[Fink-devel] Depends ignored during build phase
Both of these should fail during the build phase (notice "dlcompa" typo, giving a -dev or -shlibs package that does not exist). Package: foo Depends: dlcompat-shlibs BuildDepends: dlcompa-dev Package: bar Depends: dlcompa-shlibs BuildDepends: dlcompat-dev Using fink from CVS (HEAD, timestamp 2003/12/15 23:31:00), 'fink build foo' fails (can't resolve dependency "dlcompa-dev") as it should. But 'fink build bar' compiles cleanly. Fink eventually complains during 'fink install bar' (this warning is from fink, not dpkg). dan -- Daniel Macks [EMAIL PROTECTED] http://www.netspace.org/~dmacks --- This SF.net email is sponsored by: SF.net Giveback Program. Does SourceForge.net help you be more productive? Does it help you create better code? SHARE THE LOVE, and help us help YOU! Click Here: http://sourceforge.net/donate/ ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: [Fink-devel] Identically named packages
On Dec 15, 2003, at 6:24 AM, Peter O'Gorman wrote: Things should not be depending on python, they ought to depend on the specific version of python that they require, if the package is indeed So what then is the 'python' splitoff for? -Ben --- This SF.net email is sponsored by: SF.net Giveback Program. Does SourceForge.net help you be more productive? Does it help you create better code? SHARE THE LOVE, and help us help YOU! Click Here: http://sourceforge.net/donate/ ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: [Fink-devel] bin-dist made more regularely
Hi Dave, On Dec 12, 2003, at 2:16 PM, David R. Morrison wrote: Hi Remi. As Ben Reed pointed out, our biggest problem in setting up more frequent bindists is getting the system fully automated. In fact, as of now, only three people have ever made full bindists, and its only a partially automated process. To find out how it works, you can look in the "scripts" CVS module (although the instructions there are not complete). If somebody out there wanted to work on getting this process into better shape, that would be great. I spent a rainy Sunday looking through the scripts used to make a bin dist and started improving them. I'm not yet finished and there are sure many bugs left. However, I send here a part of the new HOWTO showing the assumptions I made and the steps needed to make a binary distribution. I'd like to hear your comments about it. The scripts are documented in more detail in a README which I do not include here due to space considerations. From the scripts/bindist/HOWTO: Q: What is needed before I can start making a binary distribution? A: - You need a working fink installation. However during the build phase all but the essential packages will be purged. Thus DON'T DO IT on a production system where you relay on certain fink packages. - The user building the binary distribution must be able to get root privilegies on the build machine using 'sudo' - The maintainer must have an account at SourceForge.net which enables her/him to upload the binary distribution. - Enough disk space to hold the local binary distribution (quantify???) - Before making a new binary distribution, the packages used must be tagged in cvs with the release number used. - If a binary distribution for the stable tree is built, only the stable tree must be enabled (at /sw/etc/fink.conf). Otherwise packages might get built against libraries existing only in the unstable tree. - Q: What scripts need to be run to make a binary distribution? A: # First edit bdenv.csh to set up the site specific settings # Then source the file to set up the environment source bdenv.csh # Create the directory structure for a new release, f.e. 7.0. This # assumes that the packages in cvs used for the binary distribution # are cvs tagged to 7.0 ./bdnewrel 7.0 # Get the sources from cvs and copy them into the dist structure # excluding those packages which may not be distributed as binaries ./bdsources # Next build the deb files for the sources copied in the previous # step. Successfully compiled packages are copied into the dist # structure, including the source files. ./bdbuild # After the previous step finished, check the $BDLOG directory for # any packages which failed to build (stored in NotBuilt). # Create index.php files in the various directories ./bdindex # Ensure all .deb's etc are uploaded ./bdsync # Make sure the files are referenced ./bdscan # Get the Package.gz files online ./bdsync --- If there are no mayor objections, I would like to start committing the changes to cvs. Shall I (can I?) branch the scripts/bindist or is it enough to tag the current status (any proposals for the tag name?). Is it wise to just tag/branch the bindist subdirectory or shall I tag/branch the full module scripts? Cheers, Remi - Failure is not an option. It comes bundled with your Microsoft product. (Ferenc Mantfeld) * Remigius K. Mommsen e-mail: [EMAIL PROTECTED] University of California, Irvine URL:http://cern.ch/mommsen c/o SLAC voice:++1 (650) 926-3595 2575 Sand Hill Road #35fax:++1 (650) 926-3882 Menlo Park, CA 94025, US home:++1 (650) 233-9041 * --- This SF.net email is sponsored by: SF.net Giveback Program. Does SourceForge.net help you be more productive? Does it help you create better code? SHARE THE LOVE, and help us help YOU! Click Here: http://sourceforge.net/donate/ ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: [Fink-devel] Bad news from the freetype2 front
David R. Morrison wrote:
Ah, I had another idea. We could create a freetype213 package which would
do all its installs in a non-standard location. You'd have to mess with
the link lines to make sure that scribus finds freetype213 and its headers,
rather than the one in /usr/X11R6, but that should be possible.
