From maciej at opencsw.org Sun Jan 4 13:56:48 2015 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Sun, 4 Jan 2015 12:56:48 +0000 Subject: How to find [easy] stuff to do Message-ID: Hello maintainers, We're well past the winter solstice, nights are getting shorter, but there are still many long evenings until the spring. If you're looking for something easy to do, you have, among other things, backgammon, and package building. Here's how to find easy (and not so easy) stuff to do: 1. Find stale packages and update them. You can look for your packages, and/or packages next to names in red: these are inactive maintainers. http://buildfarm.opencsw.org/obsolete-pkgs/stale-packages.html 2. Clean up after package renames. When we rename a package, the old package becomes a stub. To remove stubs we need to make sure nothing depends on them. We have a page which lists packages that need to be rebuilt (to free up stubs for removal): http://buildfarm.opencsw.org/obsolete-pkgs/PkgsToRebuildFrom_unstable_SunOS5.10_i386.html 3. Identify packages that don't work and can be removed. At the bottom of the stale packages report, there is a list of packages which were uploaded by someone who no longer uploads anything, and nothing depends on them. Install the package and try to use it. Does it work? If not, it should be removed. Broken packages hurt our project's reputation: People would download one package and if it's broken, they would assume all packages are like that (which we know is not true). 4. Find a small coding project http://wiki.opencsw.org/projects-available 5. Review other people's commits Subscribe to the devel@ mailing list and see what other people are doing. If you're new, you might learn something new. If you've been here a while, you might help someone else. ================================================== FAQ Q: Can I refresh package listed under somebody else's name? A: Yes! The exception are the core packages like GCC or Apache. In such case talk to people first. As we talked in October[1], we're accepting that our package database doesn't reflect reality as to who owns (cares about) which package. We only have information about who was the last person to upload each package. The two things are correlated, but not strongly. Q: What if I want a package to be assigned to my name? A: Rebuild it and re-upload. Have a happy 2015! Maciej [1] http://lists.opencsw.org/pipermail/maintainers/2014-October/019510.html From bonivart at opencsw.org Sun Jan 4 14:23:07 2015 From: bonivart at opencsw.org (Peter Bonivart) Date: Sun, 4 Jan 2015 14:23:07 +0100 Subject: How to find [easy] stuff to do In-Reply-To: References: Message-ID: On Sun, Jan 4, 2015 at 1:56 PM, Maciej (Matchek) Blizi?ski wrote: > FAQ > > Q: Can I refresh package listed under somebody else's name? > A: Yes! The exception are the core packages like GCC or Apache. In > such case talk to people first. As we talked in October[1], we're > accepting that our package database doesn't reflect reality as to who > owns (cares about) which package. We only have information about who > was the last person to upload each package. The two things are > correlated, but not strongly. > > Q: What if I want a package to be assigned to my name? > A: Rebuild it and re-upload. > > Have a happy 2015! > > Maciej > > [1] http://lists.opencsw.org/pipermail/maintainers/2014-October/019510.html Is this official because a couple of people see it that way? What are the core packages? Please give me a complete list of them. I remember trying to define tiers in 2011 (?) but we never got anywhere. Why are "core" packages an exception? As far as I know we don't handle them any different in any other way. I still expect to be contacted beforehand but the above pretty much states that anything goes. From maciej at opencsw.org Sun Jan 4 18:54:44 2015 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Sun, 4 Jan 2015 17:54:44 +0000 Subject: How to find [easy] stuff to do In-Reply-To: References: Message-ID: 2015-01-04 13:23 GMT+00:00 Peter Bonivart : > I still expect to be contacted beforehand but the above pretty much > states that anything goes. This wasn't what I meant, so Peter and I talked on IRC to clarify when and how much communication is needed. There is a dedicated page in the manual to this topic, we've updated it: http://www.opencsw.org/manual/for-maintainers/contributing-to-recipes.html Maciej From maciej at opencsw.org Mon Jan 5 00:40:00 2015 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Sun, 4 Jan 2015 23:40:00 +0000 Subject: How to find [easy] stuff to do In-Reply-To: References: Message-ID: I've found an old page listing which packages are in GAR and which aren't: http://www.opencsw.org/package-gar-status.html It's out of date. It was probably generated with updateGarSvnPackageStatus.sh[1]. I read it, and decided I don't want to use it, because it depends on a very specific environment set up: a special subversion checkout and directories. Instead I wrote a small program which uses data from pkgdb via REST[2]. It uses a simple heuristic: looks for the OPENCSW_REPOSITORY entry in pkginfo. There probably is a better way to do it, but this gives decent results for now. Here's what came back: http://buildfarm.opencsw.org/~maciej/not-in-gar.txt When I have a little more time, I'll make it run faster and I'll turn it into a cron job. Maciej [1] http://sourceforge.net/p/opencsw/code/HEAD/tree/utilities/upstream_watch/updateGarSvnPackageStatus.sh [2] https://sourceforge.net/p/gar/code/24543/ From grzemba at contac-dt.de Mon Jan 5 15:46:10 2015 From: grzemba at contac-dt.de (Carsten Grzemba) Date: Mon, 05 Jan 2015 15:46:10 +0100 Subject: Wintercamp 2014 Ilmenau? In-Reply-To: References: Message-ID: I would start the organization of the Wintercamp which should take place in Ilmenau. Or there are any other plans? Carsten -------------- next part -------------- An HTML attachment was scrubbed... URL: From grzemba at contac-dt.de Thu Jan 8 08:43:03 2015 From: grzemba at contac-dt.de (Carsten Grzemba) Date: Thu, 08 Jan 2015 08:43:03 +0100 Subject: Doodle Wintercamp 2015 In-Reply-To: References: Message-ID: I setup a doodle poll for OpenCSW Winter Camp 2015 which should take place in Ilmenau (50??41??N, 10??55??O) http://doodle.com/zkgvgc53ydqeff26 Carsten -------------- next part -------------- An HTML attachment was scrubbed... URL: From pfelecan at opencsw.org Thu Jan 8 10:46:35 2015 From: pfelecan at opencsw.org (Peter FELECAN) Date: Thu, 08 Jan 2015 10:46:35 +0100 Subject: Doodle Wintercamp 2015 In-Reply-To: (Carsten Grzemba's message of "Thu, 08 Jan 2015 08:43:03 +0100") References: Message-ID: Carsten Grzemba writes: > I setup a doodle poll for OpenCSW Winter Camp 2015 which should take > place in Ilmenau (50??41??N, 10??55??O) Can you give us a little bit more information about Ilmenau, where do you think that the camp will take place, &c. -- Peter From grzemba at contac-dt.de Thu Jan 8 11:20:51 2015 From: grzemba at contac-dt.de (Carsten Grzemba) Date: Thu, 08 Jan 2015 11:20:51 +0100 Subject: Doodle Wintercamp 2015 In-Reply-To: References: Message-ID: of course, I setup a wiki page Am 08.01.15 schrieb Peter FELECAN : > Carsten Grzemba writes: > > > I setup a doodle poll for OpenCSW Winter Camp 2015 which should take > > place in Ilmenau (50??41??N, 10??55??O) > > Can you give us a little bit more information about Ilmenau, where do > you think that the camp will take place, &c. > -- > Peter > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From dam at opencsw.org Thu Jan 15 15:26:17 2015 From: dam at opencsw.org (Dagobert Michelsen) Date: Thu, 15 Jan 2015 15:26:17 +0100 Subject: Subversion Message-ID: <324A426A-204A-42BE-BF23-125BEB840FE0@opencsw.org> Hi Rupert, in your subversion update several things are broken. I noticed you bumped the version and silenced checkpkg: https://buildfarm.opencsw.org/source/diff/opencsw/csw/mgar/pkg/subversion/trunk/Makefile?r2=%2Fopencsw%2Fcsw%2Fmgar%2Fpkg%2Fsubversion%2Ftrunk%2FMakefile%4024518&r1=%2Fopencsw%2Fcsw%2Fmgar%2Fpkg%2Fsubversion%2Ftrunk%2FMakefile%4023460 This is unfortunate as the checkpkg errors would have lead you to the issue. Properly updating subversion is difficult, after investigating for the last two days these are the prerequisites I have identified: - location of the subversion config changed back from /opt/csw/etc/subversion to /etc/subversion leading to the buildfarm config to be not pulled in. - libtool needs to be updated to honour gcc 4.9.x tags - this will probably help in allowing cc/gcc mixture of subversion and the foreign function apis - python module needs to be modulated for 2.6/2.7/3.x So, please, do not just override