[Fink-devel] Newbie: Package maintenance and bringing GNOME up to speed
Hi all, I've been a long time fink user, but only recently have my skills come up to a level where I would feel comfortable offering help. I have spent some time over the last few days bringing some of the core dependencies for GNOME 2.2.1 up to their relevant releases on my machine, and given the success and minimal patching required, I would like to push some of these forward. I have a couple of questions: 1) Is there anything not covered in the FAQs/Porting/Packaging/SourceForge docs that I need to know/learn before I can help? 2) What are the current arrangements for the maintenance of the packages? I don't want to step on anyone's toes - if updates for GNOME are already in the pipeline, I'd rather contribute to that effort than start my own. Do I just submit them to the package submission tracker and wait? I'm probably in over my head here, so feel free to tell me if I'm way off base - but I'd really like to give a little back :) Part of my drive is that I would like to get Evolution 1.4 beta up and running on Mac OS X - if I'm going to port Evolution (and GNOME 2.2.1) anyway, I might as well do it properly so everyone can enjoy it :) BTW: Another package I would like to see is Redhat's Bluecurve/artwork packages - I have them ported on my local fink setup, and so long as the licence allows it, this would be one of a few "miscellaneous extras" packages that I would push forward. thanks, Tony [EMAIL PROTECTED] http://www.iinet.net.au/~tonyarnold/ --- This SF.net email is sponsored by: ValueWeb: Dedicated Hosting for just $79/mo with 500 GB of bandwidth! No other company gives more support or power for your dedicated server http://click.atdmt.com/AFF/go/sdnxxaff00300020aff/direct/01/ ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
[Fink-devel] Fwd: [Fink-users] AM_PATH_GTKGL trouble
forwarded from fink-users: Begin forwarded message: From: Jonathan Brandmeyer <[EMAIL PROTECTED]> Date: Tue Apr 1, 2003 7:16:57 PM US/Eastern To: [EMAIL PROTECTED] Subject: [Fink-users] AM_PATH_GTKGL trouble I am a developer for a python extension module project. We are trying to support OSX using fink, due to its support for gtkglarea. We are using a slightly modified version of the autoconf macro 'AM_PATH_GTKGL' (its contents are appended to this message). The test fails with these linker errors: "_glPointParameterf referenced from libGL expected to be defined in OpenGL _glPointParameterfv referenced from libGL expected to be defined in OpenGL" Our test system is configured with: OS 10.2, Fink stable Apple X11 0.3 with SDK Apple's Python distro Developer tools dated June 2002. The following fink packages (with automatic dependancies): automake autoconf gtkglarea We attempted to build gtkglarea from its source package and recieved the same errors during a related check. Thanks for your assistance, Jonathan Brandmeyer The content of the test macro follows. I have added comments in all caps, prefixed with ##. - # configure paths for GtkGLArea # Janne Lof 1999-17-2 # AM_PATH_GTKGL([ACTION-IF-FOUND [,ACTION-IF-NOT-FOUND]]) AC_DEFUN(AM_PATH_GTKGL, [ AC_PROVIDE([AM_PATH_GTKGL]) AC_ARG_WITH(gl-prefix,[ --with-gl-prefix=PFXPrefix where OpenGL or Mesa is installed], gl_prefix="$withval", gl_prefix="") AC_ARG_WITH(gtkgl-prefix, [ --with-gtkgl-prefix=PFX Prefix where GtkGLArea is installed], gtkgl_prefix="$withval", gtkgl_prefix="") # tests for OpenGL if test x$gl_prefix != x ; then GL_CFLAGS="-I$gl_prefix/include" GL_LDOPTS="-L$gl_prefix/lib" else GL_CFLAGS="" GL_LDOPTS="" fi saved_LIBS="$LIBS" saved_CFLAGS="$CFLAGS" # test for plain OpenGL ## THIS TEST PASSES, SEVERAL ALTERNATES FOLLOW WHICH ARE NOT SEEN AC_MSG_CHECKING([GL]) LIBS="$saved_LIBS $GTK_LIBS $GL_LDOPTS -lGLU -lGL" AC_TRY_LINK( ,[ char glBegin(); glBegin(); ], have_GL=yes, have_GL=no) AC_MSG_RESULT($have_GL) if test x$have_GL = xyes; then GL_LIBS="-lGLU -lGL" else # test for GL with Pthreads. ## THIS TEST IS NOT SEEN (NOT NEEDED) AC_MSG_CHECKING([GL with threads]) LIBS="$saved_LIBS $GTK_LIBS $GL_LDOPTS -lGLU -lGL -lpthread" AC_TRY_LINK( ,[char glBegin(); glBegin(); ], have_GL=yes, have_GL=no) AC_MSG_RESULT($have_GL) if test x$have_GL = xyes; then GL_LIBS="-lGLU -lGL -lpthread" else # test for Mesa ## THIS TEST IS NOT SEEN (NOT NEEDED) AC_MSG_CHECKING([Mesa]) LIBS="$saved_LIBS $GTK_LIBS $GL_LDOPTS -lMesaGLU -lMesaGL" AC_TRY_LINK( ,[ char glBegin(); glBegin(); ], have_Mesa=yes, have_Mesa=no) AC_MSG_RESULT($have_Mesa) if test x$have_Mesa = xyes; then GL_LIBS="-lMesaGLU -lMesaGL" else # test for Mesa with threads ## THIS TEST IS NOT SEEN (NOT NEEDED) AC_MSG_CHECKING([Mesa with pthreads]) LIBS="$saved_LIBS $GTK_LIBS $GL_LDOPTS -lMesaGLU -lMesaGL -lpthread" AC_TRY_LINK( ,[ char glBegin(); glBegin(); ], have_Mesa_pthread=yes, have_Mesa_pthread=no) AC_MSG_RESULT($have_Mesa_pthread) if test x$have_Mesa_pthread = xyes; then GL_LIBS="-lMesaGLU -lMesaGL -lpthread" else #all failed LIBS="$saved_LIBS" CFLAGS="$saved_CFLAGS" GTKGL_LIBS="" GTKGL_CFLAGS="" ifelse([$2], , :, [$2]) fi fi fi fi # tests for GtkGLArea ## THIS TEST FAILS WITH LINKER ERRORS. if test x$gtkgl_prefix != x; then GTKGL_CFLAGS="-I$gtkgl_prefix/include" GTKGL_LDOPTS="-L$gtkgl_prefix/lib" else GTKGL_CFLAGS="" GTKGL_LDOPTS="" fi AC_MSG_CHECKING([GtkGLArea]) LIBS="$save_LIBS -lgtkgl $GTK_LIBS $GL_LDOPTS $GL_LIBS $GTKGL_LDOPTS" AC_TRY_LINK( ,[ char gtk_gl_area_new(); gtk_gl_area_new(); ], have_gtkgl=yes, have_gtkgl=no) AC_MSG_RESULT($have_gtkgl) if test x$have_gtkgl = xyes; then LIBS="$saved_LIBS" CFLAGS="$saved_CFLAGS" GTKGL_CFLAGS="$GTKGL_CFLAGS $GL_CFLAGS" GTKGL_LIBS="$GTKGL_LDOPTS -lgtkgl $GL_LDOPTS $GL_LIBS" ifelse([$1], , :, [$1]) else LIBS="$saved_LIBS" CFLAGS="$saved_CFLAGS" GTKGL_LIBS="" GTKGL_CFLAGS="" ifelse([$2], , :, [$2]) fi AC_SUBST(GTKGL_CFLAGS) AC_SUBST(GTKGL_LIBS) ]) --- This SF.net email is sponsored by: ValueWeb: Dedicated Hosting for just $79/mo with 500 GB of bandwidth! No other company gives more support or power for your dedicated server http://click.atdmt.com/AFF/go/sdnxxaff00300020aff/direct/01/ ___ Fink-users mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-users -chris zubrzycki - -- PGP public key: http://homepage.mac.com/beren/publickey.txt ID: 0xA2ABC070 Fingerprint: 26B0 BA6B A409 FA83 42B3 1688 FBF9 8232 A2AB C070 "Sadly, text alone cannot convey the depths of my sarcasm." --- This
Re: [Fink-devel] testing needed
With your latest gnuplot, I got: ... Enable generation of PDF files ... So it *seems* PDF support is built properly now. Also the bogus pdflib.dylib symlink is not present. Cheers, Max --- This SF.net email is sponsored by: ValueWeb: Dedicated Hosting for just $79/mo with 500 GB of bandwidth! No other company gives more support or power for your dedicated server http://click.atdmt.com/AFF/go/sdnxxaff00300020aff/direct/01/ ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: [Fink-devel] testing needed
On Apr 2,2003 00:38:09 +0200, Martin Costabel <[EMAIL PROTECTED]> wrote : >David R. Morrison wrote: >> Hello. Please help me quickly test some new versions of the following >> packages: pdflib, gnuplot, grace. The new versions are all in the >> unstable tree. It would be particularly useful to know that gnuplot-py, >> octave, and siag work correctly with the new version of gnuplot. > >Please tell me this is an April fool's joke ;-) > No, I'm afraid it was serious. >The new pdflib-shlibs package has a circular symlink >libpdf.1.dylib->libpdf.1.dylib >which makes anything depending on libpdf crash ("Too many levels of >symbolic links"). > Thanks for catching this; its now fixed in a new version. >The new gnuplot doesn't build a pdf terminal any more. This could have >been achieved easier without change to the pdflib package, but I suppose >it is not intentional? > Not