Yeah, this is how I did the new flex package. The new flex fixes lex
generation for some kinds of C++ code, but breaks other fink packages,
so I had it install in %p/lib/flex/{bin,include,lib} and then the
package(s) that depend on it need to explicitly add that to the search
paths...
--
Benjamin Reed a.k.a. Ranger Rick -- http://ranger.befunk.com/
gpg: 6401 D02A A35F 55E9 D7DD 71C5 52EF A366 D3F6 65FE
Snitchin' Is Bitchin'. -- Murphy
pgp0.pgp
Description: PGP signature
Re: [Fink-devel] Bad news from the freetype2 front
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 David R. Morrison wrote: | Martin, | | Have you tried statically linking freetype2 into scribus? | Well, I am tending to agree with Martin here, a real freetype2-dev package may be required. It could be called a different name, freetype2-real-dev and freetype2-dev could conflicts/replaces it and install nothing (as now). This would, of course, involve putting all those BuildDepends on freetype2-dev back in the tree though :/. Peter -BEGIN PGP SIGNATURE- Version: GnuPG v1.2.3 (Darwin) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iQCVAwUBP95XtbiDAg3OZTLPAQIJsAQAuHh7fBjp1npkLL+3RontG8eU9DGxoIyc oUN7fAuKnTU80Klwn8eptsj2GEGuaeDs1DWdFPj/RzSt0Q4i8U8BVsX8ufBSz1M4 FH2FE5iUO0VR3GyTrwV6QR1B1hG4cC70ybHp3UfrbwqFJLMqABlzc/Bc1Z4SUBhH OxhkznVJcPs= =W2I1 -END PGP SIGNATURE- --- This SF.net email is sponsored by: SF.net Giveback Program. Does SourceForge.net help you be more productive? Does it help you create better code? SHARE THE LOVE, and help us help YOU! Click Here: http://sourceforge.net/donate/ ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: [Fink-devel] Bad news from the freetype2 front
Ah, I had another idea. We could create a freetype213 package which would do all its installs in a non-standard location. You'd have to mess with the link lines to make sure that scribus finds freetype213 and its headers, rather than the one in /usr/X11R6, but that should be possible. -- Dave --- This SF.net email is sponsored by: SF.net Giveback Program. Does SourceForge.net help you be more productive? Does it help you create better code? SHARE THE LOVE, and help us help YOU! Click Here: http://sourceforge.net/donate/ ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: [Fink-devel] Bad news from the freetype2 front
Martin, Have you tried statically linking freetype2 into scribus? -- Dave --- This SF.net email is sponsored by: SF.net Giveback Program. Does SourceForge.net help you be more productive? Does it help you create better code? SHARE THE LOVE, and help us help YOU! Click Here: http://sourceforge.net/donate/ ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: [Fink-devel] Bad news from the freetype2 front
Martin Costabel wrote: I used to be all in favor of eliminating freetype2 from the 10.3 tree, because it seemed we finally had a decent freetype2 in X11. Unfortunately it now appears that the freetype2 in Apple's X11 just doesn't cut it. After some further investigations and discussions with the scribus developers, the situation now looks to me as follows: In Fink/Panther, we currently have 3 versions of freetype2: - version 2.1.0 in Apple's X11 - version 2.1.3 in the freetype2 package - version 2.1.4 in xfree86-4.3.99 (this version number can be found in freetype2/freetype/freetype.h) Unfortunately, there were substantial changes between 2.1.0 and 2.1.3, in particular concerning the loading of type1 fonts, so that the latest versions of scribus do not work with freetype2 versions earlier than 2.1.3. In such a situation, it would be normal for Fink to provide a complete freetype2 package whose *-dev splitoff would install the headers, the freetype.dylib symlink, and the freetype-config configuration script. We did have such a package and opted for retiring it in order to avoid incompatibilities with the X11-installed version of freetype2. In particular, we don't really want a /sw/bin/freetype-config that shadows the one in /usr/X11R6/bin. I am trying to think of a way how to have a freetype2-dev package that avoids these incompatibilities. The problem is, of course, that X11 does not behave like a standard package, in that its -dev part cannot be swapped in and out. What one would need is a kind of "BuildOnly:" field for freetype2-dev that means that this package is installed only while it is required for building another package, and it removes itself immediately