Re: editors/nvi-devel distfile recovered
Julian H. Stacey wrote: > Hi joh...@freebsd.org ( cc ports@) > as MAINTAINER= of editors/nvi-devel > BROKEN= Unfetchable > > I have made my local distfile temporarily available here: > http://berklix.com/~jhs/ftp/FreeBSD/ports/distfiles/nvi-1.81.6.tar.bz2 > Please copy it somewhere & remove BROKEN= , thanks. Thank you, I'll pick this up. Johan pgpiIZpOa2E1t.pgp Description: PGP signature
Re: Missing dependency readline in devel/libtool, wrong libreadline.so version
Hi Thomas, Thomas Mueller wrote: > Trying to update devel/libtool, I get /usr/local/bin/libtoolize with 0 bytes. > There is a message > Shared object "libreadline.so.8" not found, required by "gawk" [snip] > It looks like devel/readline, maybe also lang/gawk, should be listed > as a dependency for libtool. I think a better solution would be to make lang/gawk use the readline library from base and to make devel/libtool use awk from base. I'll look into this (soon). Regards, Johan pgpHEMpJokefY.pgp Description: PGP signature
Re: Some ports not compiling with security/libressl
Hi Jens, Jens K. Loewe wrote: > I'm currently trying to replace OpenSSL on my server by LibreSSL > (security/libressl). Anyway, I found some ports struggling to compile > with it, including (but not limited to) www/w3m, > security/p5-Net-SSLeay (which seems to query the OpenSSL version > directly?!) and lang/python27. > > Is it planned to take care of this use case? LibreSSL is explicitly not going to be fully compatible with OpenSSL. Certain functions from OpenSSL have already been removed from LibreSSL and as rewrites contine, the difference may grow. There would not be any point in adding a compatibility layer to the FreeBSD port. Application authors may chose to update their applications to work with LibreSSL, but I do not expect this to happen very quickly, especially because LibreSSL is a fast-moving target at the moment. That said, I'm not maintaining any of the ports you mentioned, and I did not look into the current compatibility issues. But in general, if you want full OpenSSL-compatibility, I suggest sticking with OpenSSL. Regards, Johan pgp2wSluykSHl.pgp Description: PGP signature
Re: gpg --refresh-keys fails with "rejected by import filter"
Hi Peter, Dr. Peter Voigt wrote: > So I played a bit and found that specifying more than one key fails, [..] > Is this related to GnuPG 2.0.24 or to the FreeBSD port? I can compare > just with a GnuPG 2.0.23 under Linux. This version works as expected. This is a confirmed bug in the GnuPG 2.0.24 release. There will probably be a bugfix-release soon. A patch candidate can be found at http://lists.gnupg.org/pipermail/gnupg-devel/2014-June/028522.html Regards, Johan pgpUfqIReUPNN.pgp Description: PGP signature
Re: ports/172837: lang/swi-pl compiles with databases/libiodbc
Johan van Selst wrote: > cpghost wrote: > > I've discovered the reason for this: > Ah, great. Now I can reproduce it as well. You are right, it is a hidden > dependency (with no sensible way of turning it off). I shall add your > patch with an explicit dependency on misc/ossp-uuid. On second thought... the potential for wrong dependency issues with conflicting ports, the lack of configuration options, the lack of shared libraries (static only) and the realisation that most people have been happy Prolog users without this feature so far, made me decide to rip it out instead. If anybody really wants to have ossp-uuid support in SWI Prolog, then please let me know. Cheers, Johan pgpDxIAVXC7gD.pgp Description: PGP signature
Re: ports/172837: lang/swi-pl compiles with databases/libiodbc
cpghost wrote: > I've discovered the reason for this: Ah, great. Now I can reproduce it as well. You are right, it is a hidden dependency (with no sensible way of turning it off). I shall add your patch with an explicit dependency on misc/ossp-uuid. Thanks, Johan pgpQsPTrLwwJZ.pgp Description: PGP signature
Re: ports/172837: lang/swi-pl compiles with databases/libiodbc
cpghost wrote: > Unfortunately, when OPTION iodbc is selected, the port now fails > to compile: Corrected now, thanks for reporting. I have also disabled the optional Java interface (JPL) for now. Will try and figure out how to get this working with varios JDKs and add it as an option later. Regards, Johan pgpucBLf6kCyA.pgp Description: PGP signature
