this to the list so that it's archived here for everyone to play > with. :) > > You can browse the project via web browser at: > http://bitbucket.org/niccarter/pocketsword/overview > > Or to grab the source and tinker, use: > $ hg clone http://bitbucket.org/niccarter/pocketsword > > I'm the only one with write access atm, but I'm the only one working on it > atm, so that's not really an issue... :( > Anyone can clone the repository on bitbucket, and commit/push their changes to bitbucker and then you can easily merge any proposed merges / patch-queues =) So it's not an issue at all =) Horray for using hg! ;-) > > Let me know if there are any issues and I'll see what I can do. :) > > > Thanks, ybic > nic... :) > > > Nic Carter > PocketSword Developer - an iPhone Bible Study app > www: http://crosswire.org/pocketsword > iTunes: http://itunes.apple.com/app/Pocketsword/id341046078 > Twitter: http://twitter.com/pocketsword > > > ___ > sword-devel mailing list: sword-devel@crosswire.org > http://www.crosswire.org/mailman/listinfo/sword-devel > Instructions to unsubscribe/change your settings at above page > ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
[sword-devel] Launchpad: libsword -> sword
FYI Launchpad.net project name for sword has been changed from "libsword" to "sword". Libsword has been kept as an alias. >From now on you can use: https://launchpad.net/sword -> to access project pages $ bzr branch lp:sword -> to get bzr branch of the svn trunk import NB NB sword does not use launchpad for upstream development NB It is used primarly for debian/ubuntu packaging of sword, and to provide automatic bzr mirror of the svn code repository. Anyone can branch and publish their personal branches of sword on launchpad. Similar "abuse" of launchpad is done by xiphos, bibletime & bibledit. For more information about launchpad and ways it can be used for sword related development, you can contact pkg-crosswire-devel debian packaging team. ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Building Clucene along with SWORD lib
On 25 July 2010 21:58, Manfred Bergmann wrote: > > Am 25.07.2010 um 21:15 schrieb Dmitrijs Ledkovs: > >> On 25 July 2010 21:13, Dmitrijs Ledkovs wrote: >>> On 25 July 2010 17:32, Manfred Bergmann wrote: >>>> Hi. >>>> >>>> I'm no automake or make specialist. >>>> Does anyone know a simple way to put Clucene sources into some folder of >>>> SWORD so that both are built/linked together? >>>> >>>> >>> >> >> Guessing from @me.com I assume you are on mac =) > > Yep, I'm on Mac. :) > >> install macports http://www.macports.org/ >> >> and then >> >> $ sudo port install clucene >> >> will do the trick =) of fetching and compiling clucene and all dependencies. > > Well, I can build SWORD against clucene either via Fink or Macports or even > compile it manually. > But I need to distribute a universal binary SWORD lib including clucene build > via makefile. > I could manually compile clucene three times and build an ub in an install > folder but that's some effort. > I've also experimented with the monolothic clucene which produces a > monolithic clucene object file. This is very easy to produce, basically just > a manual g++ call with some options. > But how can I then integrate this .o file so that SWORD will use it at link > stage? > Check ports variants clucene to see if universal variant is available. If not build it yourself as universal binary. Then you will also be able to build sword as universal binary. As long as clucene is universal and you build sword as universal as well it should all just work. Here: http://www.crosswire.org/pipermail/sword-devel/2010-June/034399.html is example how to configure sword to build as universal library straight off the bat. Hope this helps. > > Manfred > > > ___ > sword-devel mailing list: sword-devel@crosswire.org > http://www.crosswire.org/mailman/listinfo/sword-devel > Instructions to unsubscribe/change your settings at above page > ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Building Clucene along with SWORD lib
On 25 July 2010 21:13, Dmitrijs Ledkovs wrote: > On 25 July 2010 17:32, Manfred Bergmann wrote: >> Hi. >> >> I'm no automake or make specialist. >> Does anyone know a simple way to put Clucene sources into some folder of >> SWORD so that both are built/linked together? >> >> > Guessing from @me.com I assume you are on mac =) install macports http://www.macports.org/ and then $ sudo port install clucene will do the trick =) of fetching and compiling clucene and all dependencies. >> >> Manfred >> >> >> ___ >> sword-devel mailing list: sword-devel@crosswire.org >> http://www.crosswire.org/mailman/listinfo/sword-devel >> Instructions to unsubscribe/change your settings at above page >> > ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Building Clucene along with SWORD lib
On 25 July 2010 17:32, Manfred Bergmann wrote: > Hi. > > I'm no automake or make specialist. > Does anyone know a simple way to put Clucene sources into some folder of > SWORD so that both are built/linked together? > > As far as I can see there is nothing in sword's automake to support building clucene in a combined tree. You need to fetch clucene yourself and build it, install in same prefix and build sword against it. If you are on mac / linux you can install from the repositories. > > Manfred > > > ___ > sword-devel mailing list: sword-devel@crosswire.org > http://www.crosswire.org/mailman/listinfo/sword-devel > Instructions to unsubscribe/change your settings at above page > ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] OS X/iPhoneOS based development
On 15 June 2010 15:59, Greg Hellings wrote: > On Tue, Jun 15, 2010 at 9:46 AM, Dmitrijs Ledkovs > wrote: >> On 15 June 2010 14:27, Greg Hellings wrote: >>> On Mon, Jun 14, 2010 at 5:52 PM, Dmitrijs Ledkovs >>> I have included build/configuration files for a number of the >>> different iPhone SDK packs with help from >>> http://sites.google.com/site/michaelsafyan/coding/resources/how-to-guides/cross-compile-for-the-iphone/how-to-cross-compile-for-the-iphone-using-cmake. >>> The so-called "architecture files" that can be set with either the >>> CMake GUI or command-line arguments are found in the cmake/toolchains/ >> >> What's CMake GUI? and where do I get one? > > http://www.cmake.org/cmake/resources/software.html > > If you download the .dmg for OS X or the installer for Windows, you'll > get the GUI version of the program. If you're using a package > manager, most of them have a cmake-gui target, or similar. > >>> I'd be happy to work with you on doing anything like that you need. I >>> don't know too much about creating Frameworks or building the >>> bindings. I believe the current state of the CMake project files has >>> feature parity with the autotools in building the library, utilities, >>> demos and tests but I hadn't touched the bindings yet, because I had >>> no idea how those are built. Also, I wanted to get feedback from >>> others on how well the current base abilities were working for them. >>> >> >> The idea about bindings is: >> >> 1) Build sword library and headers >> 2) Take bindings pixy dust (swig) and run swig to regenerate itself >> against sword library >> 3) Take the output from swig and compile the bindings (e.g. python >> module) install & use >> >> I don't know how many swig bindings are currently in working/updated >> state. If you find any CMake/swig info please post links =) > > AFAIK the Corba and Python bindings are in regular use by the SWORDWeb > and the BP Bible developers respectively. Beyond that I'm not sure, > either. > > CMake has had support for SWIG since version 2.0, with an example in > their wiki at > http://www.cmake.org/Wiki/CMake_FAQ#How_do_I_use_CMake_to_generate_SWIG_wrapper_libraries.3F > > I'll see what I can do about adding support for that. Does anyone > else know of other SWIG wrappers, or just other wrappers, that are > actively in use? > >> >>> Feel free to push a new branch to Launchpad if you have fixes or >>> additions or to mail me patches or anything else like that. I can >>> also be found on IRC many days, though depending on work and such I >>> might be more or less available. >>> >> >> lp:sword/cmake now points to your branch =) > > lp:sword/cmake or lp:libsword/cmake? When I first signed up for > Launchpad I flagged myself as working with Sword, only to find out > that it's "Semantic Web on Rails development". > Yes. lp:libsword/cmake sorry =) sword is still taken = > --Greg > > ___ > sword-devel mailing list: sword-devel@crosswire.org > http://www.crosswire.org/mailman/listinfo/sword-devel > Instructions to unsubscribe/change your settings at above page > ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] OS X/iPhoneOS based development
On 15 June 2010 14:27, Greg Hellings wrote: > On Mon, Jun 14, 2010 at 5:52 PM, Dmitrijs Ledkovs > wrote: >> WOW >> >> I'm not quite sure but it's either CMake is amazing or Xcode is >> amazing or both of them totally rock. And probably Gregory for doing >> awesome job in doing this port. > > Dmitrijs, glad to hear you're finding it useful! You're the first > feedback I've received regarding it, so it's always nice to get > feedback, especially if the first feedback is positive! > >> >> it looks like it is doing out of the tree builds so i can have >> Makefile cmake generator in one dir and xcode generator in another dir >> and use them by mix and matching. > > This comes as a minimal-effort feature of CMake and is encouraged as > the preferred way to work with CMake. That way, if something goes > wrong during configuration or build you easily just rm -rf the > offending directory and start from scratch if you can't rectify the > situation. For SWORD I made out-of-tree mandatory with CMake. > >> >> From faq of cmake there are loads of simple parameters you can set to >> get it xcode projects for any arches/sdk's you want/need. > > I have included build/configuration files for a number of the > different iPhone SDK packs with help from > http://sites.google.com/site/michaelsafyan/coding/resources/how-to-guides/cross-compile-for-the-iphone/how-to-cross-compile-for-the-iphone-using-cmake. > The so-called "architecture files" that can be set with either the > CMake GUI or command-line arguments are found in the cmake/toolchains/ What's CMake GUI? and where do I get one? > directory. When I tested, they worked for all the 3.0+ versions that > you see in that directory. Hints for what command-line arguments to > use can be found in cmake/configure-iphone.sh and > cmake/configure-iphone-simulator.sh. I don't know how well the > resulting libraries will work, but I do know that I have made them > compile in Xcode using those settings. > >> >> Furthermore the generated xcode project is in subdir but it doesn't >> copy over source files instead they are referenced to the source dir >> from where cmake was run. So anything I edit in xcode is editing the >> svn checkout. > > The same goes for Visual Studio and Unix makefile builds with CMake > (and, I would presume, any other systems CMake supports that are > capable of it). It makes it very nice if you have multiple > configurations (e.g. Mac OS X, iPhone Simulator and iPhone Device) in > 3 different directories and you want to test how a single change would > affect all three, etc etc. A very nice feature, indeed! > >> >> The source tree in xcode project looks about the same as the checkout. >> But the targets are structured wonderfully. You get the dynlib target, >> when you click on it it shows all the source files which are needed to >> build that target only. And every single little tool has it's own >> target and it depends on libsword target and lists only it's own >> sourcecode. Naturally there is the almightly "build everything" target >> =) > > It will even generate a "make install" or equivalent on platforms that > even remotely support it (OS X, Linux, Windows, etc). > >> >> The best bit is probably that the xcode project is linked back to >> cmake. Each target has pre & post rule in xcode to talk back to cmake >> and reconfigure on the fly. >> >> And finally the actually compile uses Xcode built-in rules so I get >> all xcode features, as if I defined the xcode project from scratch / >> by hand. So i get xcode precompiled headers, pre-binding, caching and >> parallel builds and can run debugger, DTrace and all that jazz against >> it. >> >> If CMake does the same for Visual Studio then hands-down cmake rocks >> and sword really needs it if we want to seriously push into mobile >> device. > > CMake does the same for Visual Studio and Unix makefile projects, even > down to the point of auto-reconfiguring if one of the CMake files has > changed since the last time a file was loaded. The only awkward spot > is in Visual Studio the build needs to stop after the auto reconfigure > and the user is prompted to reload the project because the file has > changed on disk. For makefiles they're smart enough to pick up the > changes directly (that was in 2.6 - version 2.8 might have fixed that > awkward step in Visual Studio). I see this as one of the biggest > improvements over autotools from a user's experience - even if you > don't remember the full steps of a
Re: [sword-devel] OS X/iPhoneOS based development