afterwards. Or it could be removed by one of the xfree86 virtual packages. Can these virtual packages have "Conflicts:" and "Replaces:" fields that are activated from time to time? Too fancy, I guess. What I can easily do, once a freetype2-dev package exists, is to make scribus BuildDepend on it, but then have scribus' InstallScript remove it at the end. In this way, it wouldn't interfere with other packages that don't need and don't want it. Other ideas welcome. -- Martin --- This SF.net email is sponsored by: SF.net Giveback Program. Does SourceForge.net help you be more productive? Does it help you create better code? SHARE THE LOVE, and help us help YOU! Click Here: http://sourceforge.net/donate/ ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: [Fink-devel] Old style package name with epoch > 0
Hi Peter, On Dec 15, 2003, at 5:36 AM, Peter O'Gorman wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Remi Mommsen wrote: | | | Well, I don't really use the information from the package name, | otherwise I would have problems with the new convention anyway. I just | have a kind of lazy parsing which chokes when the package uses an epoch | in the old style version. But I'll try to improve that. | | I fail to understand why you are parsing .info files anywhere, the fink modules should be able to give you any package information you might need, use them. Indeed, and I try to use it as much as possible. But the Fink code is new to me, thus there are sure better ways to do many things. I'll appreciate any proposals for improvements once I have the first version of the code, which will take a few more days (or better evenings). The point where I still need to parse the info file directly is when I get from 'cvs update' the list of files which are newly added (or modified). Let's say I learn that foo-bla-1.3-1.info is added to the repository, then I'd like to 'fink build foo-bla-1.3.1', i.e. I just stripped the .info. This works not for packages with have an epoch defined. Anyhow, I use now Fink::Services::read_properties($pkg) to parse the file. This allows to specify the full version for 'fink build' for all packages (new and old naming scheme). Cheers, Remi - If it's green, it's biology. If it stinks, it's chemistry. If it has numbers, it's math. If it doesn't work, it's technology. (anonymous) * Remigius K. Mommsen e-mail: [EMAIL PROTECTED] University of California, Irvine URL:http://cern.ch/mommsen c/o SLAC voice:++1 (650) 926-3595 2575 Sand Hill Road #35fax:++1 (650) 926-3882 Menlo Park, CA 94025, US home:++1 (650) 233-9041 * --- This SF.net email is sponsored by: SF.net Giveback Program. Does SourceForge.net help you be more productive? Does it help you create better code? SHARE THE LOVE, and help us help YOU! Click Here: http://sourceforge.net/donate/ ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
[Fink-devel] openssl in 10.3/stable/main/finkinfo
Hi, I noticed that the info and patch file for openssl is found in 10.3/stable/main/finkinfo which are identical to 10.3/unstable/crypto/finkinfo, but differ from the 10.3/stable/crypto version. I guess this is a mistake that openssl for 10.3/stable is in the main branch and not in crypto. Cheers, Remi - Failure is not an option. It comes bundled with your Microsoft product. (Ferenc Mantfeld) * Remigius K. Mommsen e-mail: [EMAIL PROTECTED] University of California, Irvine URL:http://cern.ch/mommsen c/o SLAC voice:++1 (650) 926-3595 2575 Sand Hill Road #35fax:++1 (650) 926-3882 Menlo Park, CA 94025, US home:++1 (650) 233-9041 * --- This SF.net email is sponsored by: SF.net Giveback Program. Does SourceForge.net help you be more productive? Does it help you create better code? SHARE THE LOVE, and help us help YOU! Click Here: http://sourceforge.net/donate/ ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: [Fink-devel] The perl modules situation