Re: texinfo
Hi Baptiste, Baptiste Daroussin wrote: > I finally got it. The bug is in the texinfo port: PORTDATA= * > This globbing makes texinfo pick up all the files inside > /usr/local/share/texinfo but if texi2html is installed before texinfo > then it installs some files in that place, and those files are picked > up in the texinfo plist. Thus pkgng is right yelling about a conflict > in files. The fix would be to expand the files in PORTDATADIR directly > into the texinfo plist. Thank you for working this out: I completely missed that. I'll update the plist for texinfo now. Will look into texi2html later and see if it really makes sense to install its files in share/texinfo. Regards, Johan pgpnfIk4Babtu.pgp Description: PGP signature
Re: texinfo
Hi Mitja, ajtiM wrote: > In ports is texinfo-5.1 update but whenever is texinfo new version > (update) I have a problem: > print/texinfo-5.1 confilcts with texi2html-5.0 (installs files into > the same place. > Problematic: /usr/local/share/texinfo/init/book.init This is curious: texinfo does not install this book.init file; and neither port seems to mention a potential conflict in the port Makefile. Could this be a left-over warning from an old version? Are other people seeing this problem as well? Regards, Johan pgpSuONn46hfr.pgp Description: PGP signature
Re: FreeBSD Port: freeciv-nox11-2.3.2_1
Stephen Hurd wrote: > Re: Previous, seems to be because of this line: > INSTALLS_ICONS=yes > Also, the hicolor-icon-theme run depend should likely be only in X11 as well. You are right, well spotted. I'll update the port to correct this. Regards, Johan pgpZ6H4k2az3K.pgp Description: PGP signature
Re: texi2html problem
Johan van Selst wrote: > Doug Barton wrote: > > Building texi2html without NLS results in this: > > texi2html > > Undefined subroutine &Locale::Messages::dgettext called at > > /usr/local/bin/texi2html line 29628. > Indeed, I can confirm that the application is currently unusable when > compiled without NLS support. Will report this upstream and remove the > without-nls option until I/they can come up with a fix. I have included a fix that seems to do the trick. The port should now work when built without NLS support as well. Do let me know if it still causes trouble. I must admit that NLS support is included by default for all ports on my systems. Johan pgpp59L3hLKjJ.pgp Description: PGP signature
Re: texi2html problem
Doug Barton wrote: > >> It would also be useful if you could make real options out of > >> what you have in the file already. > > Not sure what you mean by this: the options seem real enough. > make config > ===> No options to configure The only relevant options are DOCS and NLS; they were never included in OPTIONS in the past (before I adopted this port). And indeed, many ports don't include DOCS and NLS in their OPTIONS today. My understanding is that most people who want to tweak these specific settings use a system-wide setting anyway. Is there some general policy on whether OPTIONS_DEFINE should be set for DOCS and NLS as well? Regards, Johan pgpRhst8OWbT8.pgp Description: PGP signature
Re: texi2html problem
Doug Barton wrote: > Building texi2html without NLS results in this: > texi2html > Undefined subroutine &Locale::Messages::dgettext called at > /usr/local/bin/texi2html line 29628. Indeed, I can confirm that the application is currently unusable when compiled without NLS support. Will report this upstream and remove the without-nls option until I/they can come up with a fix. > msgexec: warning: Locale charset "UTF-8" is different from > input file charset "ASCII". > Output of 'msgexec' might be incorrect. This should be harmless as UTF-8 is a superset of ASCII; but it is probably useful to enforce the ASCII character set during build. > It would also be useful if you could make real options out of what you > have in the file already. Not sure what you mean by this: the options seem real enough. Regards, Johan pgpXVZ47MVTTS.pgp Description: PGP signature
Re: Mk macros & print/texinfo/distinfo variant SHA256 SIZE