On 15 June 2010 00:30, Nic Carter wrote: > Hehe - now you see why Manfred and I quite enjoy Xcode... ;) > yeah it looks promising. All I can think about is how to run DTrace (aka Instruments) against everything I build to magically show me where to optimize my code =) > I am personally not interested in a nice framework like you are talking > about, as PocketSword wants to be as small as possible given it is installed > on a mobile device. So I don't need the cross-platform & -arch stuff that > you want to do. But I love the sound of it. :) the point is, if done right it will be easy to add objc wrappers around it. (BTW what is the state of the Objc++ bridge?) You can always build optimised framework targetting just e.g. iOS4 it would be just another build configuration =) and by building it from same source, i.e. same / similar xcode project. Are you doing MacSword as well? > But, apple have tried to make Xcode work nicely with make/cmake tools & the > cmake guys have then taken advantage of that and made cmake do what you have > seen. :) > > Sorry I'm coming in at the end of this discussion - timezones can be > interesting! But I'll try to help answer any other questions you have about > Xcode - you could also take a look at the PocketSword setup on bitbucket? But > now that you can see a nice and big and complicated Xcode project, hopefully > that helps you understand how it all works. :) > bitbucket =) hg nice. Rumor has it Xcode 4 has git integration, but it's just whispernet as everyone who has it is under NDA. > > Thanks, ybic > nic... :) > > ps: Manfred: I owe you several emails - sorry, will try to reply when I get > home tonight! :) > > On 15/06/2010, at 8:52, Dmitrijs Ledkovs wrote: > >> WOW >> >> I'm not quite sure but it's either CMake is amazing or Xcode is >> amazing or both of them totally rock. And probably Gregory for doing >> awesome job in doing this port. >> >> it looks like it is doing out of the tree builds so i can have >> Makefile cmake generator in one dir and xcode generator in another dir >> and use them by mix and matching. >> >> From faq of cmake there are loads of simple parameters you can set to >> get it xcode projects for any arches/sdk's you want/need. >> >> Furthermore the generated xcode project is in subdir but it doesn't >> copy over source files instead they are referenced to the source dir >> from where cmake was run. So anything I edit in xcode is editing the >> svn checkout. >> >> The source tree in xcode project looks about the same as the checkout. >> But the targets are structured wonderfully. You get the dynlib target, >> when you click on it it shows all the source files which are needed to >> build that target only. And every single little tool has it's own >> target and it depends on libsword target and lists only it's own >> sourcecode. Naturally there is the almightly "build everything" target >> =) >> >> The best bit is probably that the xcode project is linked back to >> cmake. Each target has pre & post rule in xcode to talk back to cmake >> and reconfigure on the fly. >> >> And finally the actually compile uses Xcode built-in rules so I get >> all xcode features, as if I defined the xcode project from scratch / >> by hand. So i get xcode precompiled headers, pre-binding, caching and >> parallel builds and can run debugger, DTrace and all that jazz against >> it. >> >> If CMake does the same for Visual Studio then hands-down cmake rocks >> and sword really needs it if we want to seriously push into mobile >> device. >> >> I think it will be easy enough to create my dream sword.framework >> using this, since by the looks of it I'll just define a separate >> project, import the library from this cmake-xcode project and play >> around with it to turn in into a framework. Although I'm gonna start >> with cmake first cause probably we can tweak it into turning >> sword.dynlib into framework =) >> >> /me is probably gonna switch xiphos to cmake as well =)) for the >> sake of buidling on mac at least =) >> >> ___ >> sword-devel mailing list: sword-devel@crosswire.org >> http://www.crosswire.org/mailman/listinfo/sword-devel >> Instructions to unsubscribe/change your settings at above page > > ___ > sword-devel mailing list: sword-devel@crosswire.org > http://www.crosswire.org/mailman/listinfo/sword-devel > Instructions to unsubscribe/change your settings at above page > ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] OS X/iPhoneOS based development
WOW I'm not quite sure but it's either CMake is amazing or Xcode is amazing or both of them totally rock. And probably Gregory for doing awesome job in doing this port. it looks like it is doing out of the tree builds so i can have Makefile cmake generator in one dir and xcode generator in another dir and use them by mix and matching. >From faq of cmake there are loads of simple parameters you can set to get it xcode projects for any arches/sdk's you want/need. Furthermore the generated xcode project is in subdir but it doesn't copy over source files instead they are referenced to the source dir from where cmake was run. So anything I edit in xcode is editing the svn checkout. The source tree in xcode project looks about the same as the checkout. But the targets are structured wonderfully. You get the dynlib target, when you click on it it shows all the source files which are needed to build that target only. And every single little tool has it's own target and it depends on libsword target and lists only it's own sourcecode. Naturally there is the almightly "build everything" target =) The best bit is probably that the xcode project is linked back to cmake. Each target has pre & post rule in xcode to talk back to cmake and reconfigure on the fly. And finally the actually compile uses Xcode built-in rules so I get all xcode features, as if I defined the xcode project from scratch / by hand. So i get xcode precompiled headers, pre-binding, caching and parallel builds and can run debugger, DTrace and all that jazz against it. If CMake does the same for Visual Studio then hands-down cmake rocks and sword really needs it if we want to seriously push into mobile device. I think it will be easy enough to create my dream sword.framework using this, since by the looks of it I'll just define a separate project, import the library from this cmake-xcode project and play around with it to turn in into a framework. Although I'm gonna start with cmake first cause probably we can tweak it into turning sword.dynlib into framework =) /me is probably gonna switch xiphos to cmake as well =)) for the sake of buidling on mac at least =) ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] OS X/iPhoneOS based development
On 14 June 2010 22:23, Manfred Bergmann wrote: > > Am 14.06.2010 um 22:01 schrieb Dmitrijs Ledkovs: > >>> For me I think it's time to switch now from manually build sword library to >>> Xcode project because now the library is build together with the ObjC >>> wrapper classes. >>> >> >> Fair enough. How are you keeping up with changes to sword? What is >> your xcode project layout? > > I don't know about changes to sword. There are actually none. :) Sorry, english is not native language here. I meant updates in sword - i.e. $ svn up If there are more files added do you manually readd them in xcode to build sword from xcode? > Xcode project includes the sword/src folder (minus some files that would not > compile like all sources in frontend, dirent for win32 and such) and clucene > as groups. > In build settings you add header file folders sword/include, icu, ... and > linker flags for libcurl, libicu, ... > Finally you add CFLAGS actually the same as in makefile. > Meah =) i won't be able to do that for gtk, glib, pango, gtkhtml, xulrunner. (i'm trying to nail xiphos build on mac) >> I didn't manage to make a project which can get have sword framework >> and xiphos getting pulled in from svn. I would also need Gtk, Glib, >> Pango and friends i have framework versions of those, but it's >> "imported" frameworks. > > I'm not exactly sure what you mean? to build xiphos i need sword, gtk, glib, gtkhtml, libxml at least. how would you go about defining all of these in the project. > At the moment there is only one Xcode target that builds everything, together > with Sword ObjC wrappers. In the xcode project tree what is the difference between "executables" and "targets". I think i really don't understand what xcode means by "target". > It would make sense to split this to have a separate targets for the Sword > dylib and wrapper. what I'm trying to understand how to package all headers & dylib into universal framework. I've seen quite a few mac specific projects which ship framework and as far as I understand it uses some kind of weak linking and relocation code. these frameworks can either be in /Library/Frameworks, ~/Library/Frameworks or inside an app-bundle. To top this off with one such package I was able to mix & match, i.e. it didn't matter where the framework was "thin" app-bundles continued to work, but bundles which had framework embedded inside them where as if statically linked and didn't take notice of the same framework installed in */Library/Frameworks. Ideally i'd like to be able to painlessly create pre-compiled sword.framework with dynlibs, headers & documentation compiled for all arches including iphone / iOS, throw it onto ftp.crosswire.org. Then anyone could grab it and start developing for mac or iDevices and e.g. xiphos and macsword could share it on the user's system saving space. (with an option to embed sword framework inside app-bundle as well) Is this what xcode calls target? Ideally sword.framework would contain icu & curl but not expose those headers/symbols to prevent abuse =) (e.g. using sword.framework to build a freaky app which actually just uses icu). In otherwords make sword.framework self-contained, flexible and abi/api stable for everyone who ports or write apps for apple platforms. > >> >>>>> How is this done with CMake? >>>>> >>>> >>>> As far as I can see CMake style source/targets are just converted into >>>> XCode xml/plst project file, so you would still need to use xcode to >>>> define debug/release/sdk/target-arches. >>> >>> "xcodebuild" can build Xcode projects via command line and accepts command >>> line build options or you can pass in a settings file. >>> >> >> Thanks, didn't know that. >> >>>> Using the sdk and gcc isysroot imho is far more UNIX-like =) >>> >>> I don't know why CMake would not be UNIX-like? >>> >> >> Don't understand CMake is a great tool, so is autotools. it's just >> I find it easier to control and inspect what is going on during >> buildtime using lower level tools. E.g. Xcode is far to abstract for >> my taste. Or maybe I just don't know how to use it properly. I never >> found enough time to use/learn it because it's not as universal as >> autotools. > > Yeah, that's right. Xcode is Mac specific and relatively abstract. However as > is generally with Mac it just hides it. You can see all relevant information. > Problems about your build or the exact gcc com
Re: [sword-devel] OS X/iPhoneOS based development
On 14 June 2010 21:38, Manfred Bergmann wrote: > > Am 14.06.2010 um 21:03 schrieb Dmitrijs Ledkovs: > >> On 14 June 2010 18:59, Manfred Bergmann wrote: >>> Hi Greg. >>> >>> I was always using a makefile compiled library. However creating >>> multi-architecture builds (i386, x86_64, ppc) was always kind of a hassle. >> >> ppc64 as well for G5 macs =) >> >>> This is very easily done by Xcode because it compiles for all targets and >>> combines the builds and I'm actually quite happy that I can use this now. >> >> FUD =) >> >> you can use isysroots to compile universal binaries using apple >> toolchain the same way XCode does it. >> >> http://developer.apple.com/mac/library/technotes/tn2005/tn2137.html > > It's called "Xcode" not "XCode". > Thanks for your pointers but I'm already doing this. > This has gotten much easier since you don't need to build for Mac OSX 10.3 > because you needed gcc 3.3 for that and gcc 4 for 10.4 and above. > True. I did do bundles packing 10.3 & 10.4 together. But that was before intel transition. > For me I think it's time to switch now from manually build sword library to > Xcode project because now the library is build together with the ObjC wrapper > classes. > Fair enough. How are you keeping up with changes to sword? What is your xcode project layout? I didn't manage to make a project which can get have sword framework and xiphos getting pulled in from svn. I would also need Gtk, Glib, Pango and friends i have framework versions of those, but it's "imported" frameworks. >>> How is this done with CMake? >>> >> >> As far as I can see CMake style source/targets are just converted into >> XCode xml/plst project file, so you would still need to use xcode to >> define debug/release/sdk/target-arches. > > "xcodebuild" can build Xcode projects via command line and accepts command > line build options or you can pass in a settings file. > Thanks, didn't know that. >> Using the sdk and gcc isysroot imho is far more UNIX-like =) > > I don't know why CMake would not be UNIX-like? > Don't understand CMake is a great tool, so is autotools. it's just I find it easier to control and inspect what is going on during buildtime using lower level tools. E.g. Xcode is far to abstract for my taste. Or maybe I just don't know how to use it properly. I never found enough time to use/learn it because it's not as universal as autotools. > > Manfred > >>> >>> Am 14.06.2010 um 18:51 schrieb Gregory Hellings: >>> >>>> Manfred, >>>> >>>> My CMake build system is supposed to support iPhone configurations. It >>>> would keep you from having to update your xcode poject every time files >>>> are changed, etc. I have built the library succesfully against the >>>> simulator files using my CMake configure, but have not had any takers on >>>> actually testing the library against an actual app. >>>> >>>> --Greg >>>> >>>> >>>> >>>> On Jun 14, 2010, at 11:52 AM, Manfred Bergmann >>>> wrote: >>>> >>>>> Troy, I'd like to put this into this bindings folder. >>>>> If you could create a "objc" folder there and give me commit rights? I'm >>>>> almost finished creating a proper Xcode project. >>>>> >>>>> >>>>> >>>>> Manfred >>>>> >>>>> >>>>> Am 16.04.2010 um 07:58 schrieb Troy A. Griffitts: >>>>> >>>>>> These all sound like good suggestions. Manfred, let me know what you >>>>>> decide and I'll either give you access to sword/bindings/objc or create >>>>>> a repository for you. >>>>>> >>>>>> Troy >>>>>> >>>>>> >>>>>> Peter von Kaehne wrote: >>>>>>> Would not the natural place be the bindings directory in the engine? >>>>>>> >>>>>>> >>>>>>> >>>>>>> Nic Carter wrote: >>>>>>>> I suggest you put it in a specific place in your MacSword SVN (or >>>>>>>> Bazaar if that's what you use now?) and I can access it from there. :) >>>>>>>> >>>>>>>> On 16/04/2010, at 2:32 AM, Manfred Bergmann wrote: >>>>>>>