intentional. I'll look into it. >The first instance of the new gnuplot I compiled failed with > >gnuplot> Expected X11 driver: >/sw/src/root-gnuplot-3.8i.0-3/sw/lib/gnuplot/3.8i/gnuplot_x11 >Exec failed: No such file or directory > >In fact, it had the path >/sw/src/root-gnuplot-3.8i.0-3/sw/lib/gnuplot/3.8i/gnuplot_x11 hardcoded >into the /sw/bin/gnuplotx executable. I cannot reproduce this any more, >because I rebuilt it and then this error disappeared. Weird. > >Otherwise, octave and siag seem to work OK with the new gnuplot (except >for pdf output, of course). > >-- >Martin > Thanks for the feedback. I'll see if I can get gnuplot fixed. -- Dave --- This SF.net email is sponsored by: ValueWeb: Dedicated Hosting for just $79/mo with 500 GB of bandwidth! No other company gives more support or power for your dedicated server http://click.atdmt.com/AFF/go/sdnxxaff00300020aff/direct/01/ ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: [Fink-devel] testing needed
David R. Morrison wrote: Hello. Please help me quickly test some new versions of the following packages: pdflib, gnuplot, grace. The new versions are all in the unstable tree. It would be particularly useful to know that gnuplot-py, octave, and siag work correctly with the new version of gnuplot. Please tell me this is an April fool's joke ;-) The new pdflib-shlibs package has a circular symlink libpdf.1.dylib->libpdf.1.dylib which makes anything depending on libpdf crash ("Too many levels of symbolic links"). The new gnuplot doesn't build a pdf terminal any more. This could have been achieved easier without change to the pdflib package, but I suppose it is not intentional? The first instance of the new gnuplot I compiled failed with gnuplot> Expected X11 driver: /sw/src/root-gnuplot-3.8i.0-3/sw/lib/gnuplot/3.8i/gnuplot_x11 Exec failed: No such file or directory In fact, it had the path /sw/src/root-gnuplot-3.8i.0-3/sw/lib/gnuplot/3.8i/gnuplot_x11 hardcoded into the /sw/bin/gnuplotx executable. I cannot reproduce this any more, because I rebuilt it and then this error disappeared. Weird. Otherwise, octave and siag seem to work OK with the new gnuplot (except for pdf output, of course). -- Martin --- This SF.net email is sponsored by: ValueWeb: Dedicated Hosting for just $79/mo with 500 GB of bandwidth! No other company gives more support or power for your dedicated server http://click.atdmt.com/AFF/go/sdnxxaff00300020aff/direct/01/ ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: [Fink-devel] Requiring MD5s
On Tuesday, April 1, 2003, at 01:16 AM, Ben Hines wrote: It has been a warning in validate for a while. just stumbled across this in the fink manpage. This would be a good thing to add to the porting guide. -- Charles Lepple <[EMAIL PROTECTED]> http://www.ghz.cc/charles/ --- This SF.net email is sponsored by: ValueWeb: Dedicated Hosting for just $79/mo with 500 GB of bandwidth! No other company gives more support or power for your dedicated server http://click.atdmt.com/AFF/go/sdnxxaff00300020aff/direct/01/ ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
[Fink-devel] testing needed
Hello. Please help me quickly test some new versions of the following packages: pdflib, gnuplot, grace. The new versions are all in the unstable tree. It would be particularly useful to know that gnuplot-py, octave, and siag work correctly with the new version of gnuplot. The change which was made was to only build static libraries for pdflib, and compile those into the other programs. This was done for licensing reasons, to enable us to distribute gnuplot and grace with our binary distribution. With this change, if I can get it pushed into stable quickly, we'll be in the following happy situation: every package which is in the binary distribution only depends on packages which are *also* in the binary distribution. No more user complaints that octave or pdftex can't be installed because of