Hi Julian, Julian H. Stacey wrote: > A 9.0-RELEASE ports fails on > cd print/texinfo ; make fetch > unless one imports newer values from current, I copied the versions used by this port to my own server to prevent this failure, specificially for the case that the primary sources updates the files. The old files should still be retrievable from ftp.stack.nl if they are modified on the GNU servers. Maybe this secondary server was temporarily unavailable? That said, it is generally a good idea to keep your ports tree up-to-date, rather than to stick with a static snapshort ports tree from the release date. Regards, Johan pgpMFartKGOW8.pgp Description: PGP signature
Re: Ports with version numbers going backwards: chinese/tin
er...@freebsd.org wrote: > ** The following ports have a version number that sorts before a previous one > ** > Please fix any errors as soon as possible. > > - *chinese/tin* : zh-tin-2.0.1_1 < zh-tin-2.0.1_4 > (master: news/tin) >| revision 1.159 >| date: 2012/02/14 12:45:28; author: mm; state: Exp; lines: +2 -1 >| Bump pcre library dependency due to 8.30 update Should be fixed now. Ciao, Johan pgpIkthBJ4AkF.pgp Description: PGP signature
Re: Adding licensing info to my ports: some questions
Nikola Lečić wrote: > Anyway, it wasn't clear from the bsd.licenses.mk that we should use > 'multi' in situations of 'any later version'. This means that all > licensing info of eg. GPL2+ ports must be updated when GPL4 appears... No, we should not use this. Not just because of the potential of having to check and correct every port when GPLv4 appears. In my book, "licenced under GPLv2 or GPLv3" is something fundamentally different from "licenced under GPLv2 or any later version". The licence framework should be able to make this distinction. Another issue is that the licence infrastructure seems to be making statements about the licence of an application, while the committers only tend to look at individual source packages. What would be the licence of an application whose source is published under BSD licence, but that is linked with both GPv3 and OpenSSL-libraries? I tend to agree with Doug and others that it is probably better to scrap the entire idea. Making assertions about licences and what is accepted is a hairy field, best left to experts. Regards, Johan pgpqnD8hLmBzH.pgp Description: PGP signature
Re: [new port] usage of shar command
Anonymous wrote: > >> Am I the only one that thinks it's odd that in 2010 we're still using > >> executable scripts to distribute text files? > > How would you do it differently? > $ diff -upNr /nonexistent mynewport I quite like this solution. It's a simple command, very similar to what's used to create a diff for port updates; the resulting text works in email; it is certainly no harder to read than a shar archive; and it not being an executable script that demands close inspection before execution is definately a plus. Ciao, Johan pgp9ka71MVN9s.pgp Description: PGP signature
Re: [new port] usage of shar command
Joe wrote: > I can not figure out just what the author was trying to say with > "output of shar `find port_dir`" If your port files are stored in the subdirectory 'newport' then you may issue the command shar `find newport` > file.shar This creates the 'file.shar' file, which contains a shell archive of all the files in the 'newport' subdirectory. You should then include this file in your new port submission. Regards, Johan pgplq8T4JsyxB.pgp Description: PGP signature
Re: bsd.licenses.mk: where is "or any later version" construct?
Alberto Villa wrote: > On Thursday 10 June 2010 17:39:52 Anonymous wrote: > > IANAL, but I think LGPL3 is applicable here, too. So, I've tried to set > > LICENSE= LGPL21+ > > but it doesn't work. I've figured this will work > > LICENSE= LGPL21 LGPL3 > > LICENSE_COMB= dual > > Is this correct usage or I'm missing smth? > that's what you're supposed to do, as far as i understand This doesn't seem right: "LGPL21 or any later version" is very different from "LGPL21 or LGPL3". Also, how should one describe the difference between a "GPL3" and a "GPL3 or any later version" licence? There might not be a newer version right now, but there will be in the future and it would be rather annoying if we'd have to check all the software licences again when this version is released. We'd then have to change the licence registration for each such port, even though neither the software nor its licence text has changed - but only because of limitations of our framework. Since this is a very common practice with GPL and LGPL licences, imho it seems sensible to make this distinction right from the start and use different keywords for software with/without the "or any later version" clause. Regards, Johan pgpy9I2ra2YbF.pgp Description: PGP signature