Re: [sword-devel] OS X/iPhoneOS based development
On 14 June 2010 18:59, Manfred Bergmann wrote: > Hi Greg. > > I was always using a makefile compiled library. However creating > multi-architecture builds (i386, x86_64, ppc) was always kind of a hassle. ppc64 as well for G5 macs =) > This is very easily done by Xcode because it compiles for all targets and > combines the builds and I'm actually quite happy that I can use this now. FUD =) you can use isysroots to compile universal binaries using apple toolchain the same way XCode does it. http://developer.apple.com/mac/library/technotes/tn2005/tn2137.html I've used 10.5 sdk but the 10.4u (NB u = universal) can be used as well, 10.6 sdk doesn't have ppc. I took current trunk, autoreconf -f -i (NB using apple autoreconf not anything you get via macports/fink) and did this: CXXFLAGS="-mmacosx-version-min=10.5 -isysroot /Developer/SDKs/MacOSX10.5.sdk -arch i386 -arch x86_64 -arch ppc -arch ppc64" CFLAGS="-mmacosx-version-min=10.5 -isysroot /Developer/SDKs/MacOSX10.5.sdk -arch i386 -arch x86_64 -arch ppc -arch ppc64" LDFLAGS="-arch i386 -arch x86_64 -arch ppc -arch ppc64" ./configure --disable-dependency-tracking --with-icusword --without-icu --prefix=/here (bogus prefix to make sure I don't pull in any non-universal dylibs i have there) $ file lib/.libs/libsword-1.6.1.dylib lib/.libs/libsword-1.6.1.dylib: Mach-O universal binary with 4 architectures lib/.libs/libsword-1.6.1.dylib (for architecture i386): Mach-O dynamically linked shared library i386 lib/.libs/libsword-1.6.1.dylib (for architecture x86_64): Mach-O 64-bit dynamically linked shared library x86_64 lib/.libs/libsword-1.6.1.dylib (for architecture ppc7400): Mach-O dynamically linked shared library ppc lib/.libs/libsword-1.6.1.dylib (for architecture ppc64):Mach-O 64-bit dynamically linked shared library ppc64 I'm not sure how to wrap this into a proper Mac OS X framework and whether there is need for that at all. Using similar options you should be able to compile for iPhone SDK, I don't have it installed. On the other note do you think 10.4u install base is still there? And do we need to support it. > How is this done with CMake? > As far as I can see CMake style source/targets are just converted into XCode xml/plst project file, so you would still need to use xcode to define debug/release/sdk/target-arches. Using the sdk and gcc isysroot imho is far more UNIX-like =) > > Manfred > > > Am 14.06.2010 um 18:51 schrieb Gregory Hellings: > >> Manfred, >> >> My CMake build system is supposed to support iPhone configurations. It would >> keep you from having to update your xcode poject every time files are >> changed, etc. I have built the library succesfully against the simulator >> files using my CMake configure, but have not had any takers on actually >> testing the library against an actual app. >> >> --Greg >> >> >> >> On Jun 14, 2010, at 11:52 AM, Manfred Bergmann >> wrote: >> >>> Troy, I'd like to put this into this bindings folder. >>> If you could create a "objc" folder there and give me commit rights? I'm >>> almost finished creating a proper Xcode project. >>> >>> >>> >>> Manfred >>> >>> >>> Am 16.04.2010 um 07:58 schrieb Troy A. Griffitts: >>> These all sound like good suggestions. Manfred, let me know what you decide and I'll either give you access to sword/bindings/objc or create a repository for you. Troy Peter von Kaehne wrote: > Would not the natural place be the bindings directory in the engine? > > > > Nic Carter wrote: >> I suggest you put it in a specific place in your MacSword SVN (or >> Bazaar if that's what you use now?) and I can access it from there. :) >> >> On 16/04/2010, at 2:32 AM, Manfred Bergmann wrote: >> >>> Alright then. >>> Can anyone with the proper rights create a Subversion folder? Or do >>> we use Bazaar? I switched over to Bazaar. :) >>> I believe I'm admin in Jira so I can create a project there myself. >>> >>> >>> Manfred >>> >>> >>> Am 15.04.2010 um 10:30 schrieb Nic Carter: >>> Hi Manfred, This all sounds good. :) Just one quick thing, though: PocketSword will be one app that works on both iPad and iPhone/iPod touch. I was looking at this today and I think this will be the best way forward. Apple have designed everything so that this is easily possible, with one codebase... :) Oh, and to answer your question about NSUserDefaults, yes, that's there! I have just added unlocking of modules to PocketSword, which will be available in v1.2.2 (currently in beta), and that works great using the code in SwordModule.mm :) Thanks for this work, Manfred! :) ybic nic... :) Nic Carter PocketSword Developer - an iPhone Bible Study app www:
Re: [sword-devel] buying xiphos? (Re: YOAH!!!!)
On 8 June 2010 23:33, Ari Constancio wrote: > On Tue, Jun 8, 2010 at 9:29 PM, Dmitrijs Ledkovs > wrote: >> On 8 June 2010 20:59, Karl Kleinpaste wrote: >>> [ not Cc'ing sword-support, from whence this came -- sword-devel only. ] >>> >>> ron81644t...@aol.com writes: >>>> Last month (May). I sent a money order to purchase the Xiphos program. >>>> I am using windows XP. This is the 3rd E-Mail to you. I love the program, >>>> but cannot istall any modules manually. PLEASE REPLY >>>> Ron Tanzola >>> >>> Where would he get the idea that Xiphos must be purchased, and who would >>> have redeemed such a purchase? >>> >> >> Xiphos is free and open source software, available for free. Please >> see http://www.fsf.org/ and http://www.gnu.org/philosophy/free-sw.html >> >> If you have send money to buy xiphos it was most likely a scam. We >> provide xiphos source code and binaries for Windows & Linux for free. >> >> Please provide as much details as you can so that we can initiate >> immediate request for take-down for violating our license terms (GPL) >> and FSF can support us with this via license violation. > > I could be wrong, but selling Xiphos doesn't count as a GPL violation. > > Ari Constancio > No it's not, as long as source code is provided by the seller as well. I am assuming the "seller" didn't do it =) ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] buying xiphos? (Re: YOAH!!!!)
On 8 June 2010 20:59, Karl Kleinpaste wrote: > [ not Cc'ing sword-support, from whence this came -- sword-devel only. ] > > ron81644t...@aol.com writes: >> Last month (May). I sent a money order to purchase the Xiphos program. >> I am using windows XP. This is the 3rd E-Mail to you. I love the program, >> but cannot istall any modules manually. PLEASE REPLY >> Ron Tanzola > > Where would he get the idea that Xiphos must be purchased, and who would > have redeemed such a purchase? > Xiphos is free and open source software, available for free. Please see http://www.fsf.org/ and http://www.gnu.org/philosophy/free-sw.html If you have send money to buy xiphos it was most likely a scam. We provide xiphos source code and binaries for Windows & Linux for free. Please provide as much details as you can so that we can initiate immediate request for take-down for violating our license terms (GPL) and FSF can support us with this via license violation. ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] finding conf file for module
On 23 May 2010 01:06, Matthew Talbert wrote:
> So, my first question is, "what is the best way to get the conf file
> for a given module?" and the second is, "could we please add this to
> the engine?"
>
The best way is probably to store sword configurations according to
XDG base-dir spec.[1]
It allows to store system-wide and per-user config files, user-data
(actual modules) & cache (search index).
If you look into your ~/.config ~/.local & ~/.cache you will notice
that a lot of programs already use it =)
On Windows gtk maps these locations to Application Data / All Users
for system & per user Application Data (or something like that, sorry
I don't use Windows)
In terms of API gtk, at least on *nix like systems, parses environment
XDG_* variables & has some defaults. The return from those functions
is semicolon separated list of paths which you then interate over to
get all the information you need.
Ofcourse if we go this route some decisions will have to be made as to
how to split it.
1) Everything is user data so we move ~/.sword -> ~/.local/sword and
coordinate with all apps to use XDG API's in their language/toolkit to
find system & user modules / configuration files
2) Split config, actual modules & search index into ~/.config/sword/,
~/.local/sword/ & ~/.cache/sword and do changes in sword where it
finds all the pieces. The idea here is that you can wipe ~/.cache and
everything will still work and cache can be regenerated, ~/.local
should store user created data (e.g. personal commentary and
downloaded modules) while ~/.config should store preferences only
(which can be wiped out to go back to default options). Read the spec
for more details.
I'm pretty sure there are ready made api's / libraries to deal with
this. Some duplication might be necessary in sword =)
e.g. I have this in my GtkBuilder branch already for xiphos for find
system data files:
1458gchar** dir_collection;
1459guint dir_collection_length;
1460guint dir_i;
1461gchar * full_fname;
1462
1463dir_collection = g_strdupv ((gchar **) g_get_system_data_dirs ());
1464dir_collection_length = g_strv_length (dir_collection);
1465
1466for (dir_i = 0; dir_i < dir_collection_length; dir_i++) {
1467full_fname = g_build_filename
(dir_collection[dir_i], PACKAGE, fname, NULL);
1468if (g_file_test (full_fname, G_FILE_TEST_EXISTS)) {
1469return full_fname;
1470}
1471
1472}
[1] http://standards.freedesktop.org/basedir-spec/basedir-spec-latest.html
___
sword-devel mailing list: sword-devel@crosswire.org
http://www.crosswire.org/mailman/listinfo/sword-devel
Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Visual Studio 2010 Professionell gratis and legal
ACM membership in USA gives you Developer Access to MSDNA and you can get VS2010 Ultimate and a lot of other goodies as well. Steps to complete 1) Sponsor a student for ACM membership =) On 18 May 2010 14:52, Wolfgang Schultz wrote: > Hello, > > MS offers a very simple way to get Visual Studio 2010 Professionell > and other developer tool at > > https://www.dreamspark.com/default.aspx > > Here my advice how to get the tools very easy. > > 1. Get an ISIC Card at: > > http://www.isic.org > > Any full-time student aged 12 years and over, at high school, college > or university is eligible for an ISIC card. As long as you're a > student, it doesn't matter whether you're 12 or 112. > > I'm not a student is there a card for me? > > Maybe you have children that could get the ISIC Card? :-) > > > 2. Register at Dreamspark and download what you need... > > I hope that helps. > > wolfgang > > http://www.zefania.de > > ___ > sword-devel mailing list: sword-devel@crosswire.org > http://www.crosswire.org/mailman/listinfo/sword-devel > Instructions to unsubscribe/change your settings at above page > ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] osis.py
On 11 May 2010 22:16, Weston Ruter wrote: > Thanks for the feedback, Dmitrijs. I'm actually taking my cues from jQuery > which is dual licensed MIT/GPL and merely says: >> >> /*! >> * jQuery JavaScript Library v1.4.2 >> * http://jquery.com/ >> * >> * Copyright 2010, John Resig >> * Dual licensed under the MIT or GPL Version 2 licenses. >> >> * http://jquery.org/license >> > http://ajax.googleapis.com/ajax/libs/jquery/1.4/jquery.min.js > > However, I just added the MIT license block in addition to the GPL and MIT > license URLs. Good that you thought of those without Internet access. > > More on why dual licensing is a good idea: > http://benalman.com/news/2010/02/on-licensing-my-code/ > Dual licensing is a bad idea cause it further spreads license profiliation =) and "they couldn't actually use it" is a false argument and there is comment about that in the blogpost already =) See the "source": http://www.gnu.org/licenses/license-list.html And better diagram from: http://www.gnu.org/licenses/quick-guide-gplv3.html About GPL compatability =) and MIT license gives you that (although a few different licenses are called that) You can take any MIT license code, change it combined it with GPL code and release as GPL. Infact MIT code can be combined with any code out there as long as attributions are given. On the other hand GPL can only be extended & re-released as GPL. So doing permissive & GPL dual license for the whole thing doesn't make sence =) Mozilla does triple-license because it is not actually licensing the whole thing under all three licenses. They still keep artwork and have levarage in telling a couple of distributions what they can / cannot include in their releases. This is a fun topic but more appropriate for debian-le...@l.d.o =) and I think I bore sword-devel enough with licensing issues already =) > Thanks! > Weston > Thank you for adding license information pkgcrosswire can include this software in Debian & Ubuntu when your software is stable enough ;-) > On Tue, May 11, 2010 at 1:45 PM, Dmitrijs Ledkovs > wrote: >> >> On 11 May 2010 19:58, Weston Ruter wrote: >> > Actually, it has to be open source because we don't have a paid GitHub >> > account :-) But you're right, I added a copyright block: >> > >> > http://github.com/openscriptures/api/commit/a731bdd68b84415f22d3a2d3a035e6ad827eaf28 >> > >> > Dual licensed MIT/GPL. >> > >> >> Kind of pointless =) MIT is compatible with GPL and MIT is less >> restrictive =) so *everyone* will choose MIT =) >> >> Plus you need to think about people without internet access who might >> receive your software. (or have internet access but live in a >> prosecuted country and get your software e.g. via printout in a >> letter). >> >> So actually it would have been absolutely amazing to have this header >> (taken from http://opensource.org/licenses/mit-license.php) >> >> >> 8<- >> >> Copyright (C) 2010 OpenScriptures.org >> >> Permission is hereby granted, free of charge, to any person obtaining a >> copy >> of this software and associated documentation files (the "Software"), to >> deal >> in the Software without restriction, including without limitation the >> rights >> to use, copy, modify, merge, publish, distribute, sublicense, and/or sell >> copies of the Software, and to permit persons to whom the Software is >> furnished to do so, subject to the following conditions: >> >> The above copyright notice and this permission notice shall be included in >> all copies or substantial portions of the Software. >> >> THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR >> IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, >> FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL >> THE >> AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER >> LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING >> FROM, >> OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN >> THE SOFTWARE. >> >> >> 8<-- >> >> > Thanks! >> > Weston >> > >> > >> > On Tue, May 11, 2010 at 11:23 AM, Dmitrijs Ledkovs >> > wrote: >> >> >> >> On 11 May 2010 17:32, Weston Ruter wrote: >> >> > I'm working on a Python module for representing osisIDs (works, >> >> > passages, >> >> >