On Dec 15, 2003, at 8:39 AM, Patrick Näf wrote: jfm said: Typically perl580 (eg) will install the same man pages, at the same place. So in order for the user not to have to force-overwrites , yes, you need such a Replaces (and perl580 too) _ if this is the case with your package. Ah, I see, the whole business is about the man pages - I didn't catch that earlier. In that case, make sure that your package is at least as up to date as the one in perl580 _ and that the corresponding man pages are substantially the same I'll do that then, thanks for the instruction. I'm afraid I've still been too sketchy, and it is not that easy _ because the perlxyz (say 580) packages need such a Replaces too, and it seems out of the question to have to update (including rev-up) this every time a maintainer of one of those +/- 20 packages checks his own _ this has to be centralized a bit one way or the other. Maybe you would be willing to do this, and then coordinate with the maintainer of perl580 ? An old list of pkgs to look at follows in PS (give and take one..). Basically, it involves installing perl580, then building each of those pkgs and installing them with dpkg --force-overwrites, so as to see the exact set of files that would get replaced, then checking for each of those files (I would expect mainly man files, but if I remember correctly there may also be sometimes scripts in /sw/bin, where obviously much more caution is required) whether it agrees substantially (ie, conveys exactly the same infomation) with the one of perl580. (Eg, by extracting also perl580 in a tmp dir, and doing, for each file, 'man full_path_to_man_file' to each of the 2 man files in 2 parallel windows). JF Mertens PS: attribute-handlers-pm cgi-pm digest-md5-pm file-spec-pm file-temp-pm filter-simple-pm filter-util-pm getopt-long-pm i18n-langtags-pm libnet-pm locale-maketext-pm memoize-pm mime-base64-pm scalar-list-utils-pm storable-pm template-pm test-harness-pm test-simple-pm text-tabs-wrap-pm time-hires-pm --- This SF.net email is sponsored by: SF.net Giveback Program. Does SourceForge.net help you be more productive? Does it help you create better code? SHARE THE LOVE, and help us help YOU! Click Here: http://sourceforge.net/donate/ ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: [Fink-devel] Identically named packages
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Ben Hines wrote: | On Dec 9, 2003, at 2:44 PM, Max Horn wrote: | |> |> Why was this done? Why not "netpbm10-bin", which then "Provides: |> netpbm-bin" ? | | | I agree, and was complaining about this the other day. Identically named | packages should be illegal. | | 'python' has a similar problem. If a user with python21 installs a | package which simply depends on 'python' (ie, the package doesn't care | what python version is installed), they will currently be forced to | install python23! Things should not be depending on python, they ought to depend on the specific version of python that they require, if the package is indeed python version neutral (i.e. it does not put anything in /sw/lib/python2.x and does not require a specific python version to run), then the package should depend on python21 | python22 | python23. The same rule should apply to pretty much all of the interpreted languages in fink. Peter -BEGIN PGP SIGNATURE- Version: GnuPG v1.2.3 (Darwin) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iQCVAwUBP93ELriDAg3OZTLPAQJkTAP/e14d7vIkZEcacJNJHERv19Hq9g2hEoxJ 3EEyZZng8kQWn+gUP8vyGNzO2mlnXrMCb1dDX5srAOo6dK4f51GitvWia1flaXpE lTu+BmpGxqTjdi8/dvYLsRaPZcgdOzb24OaQD8CC3tYQrGEPwcuRarZP2r03VmlB kGgiTIGUhEI= =762e -END PGP SIGNATURE- --- This SF.net email is sponsored by: SF.net Giveback Program. Does SourceForge.net help you be more productive? Does it help you create better code? SHARE THE LOVE, and help us help YOU! Click Here: http://sourceforge.net/donate/ ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: [Fink-devel] new fink commands
Opinions seems to me to go a bit too much in one single direction on this _ so let me try to play the devil's advocate.. On Dec 9, 2003, at 11:44 PM, Max Horn wrote: Am Dienstag, 09.12.03 um 21:56 Uhr schrieb TheSin: netpbm10 has 2 children: -> netpbm10-shlibs -> netpbm-bin so netpbm-bin has two parents Ugh! That's quite a nasty trick. In particular it means that if I do this: "fink install netpbm netpbm-bin" Fink will *not* do what I expect it to do... If you install pkgs you don't know, right. On the other hand, if you have netpbm-bin installed, and you do a blind "fink update-all", you'll stay up to date... Why was this done? Why not "netpbm10-bin", which then "Provides: netpbm-bin" ? I can't speak for the maintainers of all those packages, but one seemingly obvious motivation would be that this way, a user who has "selected" netpbm-bin and just wants to stay up to date, will stay so, automatically via "update-all". I.e., the fact that the functionality of update-all has basically been broken