Re: Need Help From Sparc64 User
James Bailie wrote: > I'm the maintainer of lang/munger, and it has been marked broken on > Sparc64 because it won't link. I need to see the compiler output > to fix the problem, but I don't have access to this architecture. There might be some useful logs on pointyhat in this case. Check out http://portsmon.freebsd.org/portoverview.py?category=lang&portname=munger Ciao, Johan pgpeq8MeRAXJN.pgp Description: PGP signature
Re: FreeBSD Port: smalltalk-3.0.5
per...@pluto.rain.com wrote: > Given the backwards compatibility issues, might it be worthwhile > to create a new smalltalk31 or smalltalk-devel port, or rename > the current port to something like smalltalk30x, so that the > older version can remain available for use with older programs? Well, 3.1 is not really a development branch: it is clearly intended to be the next stable branch. The first release might not have been perfect, but the author has made it clear that 3.1 is 'the way forward' and that the 3.0 branch will be discontinued. I will happily make a separate branch if there is even one smalltalk user who wants to keep 3.0.x around (if you read this: please respond). But otherwise I'm inclined to just drop 3.0 in the near future and let people convert their programs. For those interested, check out the 'Incompatible changes' section on http://smalltalk.gnu.org/news/gnu-smalltalk-3-1 Regards, Johan pgp25jcxuIOOp.pgp Description: PGP signature
Re: FreeBSD Port: smalltalk-3.0.5
Greetings, Herby Vojčík wrote: > I hope I am not very annoying, but I'd like to ask when will the port of > 3.1 be ready. It features star packages with new clauses, which 3.0.5 > cannot read, and better compstibility with other Smalltalk is also nice a > feature. I was pretty disappointed that I cannot install Swazoo 2.2 since > it\s packaged in 3.1 format. Let me add a little bit of background here: versions 3.0.5 and 3.1 were both released on the very same day: 3.0.5 is not older than 3.1. The former was a stable bugfix release and 3.1 came with serveral new features (and unfortunately no perfect backwards compatibility for running older programs). There were some minor issues (bugs) in 3.1 as well, although I don't remember the exact impact. At the time, there were no programs that used the new features and switching to it might cause problems for others. Right now, many programs will work with either version, and some even require the new 3.1 features (as you pointed out). I had expected that there would be a bugfix release for 3.1 by now with, so that I could upgrade to that version. But maybe the bugs aren't as important as I thought. I'll look in to these issues and will probably switch the port to 3.1(.x) soon. Regards, Johan pgpQvyMrCHe4B.pgp Description: PGP signature
Re: www/firefox35 coredumps every time at startup
Heino Tiedemann wrote: > I installed the port www/firefox35. After installation I tried to > start it > It core dumps 'every time' (100% reproducable). Did you load sem(4)? To quote UPDATING: If your Firefox crashes with the following message while viewing a HTML5 page: "Bad system call (core dumped)" you need to load the sem module (kldload sem). To load sem on every boot put the following into your /boot/loader.conf: sem_load="YES" Ciao, Johan pgpbyKSUhe8xV.pgp Description: PGP signature
Re: Pike (lang/pike78) on FreeBSD 7.2
Hi Arjan, Arjan van Leeuwen wrote: > Since I upgraded to FreeBSD 7.2 (from 7.1), my Pike installation has been > hanging regularly. When this happens, there is always one pike process in > [wait] state and one in [umtxn] (from top). I use Pike frequently on 7.2 systems, but haven't seen this problem yet. You should probably ask this on the Pike mailinglist, where the developers may come up with suggestions to further investigate this, http://pike.ida.liu.se/forums/ A new Pike 7.8 release is due "real soon now" by the way, and I will update the FreeBSD port once it's released (and it works for me). For now you might try to install the release candidate from http://pike.ida.liu.se/pub/pike/beta/7.8.316/Pike-v7.8.316.tar.gz to see if that one still has the same issues. Regards, Johan pgpG0yh9cfjkq.pgp Description: PGP signature
Re: vim ports broken.