Re: [sword-devel] osis.py
On 11 May 2010 19:58, Weston Ruter wrote: > Actually, it has to be open source because we don't have a paid GitHub > account :-) But you're right, I added a copyright block: > http://github.com/openscriptures/api/commit/a731bdd68b84415f22d3a2d3a035e6ad827eaf28 > > Dual licensed MIT/GPL. > Kind of pointless =) MIT is compatible with GPL and MIT is less restrictive =) so *everyone* will choose MIT =) Plus you need to think about people without internet access who might receive your software. (or have internet access but live in a prosecuted country and get your software e.g. via printout in a letter). So actually it would have been absolutely amazing to have this header (taken from http://opensource.org/licenses/mit-license.php) 8<- Copyright (C) 2010 OpenScriptures.org Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. 8<-- > Thanks! > Weston > > > On Tue, May 11, 2010 at 11:23 AM, Dmitrijs Ledkovs > wrote: >> >> On 11 May 2010 17:32, Weston Ruter wrote: >> > I'm working on a Python module for representing osisIDs (works, >> > passages, >> > IDs, and refs). I think I've got the OsisWork class baked pretty well >> > (at >> > least for reading/parsing), and I would appreciate your thoughts: >> > http://github.com/openscriptures/api/blob/master/osis.py#L122 >> > >> >> Currently I can only assume that it is proprietary software =) >> >> Please include copyright and license at the top of the file. >> >> Copyright (C) Name Lastname >> >> and boiler-plate for one of the OSI / FSF approved licenses. >> >> I recommend you to choose between Expat, 3-Clause BSD or LGPL / GPL >> licenses depending on how much freedom you would like to grant =) >> >> ___ >> sword-devel mailing list: sword-devel@crosswire.org >> http://www.crosswire.org/mailman/listinfo/sword-devel >> Instructions to unsubscribe/change your settings at above page > > > ___ > sword-devel mailing list: sword-devel@crosswire.org > http://www.crosswire.org/mailman/listinfo/sword-devel > Instructions to unsubscribe/change your settings at above page > ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] osis.py
On 11 May 2010 17:32, Weston Ruter wrote: > I'm working on a Python module for representing osisIDs (works, passages, > IDs, and refs). I think I've got the OsisWork class baked pretty well (at > least for reading/parsing), and I would appreciate your thoughts: > http://github.com/openscriptures/api/blob/master/osis.py#L122 > Currently I can only assume that it is proprietary software =) Please include copyright and license at the top of the file. Copyright (C) Name Lastname and boiler-plate for one of the OSI / FSF approved licenses. I recommend you to choose between Expat, 3-Clause BSD or LGPL / GPL licenses depending on how much freedom you would like to grant =) ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Xiphos : copy'n'paste to facebook produces Chinese !
On 6 May 2010 16:33, Gregory Hellings wrote: > > > > > On May 6, 2010, at 10:21 AM, Peter von Kaehne wrote: > >> On 06/05/10 13:42, David Haslam wrote: >>> >>> The quirky behaviour [..] >> >> Tried it On Firefox and Ubuntu without being able to reproduce. Don't >> have Windows right now. >> >>> This therefore points to a subtle bug in Xiphos, one that has not been >>> encountered before. >> >> Not by necessity. >> >> 1) Facebook is very buggy and known to be buggy. Facebook mangling posts >> would not surprise me. >> 2) It could be something with your specific clipboard. > > Additional software in the same stack: GTK1 as modified/used by FF, GTK used > by Xiphos, GTKHTML, Firefox, Gecko, Windows. Maybe others I'm missing? > GTK+2 =) FF on Windows doesn't use GTK. On linux distros FF-gtk flavour mostly uses system gtk. Gtkhtml here is not the culprit. When you copy you don't copy what you see, instead you get a copy object which has different represation depending on how the target application interprets the paste. This is very similar to drag and drop: when you drag a file and drop it into terminal you get full path to the file, when you drop it into GIMP it opens the file for editing if it knows how to edit that type of file etc So here the copy object is either generated incorrectly or interpreted incorrectly. It could be Xiphos (not generating copy object correctly), Windows (mangling the copy object), Firefox (mangling it again) or Facebook (interpreting the copy object wrong). On my combinations of this chain I cannot reproduce the problem. > Unicode is notoriously sticky for programs to handle because there is never > a guarantee other programs understand it or understand the same version and > encoding you're using. And many popular software tools simply aren't Unicode > aware in the first place or have limited support or require Unicode data to > be handled differently and with different APIs. > Even if the app source & app target understand you can force run the app in different encoding (or in case of a web-browser even per document). So far I haven't heard from the original reporter what are the encoding settings set to in windows regional settings, browser and xiphos. > --Greg > >> >>> There are many more thousands of users of facebook than of Xiphos. >>> It's unlikely that facebook would have not had a report filed & fixed if >>> this was their software bug. >> >> See above. >> >> So far no one has been able to reproduce it. I will try later on Windows >> if I get round to it. >> >> >> Peter >> >> ___ >> sword-devel mailing list: sword-devel@crosswire.org >> http://www.crosswire.org/mailman/listinfo/sword-devel >> Instructions to unsubscribe/change your settings at above page > > ___ > sword-devel mailing list: sword-devel@crosswire.org > http://www.crosswire.org/mailman/listinfo/sword-devel > Instructions to unsubscribe/change your settings at above page > ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Xiphos : copy'n'paste to facebook produces Chinese !
On 6 May 2010 11:43, David Haslam wrote: > > Has anyone else discovered this quirky behaviour? > > I copied a KJV verse from Xiphos direct to facebook, and it produced a whole > lot of Chinese characters! > > Workaround: Copy first to a text editor, such as Notepad++, then select > again for copy'n'paste. > > Xiphos team should explain what's going on! > > And perhaps consider providing a context menu option, "Copy as plain text". > > David Xiphos copy-n-paste is done right. Make sure you are running all of your applications in UTF-8 encoding and use normal web-browser. I'm assuming you are using windows: install google chrome, firefox, opera or safari and test with those browsers. ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Expanded keywords committed in the tree
>> propset (pset, ps): Set the value of a property on files, dirs, or revisions. >> usage: 1. propset PROPNAME PROPVAL PATH... >> 2. propset PROPNAME --revprop -r REV PROPVAL [TARGET] >> >> 1. Changes a versioned file or directory property in a working copy. >> 2. Changes an unversioned property on a repository revision. >> (TARGET only determines which repository to access.) > After re-reading this I the correct way is this: $ svn propset svn:keywords "Author Date Id Rev Url" -R . $ svn ci -m "Fixing svn:keywords property" Questions, comments, will you do this? =) ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Expanded keywords committed in the tree
Forget the patch. I found the problem. in the trunk checkout do: $ svn proplist -R -v | grep keywords -C 1 -- Properties on 'examples/classes/ciphercng.cpp': svn:keywords author date id -- Notice that generally you have only "author date id" enabled. $ svn proplist -R -v | grep keywords -C 1 | grep -i " rev" -C 2 Properties on 'utilities/osis2mod.cpp': svn:keywords Author Date Id Rev -- That's the whole output. The Rev keyword is enabled only for osis2mod.cpp But the actual Rev property is set on a few files: $ grep --exclude-dir=.svn \$Rev -r . ./utilities/mod2imp.cpp:fprintf(stderr, "\n=== mod2imp (Revision $Rev: 2234 $) SWORD module exporter.\n"); ./utilities/osis2mod.cpp: static const char* revision = ""; ./utilities/osis2mod.cpp: fprintf(stderr, "You are running osis2mod: $Rev: 2502 $\n"); ./utilities/tei2mod.cpp:fprintf(stderr, "You are running %s: $Rev: 2138 $\n", argv[0]); ./utilities/imp2vs.cpp: fprintf(stderr, "\n=== imp2vs (Revision $Rev: 2234 $) SWORD Bible/Commentary importer.\n"); ./utilities/imp2vs.cpp: "$$$Rev.0.0 for intro of the Book of Revelation of John. $$$[ Module Heading ]\n" ./configure.ac:AC_REVISION($Revision: 1.45 $) ./bindings/configure.ac:AC_REVISION($Revision$) ./bindings/swig/package/configure.ac:AC_REVISION($Revision: 1.2 $) Note that only osis2mod.cpp is correct substituion. The rest are bogus. $ grep \$Rev configure.ac AC_REVISION($Revision: 1.45 $) dmit...@surgut:~/src/sword/svn$ svn info configure.ac | grep Revision Revision: 2515 I believe that we got to this state due to ability to do propset in the working copy, or globally on the server: > propset (pset, ps): Set the value of a property on files, dirs, or revisions. > usage: 1. propset PROPNAME PROPVAL PATH... > 2. propset PROPNAME --revprop -r REV PROPVAL [TARGET] > > 1. Changes a versioned file or directory property in a working copy. > 2. Changes an unversioned property on a repository revision. > (TARGET only determines which repository to access.) So after doing this in my working copy: $ svn propset svn:keywords "Author Date Id Rev Url" -R . All keywords are correctly collapsed and would be committed as $Id$ , $Rev$ etc instead of bogus old expanded values. But to fix this on the server and for everyone who checkouts please do: propset svn:keywords --revprop -r ??? "Author Date Id Rev Url" Now if you do this for the repository root everything should be tip-top ;-) I'm not an expert in svn so i don't know whether you have to specify revision "-r NNN". I guess you should apply this property from "-r 1" for the whole repository but then it will mean that you are rewriting history cause libsword releases out there are reporting "You are running release $Rev: bogus$". Thanks ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
[sword-devel] Expanded keywords committed in the tree
Hello I'm using bzr when working on sword. There is continious import running on lp:libsword (sword name was already taken). Bzr supports svn-keywords via plugin. When I enable that function bzr finds that some keywords have been committed to the svn. Can someone please turn off keywords expansion in the svn client, commit this patch and turn keywords expansion back on? or something like that. Thanks. expanded-keywords.patch Description: Binary data ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Announcing a new SWORD front end called Xulsword
On 23 April 2010 04:01, John Austin wrote: > Xulsword (pronounced zool-sword) is now available for download and > collaboration at: > for my taste it's a bit too close to Xiphos. But maybe it's just me =) > http://code.google.com/p/xulsword/ > > Xulsword is a Bible reading and study tool which is easy and fun to > use, yet very powerful at the same time. The marriage of Xulrunner > with SWORD provides the best of both worlds by combining the power and > speed of C++ in the back-end SWORD engine with the quick and easy > programming of a Javascript front-end and CSS/Xul UI. It is designed > for quick internationalization, and potential users also include those > who are not experienced at using a computer, those who might be new to > the Bible, and those who may not have (or may not want to use) an > internet connection. > Xiphos is currently using xulrunner-embedding with Gtk+ toolkit. Xulsword using plain XULRunner for all UI rendering? Looks good =) > Xulsword may be particularly suited to places where the printed Bible > in a people's language is hard to find. > > Currently compiled for Windows, this XULRunner based program could > also be compiled without too much effort for Linux and Mac (any > takers?). > Haven't looked at it, yet. It might be "compiled without too much effort" but packaging properly does take quite a bit of effort of all XUL-based apps on unix like platforms. It's non-trivial. > New contributors are welcome, see issues list. > > Working together for His glory, > John > ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] JavaScript bindings for libsword
On 16 April 2010 14:46, Stephan Tetzel wrote: > Hello, > > I thought about making a bible app for webOS that's based on libsword. > Does anyone know if there are javascript bindings for libsword? I know > that there is the PDK (Plugin development kit), but it is still in beta. > > Best Regards, > Stephan > I'm not sure about javascript bindings. But that way you will still need to get libsword on webOS somehow, which will probably mean you will have to use PDK. If I were you I would look into jsword and run a jsword server somewhere and let webOS people access your server for content using your javascript driven app. Regars. ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] GSword source code?