some silly missing thing from the bindist. Some of our Restrictive/Distributable packages will still have Build dependencies which can't be distributed, but they'll have no actual dependencies. When the rush of getting the bindist out is over, I'll be proposing this as a policy to be followed in the future, for all of our packages: if you package depends on a Restrictive pacakge, then your package mujst also be labelled as Restrictive. But let's not debate that now! Please, instead, help me test these few and let me know your results. Thanks, Dave --- This SF.net email is sponsored by: ValueWeb: Dedicated Hosting for just $79/mo with 500 GB of bandwidth! No other company gives more support or power for your dedicated server http://click.atdmt.com/AFF/go/sdnxxaff00300020aff/direct/01/ ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: [Fink-devel] Re: Requiring MD5s
Here is the relevant discussion from #fink (edited, times in JST). Apr 01 13:42:47 anyone object to requiring md5s for all packages? Apr 01 13:42:49 no? good Apr 01 14:07:37 Clef: What do you mean requiring? Apr 01 14:08:32 Clef: If you mean that the package will not build if there is no md5.. then yes, I object Apr 01 15:25:03 tough! Apr 01 15:27:13 how about only for stable and unstable trees, not local and exp. Apr 01 15:27:18 --> Murrito has joined #fink Apr 01 15:27:28 cause we really need MD5s for all packages. It needs to be an error. Apr 01 15:29:24 regading MD5 hashes, I once encountered a package site that would for some reaosn shuffle its tar files quite regularly, so you couldn't geta stable MD5 hash Apr 01 15:29:35 we mirror those files. Apr 01 15:29:42 thats no excuse. Apr 01 15:29:43 hmm, good point Apr 01 15:29:59 and the point anyway is that the opendarwin mirror wants md5s. Apr 01 15:30:04 we'll be mirroring everything soon Apr 01 15:31:26 mirroring? Apr 01 15:31:31 * bbraun2 perks up --- This SF.net email is sponsored by: ValueWeb: Dedicated Hosting for just $79/mo with 500 GB of bandwidth! No other company gives more support or power for your dedicated server http://click.atdmt.com/AFF/go/sdnxxaff00300020aff/direct/01/ ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: glut license (was Re: [Fink-devel] MD5)
-BEGIN PGP SIGNED MESSAGE- Hash: RIPEMD160 On Dienstag, April 1, 2003, at 03:49 Uhr, Ben Hines wrote: "distribution is allowed without license". To cut things short. I ran this by our lawyers and here is what they said: "This type of license indicates that you may distribute the entity referred to as "program" in any form that it might take. Be it binary or source. There is no implicit mentioning of 'binary distribution is not allowed' and in license agreements that means that you can safely assume that it may be distributed. Licenses are so called "restrictive approach measures" everything which is not explicitly denied is assumed to be legal" There you have it.. - -d - -BEGIN GEEK CODE BLOCK- Version: 3.12 GCC d+ s: a-- C+ UB P+ L++ E--- W N+ o+++ K w-- O M+ V++ PS PE Y++ PGP t+ 5 X- R+ tv-- b DI D+ G e h+ r++ y++ - --END GEEK CODE BLOCK-- -BEGIN PGP SIGNATURE- Version: GnuPG v1.2.1 (Darwin) iD8DBQE+iVqYiW/Ta/pxHPQRAzMzAJ9qrKNTPW8OmiKZj8iv6ym9yx4gyACgt2Xg tdUrvcRIMqhij+3bQ2Scakg= =ZBjk -END PGP SIGNATURE- --- This SF.net email is sponsored by: ValueWeb: Dedicated Hosting for just $79/mo with 500 GB of bandwidth! No other company gives more support or power for your dedicated server http://click.atdmt.com/AFF/go/sdnxxaff00300020aff/direct/01/ ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: [Fink-devel] Re: Requiring MD5s