Carlos A. M. dos Santos wrote: > >> Choosing "%" to differentiate it, however, was a bad idea. It would > >> be better to use a simple underscore. > > Why is % any worse than an underscore? As I explained earlier, fetch has no > > problem dealing with it. The errors people were posting are server-side, > > ftp.freebsd.org has no problems handling it. > POLA: users get surprised by the "Bad Request" answer and think that > the package is broken. So they complain at -ports, generating lengthy > email threads. I have another suggestion: why not tell it that this specific patch file should be fetched from the FreeBSD server and not try all the other vim mirror sites. We have a mechanism to do so, which seems appropriate here. Johan --- Makefile.orig 2009-06-25 10:24:08.0 +0200 +++ Makefile2009-06-25 10:21:52.0 +0200 @@ -16,7 +16,7 @@ DISTFILES= ${RELEASE}${EXTRACT_SUFX} PATCH_SITES= ${MASTER_SITES:S|unix|patches/${PORTVERSION:C/\.[0-9a-z]*$//}|}\ - ftp://ftp.FreeBSD.org/pub/FreeBSD/ports/local-distfiles/obrien/ + ftp://ftp.FreeBSD.org/pub/FreeBSD/ports/local-distfiles/obrien/:local PATCHFILES!= /usr/bin/jot -s " " -w ${PORTVERSION:C/\.[0-9]*$//}.%03d \ ${PATCHLEVEL} 1 ${PATCHLEVEL} # bits to remove @@ -27,7 +27,7 @@ .for p in ${BADPATCHES} PATCHFILES:= ${PATCHFILES:N7.2.${p}} .endfor -PATCHFILES:= ${PATCHFILES:S/041/041%/} +PATCHFILES:= ${PATCHFILES:S/041/041%:local/} MAINTAINER?= obr...@freebsd.org COMMENT?= Vi "workalike", with many additional features pgpFFuM2zMmhp.pgp Description: PGP signature
Re: pthreads => no Bacula encryption on FreeBSD Release 7
Dan Langille wrote: > "This is to warn you that Bacula will probably not be able to be > compiled and run with encryption on Release 7 of FreeBSD. This is > because the version of pthreads in that release has pthread_t defined as > a structure, which is incompatible with OpenSSL." The proper solution here would probably be for bacula to use the newer CRYPTO_THREADID features of OpenSSL, which do not have this restriction. The API is described at http://www.openssl.org/docs/crypto/threads.html Unfortunately these are only available with a recent OpenSSL source - and not with the version that is included in the FreeBSD 7 base system. Ciao, Johan pgpbcfGLXPGYT.pgp Description: PGP signature
Re: [RFC] New category proposal, i18n
Doug Barton wrote: > > Should this new category come to being, the self identified ports in misc > > would get relocated. All other ports would simply be extended with the new > > virtual category name. > You've probably already covered this, but are you making a distinction > between ports that are used to _do_ localization-related tasks, and > ports that are localized versions of existing ports? I think that this is an important distinction; and personally I would expect only the previous ports to go into this category. That is, putting ports into categories primarily based on the functionality of a port. For example when looking for a port like firefox-i18n, I'd expect to find it in the www/ category, rather than an i18n/ subdir. But ports like gettext or other localization tools could be in the i18n category. Ciao, Johan pgpX6HsiUbOha.pgp Description: PGP signature
Re: shells/bash-4.0 port horribly broken
Jerry wrote: > Was this some sort of 'improvement' by the Bash developers, or is it > an un-squashed bug? It seems that this might actually be a feature. Quoting the COMPAT document of bash4: 38. Since bash-4.0 now follows Posix rules for finding the closing delimiter of a $() command substitution, it will not behave as previous versions did, but will catch more syntax and parsing errors before spawning a subshell to evaluate the command substitution. Ciao, Johan pgpowQxOjpVNe.pgp Description: PGP signature
Re: Using sed in a Makefile
Jerry wrote: > @${REINPLACE_CMD} -e > 's|CONFIG_FILE=${CONFIG_FILE:-/etc/${SCRIPT_NAME}.conf}|CONFIG_FILE=${CONFIG_FILE:-/usr/local/etc/${SCRIPT_NAME}.conf}|g' > ${WRKSRC}/${PORTNAME}.sh > > Syntax error: Unterminated quoted string ${..} will be interpreted by make - even when it is used between '..' which is probably not what you want here. If you want to use a litteral $-sign in your command, use $$ in the Makefile. Ciao, Johan pgpVOiSbnSoTj.pgp Description: PGP signature
Re: Call for potential ports maintainers