Note that you cannot demand this though =) it is licensed under lgpl to have possibility of doing commercial software. And anyone can extend gpl portions of code privately without publishing code as long as didn't release it =) On 06/04/2010, David Haslam wrote: > > If http://www.crosswire.org/wiki/Frontends:GSword GSword is based on > JSword, please state where the GSword source code is maintained. > > -- David > -- > View this message in context: > http://n4.nabble.com/GSword-source-code-tp1753105p1753105.html > Sent from the SWORD Dev mailing list archive at Nabble.com. > > ___ > sword-devel mailing list: sword-devel@crosswire.org > http://www.crosswire.org/mailman/listinfo/sword-devel > Instructions to unsubscribe/change your settings at above page > -- Sent from my mobile device ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Improvements to osis2mod to handle XML comments and correctly
On 5 April 2010 13:55, Manfred Bergmann wrote: > Hi DM. > > Am 05.04.2010 um 13:21 schrieb DM Smith: > >> Regarding using a "real" parser, it is a good idea. But we don't want SWORD >> to be dependant on an external parser. > > What's the reason for that? > I could understand if it would mean for the user to install certain libraries > manually but when the sources can be integrated into the project and has the > appropriate licence then why not? > > > Manfred > IMHO there is no harm in bringing in libxml or a much more lightweight parser like GMarkup. The build system just needs to be adjusted to link e.g. libxml for the osis2mod binary and not shared sword library. in can be even called a new tool osisxml2mod for example and make it be build optionally such that you can still have full sword dev environment without libxml. Tools for creating modules do not have be linked with sword or even live in sword taball / svn. Although it does help consistent distribution of tools. ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Mac OSX Development Setup
2010/1/13 Troy A. Griffitts : > Trying to help a friend get their Mac setup to compile SWORD, I did some > hunting and discovered that searching our wiki and also a google search on > crosswire.org does not turn up any results. Could someone respond to this > thread with a quickstart for development from a vanilla Mac? It would be > most appreciated by me, and by anyone else searching our site for: Mac OSX > Development Setup :) > > Thanks! > > Troy > I have a jhbuild moduleset to build sword & xiphos. It's not perfect yet but works beautifully on mac. (it's based off gtk-osx modules). But if you just want to build sword you don't need to build any of the gtk stuff ;-) -- With best regards Dmitrijs Ledkovs (for short Dima), Ледков Дмитрий Юрьевич () ascii ribbon campaign - against html e-mail /\ www.asciiribbon.org - against proprietary attachments ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
[sword-devel] -Wall -Werror revisited
Hello all
Applying first patch. Do clean rebuild:
$ ./autogen.sh && ./configure --enable-debug && make clean && make
distclean && make -j5
Notice that the build fails. I believe those two lines in configure.ac
that touch user variables are somehow significant to FLAGS processing.
(No clue/wildguess, in debian package build we pass -Wall -Werror
flags multiple times which gets through this barrier somehow, making
our builds fail straight away)
Apply second patch, to pass even more -Wall -Werror (for the cc
compiles as well).
Apply third patch and notice that you can do a clean rebuild and not
have any warnings at all.
Please apply these patches to trunk.
ps third patch is the jmarsden's warnings patch updated against trunk
--
With best regards
Dmitrijs Ledkovs (for short Dima),
Ледков Дмитрий Юрьевич
() ascii ribbon campaign - against html e-mail
/\ www.asciiribbon.org - against proprietary attachments
=== modified file 'configure.ac'
--- configure.ac 2010-01-02 04:51:05 +
+++ configure.ac 2010-01-10 22:33:07 +
@@ -28,8 +28,6 @@
# -
# Check Programs
# -
-CFLAGS=" $CFLAGS"
-CXXFLAGS=" $CXXFLAGS"
AC_LANG(C++)
AC_PROG_CC
AC_PROG_CXX
=== modified file 'configure.ac'
--- configure.ac 2010-01-10 22:33:07 +
+++ configure.ac 2010-01-10 23:04:09 +
@@ -286,9 +286,8 @@
AC_SUBST(enable_debug)
AC_SUBST(enable_profile)
-# TODO: WHY ARE WE SUBST CXXFLAGS BUT NOT CFLAGS? And why, when I uncomment out the cflags subst do we get compile errors?
AC_SUBST(AM_CXXFLAGS)
-#AC_SUBST(AM_CFLAGS)
+AC_SUBST(AM_CFLAGS)
AC_SUBST(target_cpu)
AC_SUBST(target_vendor)
=== modified file 'src/mgr/filemgr.cpp'
--- src/mgr/filemgr.cpp 2009-12-24 22:22:16 +
+++ src/mgr/filemgr.cpp 2010-01-10 22:33:58 +
@@ -408,7 +408,7 @@
int FileMgr::copyFile(const char *sourceFile, const char *targetFile) {
- int sfd, dfd, len;
+ int sfd, dfd, len;
char buf[4096];
if ((sfd = ::open(sourceFile, O_RDONLY|O_BINARY, S_IREAD|S_IWRITE|S_IRGRP|S_IROTH)) < 1)
=== modified file 'src/utilfuns/zlib/untgz.c'
--- src/utilfuns/zlib/untgz.c 2001-03-23 09:00:15 +
+++ src/utilfuns/zlib/untgz.c 2010-01-10 22:33:58 +
@@ -13,6 +13,8 @@
#include
#ifdef unix
# include
+# include
+# include
#else
# include
# include
@@ -80,7 +82,7 @@
enum { TGZ_EXTRACT = 0, TGZ_LIST };
-static char *TGZfname OF((const char *));
+// static char *TGZfname OF((const char *));
void TGZnotfound OF((const char *));
int getoct OF((char *, int));
@@ -105,22 +107,23 @@
/* Return the real name of the TGZ archive */
/* or NULL if it does not exist. */
-static char *TGZfname OF((const char *fname))
-{
- static char buffer[1024];
- int origlen,i;
-
- strcpy(buffer,fname);
- origlen = strlen(buffer);
-
- for (i=0; TGZprefix[i]; i++)
-{
- strcpy(buffer+origlen,TGZprefix[i]);
- if (access(buffer,F_OK) == 0)
- return buffer;
-}
- return NULL;
-}
+// COMMENTED OUT 20090516 for SWORD
+// static char *TGZfname OF((const char *fname))
+// {
+// static char buffer[1024];
+// int origlen,i;
+//
+// strcpy(buffer,fname);
+// origlen = strlen(buffer);
+//
+// for (i=0; TGZprefix[i]; i++)
+// {
+//strcpy(buffer+origlen,TGZprefix[i]);
+//if (access(buffer,F_OK) == 0)
+// return buffer;
+// }
+// return NULL;
+// }
/* error message for the filename */
@@ -277,7 +280,7 @@
intremaining = 0;
FILE *outfile = NULL;
char fname[BLOCKSIZE];
- time_t tartime;
+ time_t tartime = 0;
while (1) {
len = gzread(in, &buffer, BLOCKSIZE);
=== modified file 'tests/testblocks.cpp'
--- tests/testblocks.cpp 2009-02-07 08:42:00 +
+++ tests/testblocks.cpp 2010-01-10 22:33:58 +
@@ -29,14 +29,18 @@
void addEntry(EntriesBlock *eb) {
string input;
string body;
- char line[1024];
+ char line[1024], *c;
std::cout << "\nEnter new Entry's text. '.' on an empty line to finish:\n";
do {
std::cout << "> ";
- fgets(line, 1000, stdin);
- input = line;
- if (input.compare("."))
- body.append(input);
+ c = fgets(line, 1000, stdin);
+ if (c == NULL) {
+ std::cerr << "ERROR: fgets failed during addEntry\n";
+ } else {
+ input = line;
+ if (input.compare("."))
+body.append(input);
+ }
}
while (input.compare("."));
std::cout << "Adding new entry. Index is: " << eb->addEntry(body.c_str()) << "\n\n";
@@ -72,29 +76,33 @@
EntriesBlock *eb = new EntriesBlock();
string input;
- char line[1024];
+ char line[1024], *c;
std::cout << "Initial entry count should be 0: " << eb->getCount() << "\
Re: [sword-devel] Twitter
2009/12/30 Matthew Talbert : >> Does anyone think this is just a stupid idea? With Twitter's character >> limit, most of our updates will be cut off to some degree, but at least it >> gets (some of) the information out there. > > Not sure if you're asking about Twitter in general or twitterfeed, but > anyway. I used to think twitter was pretty pointless, but I'm starting > to get what it's all about now. With Google now getting into the live > search thing, it definitely raises awareness on the web in general to > have tweets about a project. > > I'd suggest a shorter hash tag; #sword seems fairly unused. Perhaps we > can all use that for sword and frontend related tweets? > > There is also a twitter feed for Xiphos, > http://twitter.com/XiphosBible. I can't promise it will have anything > useful or even get updated, but we'll see :) > > Matthew > I love identi.ca (although twitter is more popular) and facebook and the whole micro-blogging stuff =) Although it seems to me a good idea to post updates on ohloh and relay that onto twitter ;-) Can you aggregate ohloh xiphos updates? [1] I post stuff there about Xiphos now. Cause I can do it from my Jabber account ;-) On related stuff can we have a Planet Sword? Similar to Ubuntu Planet[2], Debian Plane[3]t etc? Cause that would be really cool ;-) A few xiphos dev's use google reader but it doesn't join all developers & writers & users across our ecosystem. [1] https://www.ohloh.net/p/xiphos/messages.rss or https://www.ohloh.net/p/xiphos/messages (web) [2] http://planet.ubuntu.com/ [3] http://planet.debian.org/ -- With best regards Dmitrijs Ledkovs (for short Dima), Ледков Дмитрий Юрьевич () ascii ribbon campaign - against html e-mail /\ www.asciiribbon.org - against proprietary attachments ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] 1.6.1 final call
2009/12/25 Matthew Talbert : > OK, I am still not understanding why there is an issue, or what the > real cause of the issue is. However, this line I think will work: > > const unsigned int MAX_CONV_SIZE = 6536 * sizeof(wchar_t) * sizeof(wchar_t); > > If somebody can come up with an actual explanation for why there is a > problem, and a non-hackish solution, that would be great. > > Just for the record, wchar_t is 16 bits on win32 and 32 bits on *nix. > So, if I'm thinking correctly (and I won't guarantee that right now), > this should give the equivalent of 1024 * 1024; > > Matthew > BTW sword is compiled on 12+ architectures for Debian. Can you please provide the shell test suite. I'll include it in debian build and you will be able to gather all kind of information & failures from different architectures. I strongly belief saying on *nix is quite over generalisation. And i'm a bit skeptical as to why do we need "magic" numbers in core library. Just a thought =) -- With best regards Dmitrijs Ledkovs (for short Dima), Ледков Дмитрий Юрьевич () ascii ribbon campaign - against html e-mail /\ www.asciiribbon.org - against proprietary attachments ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] ESV clucene search oddities
2009/12/24 Dmitrijs Ledkovs : > 2009/12/24 Matthew Talbert : >>> >>> On Fedora11, I have 0.9.21 (not "b"), and Xiphos finds "crystal" in Rev >>> 9:21 with latest Sword SVN. I didn't even rebuild the index. >> >> So, it's looking like it's very likely a clucene issue which is now resolved. >> > > Would you want me to backport clucene from lucid (0.9.21b) to a > developer PPA for all supported ubuntu releases for testing and later > push it to our stable crosswire-ppa as well? > Never mind this, I'm talking rubish =) it will make all the other apps that link against clucene uninstallable on user's machine. -- With best regards Dmitrijs Ledkovs (for short Dima), Ледков Дмитрий Юрьевич () ascii ribbon campaign - against html e-mail /\ www.asciiribbon.org - against proprietary attachments ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] ESV clucene search oddities
2009/12/24 Matthew Talbert : >> >> On Fedora11, I have 0.9.21 (not "b"), and Xiphos finds "crystal" in Rev >> 9:21 with latest Sword SVN. I didn't even rebuild the index. > > So, it's looking like it's very likely a clucene issue which is now resolved. > Would you want me to backport clucene from lucid (0.9.21b) to a developer PPA for all supported ubuntu releases for testing and later push it to our stable crosswire-ppa as well? -- With best regards Dmitrijs Ledkovs (for short Dima), Ледков Дмитрий Юрьевич () ascii ribbon campaign - against html e-mail /\ www.asciiribbon.org - against proprietary attachments ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Updated regex.patch