Am Dienstag, 01.04.03 um 08:27 Uhr schrieb Ben Hines: On Monday, March 31, 2003, at 10:16 PM, Ben Hines wrote: For the mirroring setup with opendarwin we are going to need MD5s for all packages. I'm going to make fink refuse to unpack packages with no MD5 specified. It has been a warning in validate for a while. Maintainers, please add MD5s to everything that does not have it. There are quite a few packages in unstable that do not have them. ok ok, if people really object to this, perhaps I could exclude trees other than unstable and stable? (not sure if a package even knows what tree its in.. i could grep the path to the info file, though) Well, just going to the list and announcing "Hey folks, I am going to change Fink so that a couple dozen packages won't work anymore" w/o getting a prior consent isn't a very good idea in my eyes. I guess you talked about this on IRC, but for major change like this it would be nice if at the very least the Fink admings (RR, drm and me) were notified first. Also, posting a link to the channel logs with the discussion would be good. best would be to lead the discussion on the mailing list, or give a summary of it here. I think this is the right thing on the long term, but *not* something that should be done from one day to the next. Rather, first the offending packages should be determined, and their maintainers kicked one last time. Then we can make this change. And yeah I realize that any change in CVS will only be visible with the next PM release. Max --- This SF.net email is sponsored by: ValueWeb: Dedicated Hosting for just $79/mo with 500 GB of bandwidth! No other company gives more support or power for your dedicated server http://click.atdmt.com/AFF/go/sdnxxaff00300020aff/direct/01/ ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel
Re: glut license (was Re: [Fink-devel] MD5)
Am Dienstag, 01.04.03 um 03:49 Uhr schrieb Ben Hines: On Monday, March 31, 2003, at 05:33 PM, Matt Stephenson wrote: On Tuesday, Apr 1, 2003, at 11:17 Australia/Sydney, Ben Hines wrote: On Monday, March 31, 2003, at 02:16 AM, Matt Stephenson wrote: I tend towards 'Restrictive/Distributable' because nothing is mentioned specifically in the glut license about distributing binaries, its just a general statement or maybe I'm being to picky. Many licenses don't mention binaries. I still don't see what is restrictive about this license. What is restricted? So the only other license field we can use is 'OSI-Approved' which quote from our packaging manual "One of OSI's requirements is that free distribution of binaries and sources is allowed." I'm not saying the license is restrictive as such I just don't think it falls under OSI-Approved. It doesn't say that distribution of binaries and sources is not allowed, though. It says "distribution is allowed without license". Uhm, where does it say that? I only see: "The programs are not in the public domain, but they are freely distributable without licensing fees." Which is a difference. In fact, the only alternativs I see are: a) Public Domain (which they exclude) b) Some license (if that is the case, they don't mention it), be it OSI complian, restrictive or what c) Nothing - a legal void Right now it seems we are at c), with the exceptions that at least they tell us we can distribute it w/o a licensing fee. Alas, that doesn't actually mean that they grant us the *right* to distribute it. There are quite some software packages which are free to use but you are not allowed to distribute them, for example. So as long as the upstream maintainer leaves us at c), a legal void, we have to assume the worst and treat a package as "Restricted". The best solution in my eyes here would be to contact the glut creators and request a clarification by them. In the meantime, they are not OSI-Approved, because clearly an unknown license or a legal void can't be, not even in "spirit", OSI compliant. Cheers, Max --- This SF.net email is sponsored by: ValueWeb: Dedicated Hosting for just $79/mo with 500 GB of bandwidth! No other company gives more support or power for your dedicated server http://click.atdmt.com/AFF/go/sdnxxaff00300020aff/direct/01/ ___ Fink-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/fink-devel