by the spitoffs venture is largely resolved by this 'trick', at least in as much concerns the users. (Of course it doesn't help to update the deps of other packages _ but that's an other issue _ at least the user's problem is largely solved by such a system.) The fact that a command what not originally conceived with a specific use in mind is no argument against such use: specs specify what the command should do, not what the user should, and as long as the use conforms with them, the better if it is creative _ it only shows addtional merit in the original design. So I could see an argument for not tightening the specs, even if minor adjustments in fink are required to fully support such use, because the alternatives to fully support an 'update-all' function _ from the user's point of view _ seem so much heavier, and hence further away. JF Mertens --- This SF.net email is sponsored by: SF.net Giveback Program. Does SourceForge.net help you be more productive? Does it help you create better code? SHARE THE LOVE, and help us help YOU! Click Here: http://sourceforge.net/donate/ ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: [Fink-devel] A few interesting facts about the Fink Webv-Site.
On 15 Dec 2003, at 7:24 PM, D. Höhn wrote: While this attempt at tracking was and is in no way complete it does show us where we could invest some more time. Most of our referrers seem to come from google , some are pouring in from apple.com now and a lot are coming from the guy who owns the fink.org domain. I suspect, like me there are a lot of people who can never remember the address. I always come via google, because I find it easier to type "fink" into Safari's google search than to try to remember where it is. -- Rohan Lloyd --- This SF.net email is sponsored by: SF.net Giveback Program. Does SourceForge.net help you be more productive? Does it help you create better code? SHARE THE LOVE, and help us help YOU! Click Here: http://sourceforge.net/donate/ ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
[Fink-devel] A few interesting facts about the Fink Webv-Site.
-BEGIN PGP SIGNED MESSAGE- Hash: RIPEMD160 Good morning Community. Over the last two weeks I run a little remote tracking tool on their Fink web-site. The tool has brought up some new, quite interesting results and it leaves room for educated guesses. I have been tracking only a few pages, the index page, the download area and the package database area and even though those pages aren't the whole web-site it shows that it is well frequented. The average daily unique User count lies around 3950 User. With about 2 page impressions per user that makes around 8000 page impressions a day. The tracking period wrapped over November and December showing that these values are pretty consistent. Some days the unique User count goes above 4000 yet most of the time it stays around 3900. With over 80% most users come from The USA, around 10% from japan, 5% from Germany and then it dwindles off, There have been recorded hits from the Arab Emirates and the Fidschies. Even the guys in Down-under and Oz seem to use Fink. While this attempt at tracking was and is in no way complete it does show us where we could invest some more time. Most of our referrers seem to come from google , some are pouring in from apple.com now and a lot are coming from the guy who owns the fink.org domain. The main batch of hits does come from direct hits, which is an indicator that Fink has spread to the common day discussion, it is being passed on as a tip by mouth propaganda and that is the most rewarding thing we could ever ask for. I hope that I can present some more detailed information soon. The indicators shown through this tracking attempt do have value, it shows that we should also invest more time into the Asian market. I shall be looking for experts in Japanese soon enough, so be prepared, I#ll ask all of you :) - -d -BEGIN PGP SIGNATURE- Version: GnuPG v1.2.3 (Darwin) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iD8DBQE/3W+3PMoaMn4kKR4RA4eIAJ9/RpYs/KcuzH47OrN+R/551DgkJACaA8pQ p5QwvDA5yRgH2O8Yfp2PC0M= =sb1m -END PGP SIGNATURE- --- This SF.net email is sponsored by: SF.net Giveback Program. Does SourceForge.net help you be more productive? Does it help you create better code? SHARE THE LOVE, and help us help YOU! Click Here: http://sourceforge.net/donate/ ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