2009/12/18 Troy A. Griffitts : > Dima (sorry, on my phone), in my previous email I posted the 2 lines of the > patch I didn't understand why you changed. > > I thought the line you removed added '-I ../include' to the g++ compile line, > which tell g++ where to look for include files. > > You replaced this line with an autotools directive 'include ...', or so I > thought. > > Does this clear up my question? > > Troy They are almost equivalent. '-l ../include' adds g++ compile line and it includes every single file from ../include. The include makefile add's does almost the same (adds ../include) *except for one file*: the ../include/regex.h is added conditionally if the configure steps decides that it is required. -- With best regards Dmitrijs Ledkovs (for short Dima), Ледков Дмитрий Юрьевич () ascii ribbon campaign - against html e-mail /\ www.asciiribbon.org - against proprietary attachments ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Updated regex.patch
2009/12/18 Troy A. Griffitts : > Thanks Dima, Your welcome ;-) > This sounds like a good problem to find. So, the configure 'include' > directive will put all of the files listed in the included file on the > -I compile include path? > ?? you mean the automake include statement which is similar to C #include - e.g. copy everything into this file from that file? There is no include directives in configure.ac nor generated configure > Would love for the help. We have not had a autotools maintainer for a > couple years now, and I just fumble my way around when I need to change > something. > I will work in a bzr branch hosted on launchpad such that I don't break trunk and everyone will still be able to test it ;-) -- With best regards Dmitrijs Ledkovs (for short Dima), Ледков Дмитрий Юрьевич () ascii ribbon campaign - against html e-mail /\ www.asciiribbon.org - against proprietary attachments ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Updated regex.patch
2009/12/17 Troy A. Griffitts : > Dima and Matthew, > > Thank you guys for working on this! Any final status on the situation? > > Dima, could you explain to me the changes you've made with regard to > changing: > > -INCLUDES = -I $(top_srcdir)/include > +include $(top_srcdir)/include/Makefile.am > > > I'm no configure expert and so this looks very different to me. > The first seems to add a -I sword/include to the include path during > compile. > The second seems to include the include/Makefile.am during autoconf > processing. > > Again, I'm ignorant to these matters, but just wanted us to check before > committing code that might not be related to the problem at hand and > might break something else right before a release. > > Thanks for bearing with my scrutiny. > > Troy > After I have patched the configure.in I have noticed that *some* files use system regex.h header and *some* files use the sword one. Which is inconsistent resulting in getting linked to both. If you look at $(top_srcdir)/include/Makefile.am you will notice that it is defining all the sword headers. By using this Makefile.am everywhere I could consistently get all other objects to either use the sword regex.h or not. Hope this makes sence. =) Anyhow there are a lot of really strange stuff in configure.ac & makefile.am's. I might work on it when I have time and send some patches but can't promise I will work on it any time soon. -- With best regards Dmitrijs Ledkovs (for short Dima), Ледков Дмитрий Юрьевич () ascii ribbon campaign - against html e-mail /\ www.asciiribbon.org - against proprietary attachments ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
[sword-devel] Updated regex.patch
This time around when searching for regex it will add additional -llibrary flags if required. For example on my linux it doesn't add any (prints none required) on windows it should add a flag. -- With best regards Dmitrijs Ledkovs (for short Dima), Ледков Дмитрий Юрьевич () ascii ribbon campaign - against html e-mail /\ www.asciiribbon.org - against proprietary attachments revno: 2436 committer: Dmitrijs Ledkovs branch nick: trunk timestamp: Thu 2009-12-17 01:53:52 + message: Adding option to build with/without sword regex diff: === modified file 'configure.ac' --- configure.ac 2009-12-10 03:20:14 + +++ configure.ac 2009-12-17 01:53:52 + @@ -19,8 +19,6 @@ case $host_os in mingw*) - # if we don't have system regex, use SWORD's copy - AC_CHECK_LIB(regex, regexec,,target_mingw32="yes") # both lines necessary to convince gcc to build dll LIBS="$LIBS -no-undefined"; export lt_cv_deplibs_check_method=pass_all; @@ -66,6 +64,8 @@ AC_HELP_STRING([--with-curl],[build install manager support using libcurl (default=yes)]),,with_curl=yes) #AC_ARG_WITH(lucene, # AC_HELP_STRING([--with-lucene],[include lucene support for searching (default=no)]),,with_lucene=no) +AC_ARG_WITH([regex], + AS_HELP_STRING([--with-regex], [Compile using internal copy of regex])) # - @@ -123,6 +123,17 @@ AM_CXXFLAGS="$AM_CXXFLAGS -DEXCLUDEZLIB" fi +AS_IF([test "x$with_regex" = "xyes"], + [have_systemregex=no], + [AC_SEARCH_LIBS(regexec, regex, [have_systemregex=yes], [have_systemregex=no])]) + +AS_IF([test "x$have_systemregex" = "xyes"], + [AC_MSG_NOTICE([Using system regex.h])], + [AS_IF([test "x$with_regex" = "xno"], + [AC_MSG_ERROR([System regex forced but not found])], + [AC_MSG_NOTICE([Using internal regex.h])])]) + +AM_CONDITIONAL(SYSTEM_REGEX, test "$have_systemregex") # - # Find CppUnit === modified file 'examples/cmdline/Makefile.am' --- examples/cmdline/Makefile.am 2009-02-17 02:31:04 + +++ examples/cmdline/Makefile.am 2009-12-17 01:53:52 + @@ -1,5 +1,5 @@ AUTOMAKE_OPTIONS = 1.6 -INCLUDES = -I $(top_srcdir)/include +include $(top_srcdir)/include/Makefile.am LDADD = $(top_builddir)/lib/libsword.la noinst_PROGRAMS = lookup search threaded_search listoptions verserangeparse outplain outrender === modified file 'include/Makefile.am' --- include/Makefile.am 2009-12-03 22:36:02 + +++ include/Makefile.am 2009-12-17 01:53:52 + @@ -75,7 +75,8 @@ pkginclude_HEADERS += $(swincludedir)/rawstr4.h pkginclude_HEADERS += $(swincludedir)/rawtext.h pkginclude_HEADERS += $(swincludedir)/rawverse.h -if MINGW +if SYSTEM_REGEX +else pkginclude_HEADERS += $(swincludedir)/regex.h endif pkginclude_HEADERS += $(swincludedir)/roman.h === modified file 'include/config.h' --- include/config.h 2009-12-07 04:02:43 + +++ include/config.h 2009-12-17 01:53:52 + @@ -59,6 +59,9 @@ /* Define to the one symbol short name of this package. */ #define PACKAGE_TARNAME "sword" +/* Define to the home page for this package. */ +#define PACKAGE_URL "" + /* Define to the version of this package. */ #define PACKAGE_VERSION "1.6.0" === modified file 'lib/Makefile.am' --- lib/Makefile.am 2009-12-03 22:36:02 + +++ lib/Makefile.am 2009-12-17 01:53:52 + @@ -1,6 +1,5 @@ AUTOMAKE_OPTIONS = 1.6 -AM_CPPFLAGS = -I$(top_srcdir)/include -I$(includedir) -AM_CPPFLAGS += -DUSE_AUTOTOOLS -DUNIX -Dunix -D__unix__ +AM_CPPFLAGS = -DUSE_AUTOTOOLS -DUNIX -Dunix -D__unix__ # Global config directory globalconfdir := @sysconfdir@ === modified file 'src/utilfuns/Makefile.am' --- src/utilfuns/Makefile.am 2009-12-03 22:36:02 + +++ src/utilfuns/Makefile.am 2009-12-17 01:53:52 + @@ -29,10 +29,10 @@ libsword_la_SOURCES += $(UNTGZ) libsword_la_SOURCES += $(ftpsrc) -if MINGW +if SYSTEM_REGEX +SWREGEX = +else SWREGEX = $(utilfunsdir)/regex.c -else -SWREGEX = endif libsword_la_SOURCES += $(SWREGEX) === modified file 'tests/Makefile.am' --- tests/Makefile.am 2009-12-03 22:36:02 + +++ tests/Makefile.am 2009-12-17 01:53:52 + @@ -1,5 +1,5 @@ AUTOMAKE_OPTIONS = 1.6 -INCLUDES = -I $(top_srcdir)/include +include $(top_srcdir)/include/Makefile.am LDADD = $(top_builddir)/lib/libsword.la SUBDIRS = cppunit === modified file 'tests/cppunit/Makefile.am' --- tests/cppunit/Makefile.am 2005-04-11 17:48:28 + +++ tests/cppunit/Makefile.am 2009-12-17 01:53:52 + @@ -8,9 +8,10 @@ endif check_PROGRAMS = $(TESTS) LibSword_SOURCES = main.cpp stringmgr_test.cpp swbuf_test.cpp url_test.cpp versekey_test.cpp -Lib
[sword-devel] Patch for regex building
This patch add new option --with-regex 1) No option given. Tries to use system regex, falls back to sword. 2) --with-regex. Uses sword regex. 3) --without-regex. Uses system regex error's out if it can't find a system regex. Please test is on windows ;-) -- With best regards Dmitrijs Ledkovs (for short Dima), Ледков Дмитрий Юрьевич () ascii ribbon campaign - against html e-mail /\ www.asciiribbon.org - against proprietary attachments Index: tests/cppunit/Makefile.am === --- tests/cppunit/Makefile.am (revision 1774) +++ tests/cppunit/Makefile.am (working copy) @@ -8,9 +8,10 @@ endif check_PROGRAMS = $(TESTS) LibSword_SOURCES = main.cpp stringmgr_test.cpp swbuf_test.cpp url_test.cpp versekey_test.cpp -LibSword_CXXFLAGS = $(CPPUNIT_CFLAGS) -I$(top_srcdir)/include +LibSword_CXXFLAGS = $(CPPUNIT_CFLAGS) LibSword_LDFLAGS = $(CPPUNIT_LIBS) swcppunitpdir = $(top_srcdir)/tests/cppunit +include $(top_srcdir)/include/Makefile.am #all: check Index: src/utilfuns/Makefile.am === --- src/utilfuns/Makefile.am (revision 2478) +++ src/utilfuns/Makefile.am (working copy) @@ -29,10 +29,10 @@ libsword_la_SOURCES += $(UNTGZ) libsword_la_SOURCES += $(ftpsrc) -if MINGW +if SYSTEM_REGEX +SWREGEX = +else SWREGEX = $(utilfunsdir)/regex.c -else -SWREGEX = endif libsword_la_SOURCES += $(SWREGEX) Index: utilities/Makefile.am === --- utilities/Makefile.am (revision 2478) +++ utilities/Makefile.am (working copy) @@ -1,6 +1,6 @@ AUTOMAKE_OPTIONS = 1.6 SUBDIRS = diatheke -INCLUDES = -I $(top_srcdir)/include +include $(top_srcdir)/include/Makefile.am LDADD = $(top_builddir)/lib/libsword.la noinst_PROGRAMS = cipherraw lexdump \ stepdump step2vpl gbfidx modwrite addvs emptyvss \ Index: configure.ac === --- configure.ac (revision 2488) +++ configure.ac (working copy) @@ -19,8 +19,6 @@ case $host_os in mingw*) - # if we don't have system regex, use SWORD's copy - AC_CHECK_LIB(regex, regexec,,target_mingw32="yes") # both lines necessary to convince gcc to build dll LIBS="$LIBS -no-undefined"; export lt_cv_deplibs_check_method=pass_all; @@ -66,6 +64,8 @@ AC_HELP_STRING([--with-curl],[build install manager support using libcurl (default=yes)]),,with_curl=yes) #AC_ARG_WITH(lucene, # AC_HELP_STRING([--with-lucene],[include lucene support for searching (default=no)]),,with_lucene=no) +AC_ARG_WITH([regex], + AS_HELP_STRING([--with-regex], [Compile using internal copy of regex])) # - @@ -123,6 +123,17 @@ AM_CXXFLAGS="$AM_CXXFLAGS -DEXCLUDEZLIB" fi +AS_IF([test "x$with_regex" = "xyes"], + [have_systemregex=no], + [AC_CHECK_HEADER([regex.h], [have_systemregex=yes], [have_systemregex=no])]) + +AS_IF([test "x$have_systemregex" = "xyes"], + [AC_MSG_NOTICE([Using system regex.h])], + [AS_IF([test "x$with_regex" = "xno"], + [AC_MSG_ERROR([System regex forced but not found])], + [AC_MSG_NOTICE([Using internal regex.h])])]) + +AM_CONDITIONAL(SYSTEM_REGEX, test "$have_systemregex") # - # Find CppUnit Index: examples/cmdline/Makefile.am === --- examples/cmdline/Makefile.am (revision 2263) +++ examples/cmdline/Makefile.am (working copy) @@ -1,5 +1,5 @@ AUTOMAKE_OPTIONS = 1.6 -INCLUDES = -I $(top_srcdir)/include +include $(top_srcdir)/include/Makefile.am LDADD = $(top_builddir)/lib/libsword.la noinst_PROGRAMS = lookup search threaded_search listoptions verserangeparse outplain outrender Index: tests/Makefile.am === --- tests/Makefile.am (revision 2478) +++ tests/Makefile.am (working copy) @@ -1,5 +1,5 @@ AUTOMAKE_OPTIONS = 1.6 -INCLUDES = -I $(top_srcdir)/include +include $(top_srcdir)/include/Makefile.am LDADD = $(top_builddir)/lib/libsword.la SUBDIRS = cppunit Index: lib/Makefile.am === --- lib/Makefile.am (revision 2478) +++ lib/Makefile.am (working copy) @@ -1,6 +1,5 @@ AUTOMAKE_OPTIONS = 1.6 -AM_CPPFLAGS = -I$(top_srcdir)/include -I$(includedir) -AM_CPPFLAGS += -DUSE_AUTOTOOLS -DUNIX -Dunix -D__unix__ +AM_CPPFLAGS = -DUSE_AUTOTOOLS -DUNIX -Dunix -D__unix__ # Global config directory globalconfdir := @sysconfdir@ Index: include/config.h === --- include/config.h (revision 2486) +++ include/config.h (working copy) @@ -59,6 +59,9 @@ /* Define to the one symbol sho