Thomas Abthorpe wrote: > The gauntlet has been thrown down, who among you is prepared to pick it up? Please, when volunteering to maintain up a port, also check the open error reports and problem reports for this port - as listed on Portsmon, http://portsmon.freebsd.org/ and other locations. Thanks, Johan pgpCUOjAH1PNB.pgp Description: PGP signature
Re: FreeBSD Port: sks-1.1.0
Hi Joe, Joseph Oreste Bruni wrote: > I see that you are the maintainer of the SKS package within the FreeBSD > ports tree. I was wondering if you could add one of the patches that is > present in the SKS revision control repository but has not yet made its > way into a release. Thanks. I'll add the patch and update the webpage. Regards, Johan pgpYR5k6Ivexd.pgp Description: PGP signature
Re: FreeBSD Port: inn-2.4.5
Howard Goldstein wrote: > Michael Grimm wrote: >> The recent Perl-upgrade to 5.8.9 crashes innd (core dump) at start-up, >> reproducible. I did re-compile all ports including inn, same result. >> Then I removed Perl support during configuration, re-compiled, and now >> innd starts and runs as expected. > For what it's worth, same situation and resolution here. This has also been discussed on the INN mailinglist. The changes in perl 5.8.9 are not compatible with 5.8.8 and current versions of INN simple don't work with perl 5.8.8 yet (although perl 5.10 is reported to work). For now the only workaround appears to either stick with perl 5.8.8 or disable perl support altogether in INN. Ciao, Johan pgpr5rSMZcT9r.pgp Description: PGP signature
Re: Why is security/pinentry not a dependency of security/gnupg
Tobias Rehbein wrote: > Or is there some way to use gpg without pinentry? You could use security/gnupg1 instead (which is still developed, just a seperate branch), which doens't require pinentry. Ciao, Johan pgp2IwUtXegLB.pgp Description: PGP signature
Re: Ion3 license violation
Stefan Sperling wrote: > An alternative is to simply keep the last released version > that had a sane license. AFAIK OpenBSD did that, see: > http://marc.info/?l=openbsd-ports&m=119522869306969&w=2 Sounds like a fair solution. Ciao, Johan pgpKKzw6X0dHN.pgp Description: PGP signature
Re: duration of the ports freeze
Aryeh M. Friedman wrote: > >> What is the current ETA on when the ports freeze will end? > > See the freebsd.org release pages > I know you will kill me for this but the URL? http://www.freebsd.org/releases/6.3R/schedule.html http://www.freebsd.org/releases/7.0R/schedule.html "Ports tree unfrozen7 Nov 2007" Although I guess this date is somewhat... ehm... ambitious. This should probably be 7 December. Johan pgptcMdLlGhu1.pgp Description: PGP signature
Re: security/gnupg (or one dependance) needs xorg**** now?
Gaye Abdoulaye Walsimou wrote: > security/gnupg (or one dependance) need xorg now? gnupg2 requires with a standard interface to read passphrases and PIN codes (in combination with smartcards) in a secure way. For this security/pinentry is needed. It comes in several flavours: curses, gtk, gtk2 or qt. I believe the default for the pinentry port is to build all variations. gnupg2 is not fully functional without pinentry installed, but you can go for the curses-only interface by setting certain make variables. Ofcourse, you could also stay with gnupg1 (which is still developed in parallel) and doesn't require additional libraries and external programs. Gnupg1 doesn't include some more exotic new features, but it's functional. Greetings, Johan pgpCw0ILeeLio.pgp Description: PGP signature
Re: problem with send-pr
Vizeli Pascal wrote: > I've problem with send-pr. > Send-pr doesn't send the message and I don't know why. This might be a mail configuration problem (on either end). You can try the webinterface, http://www.freebsd.org/send-pr.html > In the attachment is a new port (gpac mp4box). I don't see any attachment - that probably got removed by the mailinglist configuration. Greetings, Johan pgp0eMUBuEX0z.pgp Description: PGP signature
Re: FreeBSD Subversion Network access problem
Siju George wrote: > I installed Subversion on FreeBSD 6.1 from ports. > Started svnserve to listen on 3690 > How do I make it listen on tcp4. You can add the option --listen-host=0 to listen on IPv4 only. Greetings, Johan pgpaeaEpCbpGo.pgp Description: PGP signature