Re: [sword-devel] SWORD 1.6.1
2009/12/17 Jonathan Marsden : > Matthew Talbert wrote: > >>> I've not tested this except under Linux, and that only very briefly, but >>> doesn't adding a similar call to AC_CHECK_LIB outside >>> the case...esac solve this, by making sure the check is always >>> made (and so can always be tested for) on any host platform? I'm probably >>> missing some corner case where this doesn't do what we >>> need, but maybe it is a start? :) > >> Yes, but that would make it check twice, correct? > > Yes. If there's a better way, go for it. This way works for me, and is > easy to understand, so I did it (in my local copy only, at this point). A > SWORD source tree that won't build is no use to me, so I fixed it :) > > Are you aware of any negative side-effects of checking this twice? > > Jonathan > I'm almost done with nice fix including all corner cases and fixing a few Makefile.am's Please bare with me =) -- With best regards Dmitrijs Ledkovs (for short Dima), Ледков Дмитрий Юрьевич () ascii ribbon campaign - against html e-mail /\ www.asciiribbon.org - against proprietary attachments ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] SWORD 1.6.1
2009/12/16 Matthew Talbert : >> configure still failing on a mac? > > It fails on linux too. The cause is the patch I sent that Chris > applied. I nearly have it fixed (I think), but someone who actually > knows something about autotools could probably do a better job. > Shall I take a look at it? Has this already landed in sword svn? Or is there a patch/bzr branch I should look at? -- With best regards Dmitrijs Ledkovs (for short Dima), Ледков Дмитрий Юрьевич () ascii ribbon campaign - against html e-mail /\ www.asciiribbon.org - against proprietary attachments ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] New dev signing in
2009/12/16 Eeli Kaikkonen : > On Wed, 16 Dec 2009, Christopher Svanefalk wrote: >> Hello all, >> >> my name is Christopher, I started calling on CHRIST in my late >> teens, and began programming some time later. Currently I am still >> in college (computer science, University of Gothenburg). I would >> like to use my skills to serve GOD and bring good things to my >> fellow human beings, and I believe helping out with this project >> could be one such way. >> >> My main proficiency is in the C-family (C/C++/C#). I have done >> *some* work in other languages (mainly Java, Python) as well. >> >> I am rather inexperienced and have never participated in a major >> project, at least not in any significant way. However, I believe I >> have many gifts from GOD, and should be able to catch up on any >> necessary skills and produce good results. >> >> I hope I can be of use. Welcome =) I'm part of the Xiphos developer team which is mostly-C-gnome frontend to sword. If you are interested in Gnome more there are a few things we are going to do shortly after 3.1.2 release (due Christmas). We will clean up old dependencies and will try to do a few cool things. So quite a few tasks you can start working on. See http://xiphos.org/ and http://sourceforge.net/projects/gnomesword/ and http://sourceforge.net/apps/trac/gnomesword/ and http://launchpad.net/xiphos Also you can joins us on #xiphos on freenode. -- With best regards Dmitrijs Ledkovs (for short Dima), Ледков Дмитрий Юрьевич () ascii ribbon campaign - against html e-mail /\ www.asciiribbon.org - against proprietary attachments ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] FFe application for sword being considered, and patches split out
heya all I've just read motu release meeting minutes and from that they point out that a clear bug-fix release from upstream needs a bug in Launchpad and sponsor (not the big and scary ffe). So if clear bugs with clear fixes will be crafted into 1.6.0.1 tarball might work instead of carrying in more patches into karmic. Posting from the phone don't have link to minutes I'm referring to. I haven't looked into patches yet but one of my concerns is api/abi. I hope none of them brake that =) if they do than a ffe will be required. Cheers Dima. -- With best regards Dmitrijs Ledkovs (for short Dima), Ледков Дмитрий Юрьевич ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Systemwide clucene indexes
2009/7/16 Raphaël Pinson : > Hello, > > I have a proposal to improve the usability of BibleTime for end users. > Currently, when you install modules, whether locally or systemwide, > you need to build the indexes before peforming searches. In > Debian/Ubuntu, quite a few modules are packaged and shipped as .deb, > making them available to all users. I think it would be great if the > indexes could be compiled and installed systemwide aswell. From what I > see, this would require 2 things : > * that BibleTime search for clucene indexes in /usr/share/foo in > addition to ~ ; > * that there exists a CLI tool to generate the indexes from the modules. > > > With these two conditions, packagers could generate the indexes from > the modules as they package them, and ship them together for > systemwide installation, which would allow all users on the system to > benefit from the indexes without building them. > > > Feedback on the idea is welcome :-) > > > Raphaël Pinson > Sure. On the debian/ubuntu side we were getting around to look into packaging modules (it is possible they are out-of date) One request (which was heatedly discussed before) is to show users 1) which modules are system-wide and hence need to use Kpackagekit to remove/add 2) for user to be able to show / hide those modules Eg. If for example one user wants to study / view all of the modules while another user just wants to work with smaller amount of modules. From me additional request would be if Bibletime can hook into Kpackagekit to show which modules are available for system-wide install and optionally install them. Kpackagekit is distro-agnostic and will work in Debian, Ubuntu and Fedora if they will have modules packaged. more info at http://packagekit.org/ -- With best regards Dmitrijs Ledkovs (for short Dima), Ледков Дмитрий Юрьевич ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] tests/*.cpp need
2009/6/13 Karl Kleinpaste : > Newly-installed Fedora11, rebuilding my environment, includes building > tests, found that 4 files fail to compile because they lack #include > . > > Let me guess Gcc4.4? I remember some fedora developers saying "oh, it compiles fine with gcc4.4" = We had to do similar thing in Debian/Ubuntu. Me and Jonathan Marsden patched it and jonathan also produced a patch against trunk which included fixes for -Wall -Werror. Jonathan / SWORD developers was that patch applied in trunk??? -- With best regards Dmitrijs Ledkovs (for short Dima), Ледков Дмитрий Юрьевич ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] How broadly do we define "API" (was: Re: which engine sources to use )
2009/6/10 Greg Hellings : > > I know the question has been raised before about separating utilities > from the library but nothing has ever shaken out of it. To me, this > again makes sense in this category. If the utilities were placed into > their own SVN repository they could easily be released on their own > schedule with their own requirements. An svn:externals could force > the source to be included with an SVN checkout of the library, but > could allow the utilities to be conceptually "operated" as a set of > highly specialized front-ends (which is really what they are) for the > library and released on their own schedule. > > --Greg > Keep the same svn. With a little bit of auto-foo magic you can generate two different tarballs and release either of them at their respective schedules. IMHO this should be at least done for the bindings. Because python bindings autofoo assumes that the libsword is already installed on the system during build-time. This is very hard to satisfy on buildd / chroot. On the other hand if bindings were a separate tarball it could easily build-depend on libsword such that we (as is packagers) create libsword package first and then create bindings package. Maybe I'm wrong. In that case could you please suggest how to build python bindings when all you have is compiled sword in the current directory, or you have libsword installed into $DESTDIR eg. in debian case ./debian/libsword/usr/lib/ and other similar paths. -- With best regards Dmitrijs Ledkovs (for short Dima), Ледков Дмитрий Юрьевич ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] How broadly do we define "API" (was: Re: which engine sources to use )
2009/6/10 Jonathan Marsden : > > At the level of utilities that are used in scripts, and whose command > line parameters are therefore in some sense an "API" for those who write > scripts using them, the very recent (post 1.6.0) change to osis2mod that > adds the -d option *already* broke "osis2mod API compatibility" :) > Well *adding* function calls, flags etc does not break API. *Changing* behaviour of existings flags, function calls or removing them breaks api. My 2 UK pence -- With best regards Dmitrijs Ledkovs (for short Dima), Ледков Дмитрий Юрьевич ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] verseparsing-utf8
2009/5/24 Jonathan Marsden : > Dmitrijs Ledkovs wrote: > >> Verseparsing-utf8 is failing. > > Confirmed. I saw that too a while back... > >> ../src/mgr/filemgr.cpp: In member function ‘signed char >> sword::FileMgr::trunc(sword::FileDesc*)’: >> ../src/mgr/filemgr.cpp:276: warning: ignoring return value of ‘ssize_t >> write(int, const void*, size_t)’, declared with attribute >> warn_unused_result > > These are annoying to read, but relatively harmless (unless you do > -Werror of course :) . They do mean that the code concerned is not > really doing sufficient error checking, which should IMO be fixed; I > think I provided some ideas/a patch for at least some of these write() > calls earlier. > So if it fails to write (e.g. out of memory / HDD) then sword will still think everything succeeded fine? Seems quite dangerous to me I've had my system running with 0MB free space left on HDD for a while. It wasn't fun when programs assumed that certain things happened >> LocaleMgr::getLocale failed to find de > > These mean the SWORD library can't find the SWORD de locale data on your > system. I'm not sure it will be able to do so, unless you actually have > SWORD installed. In your example you did: > >> $ ./configure --with-icu --enable-tests > > so you'd need the SWORD locale data under /usr/local/share/sword/ (as I > understand it) for these tests to work as designed. > well I installed our debian package and yes _now_ the test works. > If these tests are supposed to work "out of the tarball" with no SWORD > library and locale data yet installed, then they probably need fixing so > that they are able to "find" the SWORD locale data inside the > tarball/build area?? > Sounds reasonable to me. That's how every other well-known test suites work they test what they have just build not was is installed and maybe is functional. And the previous cppunittest was working with the build data, not installed data. Again I'm not a C++ expert so I can only report the bug. On the bright side i now now how to run both cppunittest and the shell script based one at runtime. So dear sword developers if you want testsuite to be run on 12+ architectures by debian buildd machines please write one which works ;-) either by updating cppunittest so that it doesn't fail or move to a shell script one. -- With best regards Dmitrijs Ledkovs (for short Dima), Ледков Дмитрий Юрьевич ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
[sword-devel] verseparsing-utf8
Verseparsing-utf8 is failing. $ wget -q http://crosswire.org/ftpmirror/pub/sword/source/v1.6/sword-1.6.0.tar.gz $ sha256sum sword-1.6.0.tar.gz 8f1348c98015c656414f89ffa235afb3b1860cce2b242a81baeb16fde7678f4a sword-1.6.0.tar.gz $ tar xf sword-1.6.0.tar.gz $ cd sword-1.6.0/ $ ./configure --with-icu --enable-tests 1>/dev/null $ make 1>/dev/null ../src/mgr/filemgr.cpp: In member function ‘signed char sword::FileMgr::trunc(sword::FileDesc*)’: ../src/mgr/filemgr.cpp:276: warning: ignoring return value of ‘ssize_t write(int, const void*, size_t)’, declared with attribute warn_unused_result ../src/mgr/filemgr.cpp: In static member function ‘static int sword::FileMgr::copyFile(const char*, const char*)’: ../src/mgr/filemgr.cpp:409: warning: ignoring return value of ‘ssize_t write(int, const void*, size_t)’, declared with attribute warn_unused_result $ cd tests/testsuite/ && ./runtest.sh verseparsing-utf8 LocaleMgr::getLocale failed to find de LocaleMgr::getLocale failed to find de LocaleMgr::getLocale failed to find de LocaleMgr::getLocale failed to find de LocaleMgr::getLocale failed to find de LocaleMgr::getLocale failed to find de LocaleMgr::getLocale failed to find de Script failed at: (- bad output; + should have been) --- verseparsing-utf8.try 2009-05-24 16:53:53.0 +0100 +++ verseparsing-utf8.good 2008-07-22 16:08:04.0 +0100 @@ -1,7 +1,7 @@ -Matthäus 2:3-12 de: -Römer 2:13 de: -Matthäus 1:2-Röm 3:13 de: -1. Könige 2 de: -1. Könige - 2. Könige; Markus 1:1; Matthäus 2:1; Matthäus; 1.Kön de: -1. Könige - 2. Könige; Markus 1:1; Matthäus 2:1; Matthäus; 1.Kön-2.Kön;I Kings-Matthäus de: I Kings 1:1-Revelation of John 22:21 -Maleachi 1:1 - Matthäus 2:1 de: +Matthäus 2:3-12 de: Matthäus 2:3-Matthäus 2:12 +Römer 2:13 de: Römer 2:13 +Matthäus 1:2-Röm 3:13 de: Matthäus 1:2-Römer 3:13 +1. Könige 2 de: 1. Könige 2:1-1. Könige 2:46 +1. Könige - 2. Könige; Markus 1:1; Matthäus 2:1; Matthäus; 1.Kön de: 1. Könige 1:1-2. Könige 25:30; Markus 1:1; Matthäus 2:1; Matthäus 1:1-Matthäus 28:20; 1. Könige 1:1-1. Könige 22:53 +1. Könige - 2. Könige; Markus 1:1; Matthäus 2:1; Matthäus; 1.Kön-2.Kön;I Kings-Matthäus de: 1. Könige 1:1-2. Könige 25:30; Markus 1:1; Matthäus 2:1; Matthäus 1:1-Matthäus 28:20; 1. Könige 1:1-2. Könige 25:30; 1. Könige 1:1-Matthäus 28:20 +Maleachi 1:1 - Matthäus 2:1 de: Maleachi 1:1-Matthäus 2:1 -- With best regards Dmitrijs Ledkovs (for short Dima), Ледков Дмитрий Юрьевич ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] linker problems with 1.6.0? - help needed
2009/5/24 Martin Gruner : > Hi all, > > we use the following stupid program to check that linking BibleTime against > Sword is going to work: > > sword_linker_check.cpp:(.text+0xd7): undefined reference to > `sword::InstallMgr::InstallMgr(char const*, sword::StatusReporter*, > sword::SWBuf, sword::SWBuf)' > Dejavu I've been trying to change configure.ac to use pkgconfig to add compile flags of the libcurl and I was getting exactly the same error at buildtime of actual libsword on their "dummy" program buildtest.cpp -- With best regards Dmitrijs Ledkovs (for short Dima), Ледков Дмитрий Юрьевич ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
[sword-devel] About what gets linked.....
I did some digging and analysis about that link flags are added while building libsword. For libcurl currently `curl-config --libs` is used. On my machine (Ubuntu jaunty) this command emits following: $ curl-config --libs -lcurl -lidn -llber -lldap -lrt -lgssapi_krb5 -lgssapi_krb5 -lz -lgnutls Similarly with libicu - currently `icu-config --ldflags` is used which on my machienve emits: $ icu-config --ldflags -lm -L/usr/lib -licui18n -licuuc -licudata -lm I've used dpkg-shlibdeps which checks executables / libraries for symbols they are using and what they are linked against. And it does confirm that not all of these libraries are actually used dpkg-shlibdeps: warning: dependency on libgcrypt.so.11 could be avoided if "debian/tmp/usr/lib/libsword.so" were not uselessly linked against it (they use none of its symbols). dpkg-shlibdeps: warning: dependency on libtasn1.so.3 could be avoided if "debian/tmp/usr/lib/libsword.so" were not uselessly linked against it (they use none of its symbols). dpkg-shlibdeps: warning: dependency on libicuio.so.38 could be avoided if "debian/tmp/usr/lib/libsword.so" were not uselessly linked against it (they use none of its symbols). dpkg-shlibdeps: warning: dependency on libidn.so.11 could be avoided if "debian/tmp/usr/lib/libsword.so" were not uselessly linked against it (they use none of its symbols). dpkg-shlibdeps: warning: dependency on libldap_r-2.4.so.2 could be avoided if "debian/tmp/usr/lib/libsword.so" were not uselessly linked against it (they use none of its symbols). dpkg-shlibdeps: warning: dependency on liblber-2.4.so.2 could be avoided if "debian/tmp/usr/lib/libsword.so" were not uselessly linked against it (they use none of its symbols). dpkg-shlibdeps: warning: dependency on librt.so.1 could be avoided if "debian/tmp/usr/lib/libsword.so" were not uselessly linked against it (they use none of its symbols). dpkg-shlibdeps: warning: dependency on libicudata.so.38 could be avoided if "debian/tmp/usr/lib/libsword.so" were not uselessly linked against it (they use none of its symbols). dpkg-shlibdeps: warning: dependency on libgnutls.so.26 could be avoided if "debian/tmp/usr/lib/libsword.so" were not uselessly linked against it (they use none of its symbols). dpkg-shlibdeps: warning: dependency on libgssapi_krb5.so.2 could be avoided if "debian/tmp/usr/lib/libsword.so" were not uselessly linked against it (they use none of its symbols). There are a few solutions on how to reduce amount of libraries that are linked. 1) Curl. Instead of using curl-config to use PKG_CHECK_MODULES macro which is using information provided by the upstream in their libcurl.pc Cause according to libcurl.pc for a shared library build it will only add -lcurl flag. 2) Icu. The option I see is to filter the output such that libicuio and libicudata are not linked. (although some of the code suggest that in the past these were explicitly added) In addition to library all of these link flags are passed to all of the utilities. Although most of the utilities are using these headers. So I would in addtion suggest to reset / override link flags in the utilities/Makefile.am to make sure that they link in only what they need (e.g. zlib for a couple of them) I've tried to create a patch to build system to fix some of these but i didn't quite succeed. The only piece I did come up with is a work around to patch ltmain.sh such that it accepts --as-needed option. This results in ld thinking and not linking a library if it is not used by executable. Such patch can be applied to a tarball but I can't see a way to keep it constant even after "autogen.sh". And again --as-needed is just a workaround cause the real reason is that curl-config and icu-config are outputing too many libraries. I really would love to have these "extra" dependencies libraries to be dropped from debian packages. I can put together packages of sword with "--as-needed" flag for some more testing but I really hope that it could be sorted out the correct way. Some references: 3 - has example how to check for libraries / set flags using pkg-config autoconf macros 4 - is a discussion about --as-needed flag from gentoo [1] http://www.openismus.com/documents/linux/using_libraries/using_libraries.shtml [2] http://www.openismus.com/documents/linux/automake/automake.shtml [3] http://library.gnome.org/users/anjuta-build-tutorial/stable/library-autotools.html.en [4] http://www.gentoo.org/proj/en/qa/asneeded.xml -- With best regards Dmitrijs Ledkovs (for short Dima), Ледков Дмитрий Юрьевич ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] SWORD utils frontends
2009/5/16 Jonathan Marsden : > I'm wondering whether it might end up being simpler to create a Python > (or whatever your chosen portable scripting language is) GUI wrapper for > the SWORD utilities, that people run locally on their own machine, > rather than a web UI for a central server? Login authentication already > exists on the user's machine, there is no need to move possibly large > files back and forth, no worries about whose files they are, multiple > users at once are handled by the users OS (on multi-user OSes only, of > course), etc. > I guess we love python too much Cross-platform with correct widgets toolkit it will look native on all platforms. > Dmitrijs Ledkovs wrote: > >> OpenId is quite neat. ... > > It is, but (unless something changed in the last few months) it has one > drawback that is hindering its usefulness: for it to work well, you need > to trust that the security of each OpenID provider you accept is solid. > And in the security world, the default is that no-one trusts anyone else! > > So what is happening is that everyone is an OpenID provider, and > everyone accepts their own provider as being sufficiently secure... so > it looks like their site is trendy and accepts OpenID. However, many > places do not accept other providers. Or did not, when I tried it! > > At least last time I tried this, an OpenID created on LaunchPad was not > accepted by SourceForge, for example. Because of this, I just gave up > and continue to remember and use about 50+ passwords! Launchpad defaults to https: openID that indeed doesn't work on SourceForge http://launchpad.net/~you does work on SourceForge ;-) > > Hmmm, I just tried again, and after a couple of odd things happening > (like being asked to associate my OpenID with my SF account, doing so, > and getting sent to a blank web page!), it now seems to be working :) > So maybe things really are improving, or I just had back luck the first > time I tried out OpenID? > Well that's the strange thing that a lot of them do ask you to "associate" or "create" local account. But nonetheless openId is the other method to login. > Jonathan -- With best regards Dmitrijs Ledkovs (for short Dima), Ледков Дмитрий Юрьевич ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Use of -Werror (cause found)
2009/5/17 Jonathan Marsden : > Jonathan Marsden wrote: > >> I'll try a build of SWORD 1.6.0 with --enable-debug in a Jaunty 32bit >> (x86) VM here... hmmm, it works fine for me! However, uncommenting the >> --enable-debug in debian/rules gets me FTBFS, on 32bit and 64bit Jaunty. >> >> Dmitrij: This looks more and more like a packaging bug of some sort? > > (1) This is not a packaging bug, it is apparently a SWORD build system bug! > Thanks Jonathan for such a long email. The more random poking the more things can be found. =D -- With best regards Dmitrijs Ledkovs (for short Dima), Ледков Дмитрий Юрьевич ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Is --enable-debug what I want?
2009/5/17 Greg Hellings : >> You could patch that out in configure.ac (so it just sets -O0 and -g3) >> and then regenerate configure.in and configure, to test whether that >> solves the FTBFS issue. I'm pretty sure it would, but tests are better >> than guesses. >> >> Jonathan >> I went the crude way and I've patched ./configure and removed -Werror there (didn't want to regenerate all the configures) and It works. At this point I'm confused a little bit, I'll play around compiling directly without packaging and I'll see how far I go. -- With best regards Dmitrijs Ledkovs (for short Dima), Ледков Дмитрий Юрьевич ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
Re: [sword-devel] Is --enable-debug what I want?
2009/5/16 Dmitrijs Ledkovs : > 2009/5/16 Karl Kleinpaste : >> Dmitrijs Ledkovs writes: >>> As far as I remember all Debian Packages are compiled with -Wall >>> -Werror like always. >> >> FYI, Xiphos cannot be and never has been built with -Wall -Werror >> because standard xulrunner headers are not compliant. We know, we tried >> recently; only recent Fedora (9 & later) xulrunner has adequately modern >> headers. >> > > At this point I'm seriously doubting my skills & knowledge. > You are right Karl by default debian adds "-Wall" but not "-Werror" -- With best regards Dmitrijs Ledkovs (for short Dima), Ледков Дмитрий Юрьевич ___ sword-devel mailing list: sword-devel@crosswire.org http://www.crosswire.org/mailman/listinfo/sword-devel Instructions to unsubscribe/change your settings at above page