the checkpkg errors without investigation or we will have broken packages. Thanks ? Dago -- "You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process." - xkcd #896 -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 2418 bytes Desc: not available URL: From rupert at opencsw.org Thu Jan 15 21:54:07 2015 From: rupert at opencsw.org (rupert THURNER) Date: Thu, 15 Jan 2015 21:54:07 +0100 Subject: Subversion In-Reply-To: <324A426A-204A-42BE-BF23-125BEB840FE0@opencsw.org> References: <324A426A-204A-42BE-BF23-125BEB840FE0@opencsw.org> Message-ID: I m sorry for this. What would be the best strategy to avoid a change in the recipe when doing a patch level upgrade? I d be glad if this could be done by just changing the version number and maybe the names of referenced dependencies. Rupert On Jan 15, 2015 3:26 PM, "Dagobert Michelsen" wrote: > Hi Rupert, > > in your subversion update several things are broken. I noticed you bumped > the > version and silenced checkpkg: > > https://buildfarm.opencsw.org/source/diff/opencsw/csw/mgar/pkg/subversion/trunk/Makefile?r2=%2Fopencsw%2Fcsw%2Fmgar%2Fpkg%2Fsubversion%2Ftrunk%2FMakefile%4024518&r1=%2Fopencsw%2Fcsw%2Fmgar%2Fpkg%2Fsubversion%2Ftrunk%2FMakefile%4023460 > > This is unfortunate as the checkpkg errors would have lead you to the > issue. > Properly updating subversion is difficult, after investigating for the > last two > days these are the prerequisites I have identified: > > - location of the subversion config changed back from > /opt/csw/etc/subversion to > /etc/subversion leading to the buildfarm config to be not pulled in. > - libtool needs to be updated to honour gcc 4.9.x tags > - this will probably help in allowing cc/gcc mixture of subversion and the > foreign function apis > - python module needs to be modulated for 2.6/2.7/3.x > > So, please, do not just override the checkpkg errors without investigation > or > we will have broken packages. > > > Thanks > > ? Dago > > -- > "You don't become great by trying to be great, you become great by wanting > to do something, > and then doing it so hard that you become great in the process." - xkcd > #896 > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From dam at opencsw.org Thu Jan 15 22:46:47 2015 From: dam at opencsw.org (Dagobert Michelsen) Date: Thu, 15 Jan 2015 22:46:47 +0100 Subject: Subversion In-Reply-To: References: <324A426A-204A-42BE-BF23-125BEB840FE0@opencsw.org> Message-ID: <6327C5F6-E1F3-4060-A022-A9FE91FA318C@opencsw.org> Hi Rupert, > Am 15.01.2015 um 21:54 schrieb rupert THURNER : > > I m sorry for this. What would be the best strategy to avoid a change in the recipe when doing a patch level upgrade? I d be glad if this could be done by just changing the version number and maybe the names of referenced dependencies. Short answer: You can?t. It may be possible if you bump the version within a week or so, but not after several month. There are multiple reasons for this: - dependency chains for updates of other packages. This includes updates from gcc 4.8 to 4.9 requiring libtool tag changes which are difficult and have been missed during the last update. - compiler changes from Sun Studio to GCC. We are transitioning all packages to GCC, this requires adjustments in flags in mixed environments where some things are still compiled with Sun Studio and the new package with gcc. - multiple python versions requires modulations (also long standing project) - incompatible library versions upgrades - upstream changes in functionality, e.g. switches from OpenSSL to GnuTLS, cares to adns etc. - patches no longer apply because of changed source structure - different linker result on sparc/x86 (open ticket at Oracle), should be mostly resolved now. The times that checkpkg changed so often the overrides need to be adjusted are mostly over, so if checkpkg reports something it is worth investigating because it may be any of the above real problems or even other things. Checkpkg helps you to find the most obvious problems during version bumps, it is not your enemy. Best regards ? Dago -- "You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process." - xkcd #896 -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 2418 bytes Desc: not available URL: From dam at opencsw.org Wed Jan 21 17:42:01 2015 From: dam at opencsw.org (Dagobert Michelsen) Date: Wed, 21 Jan 2015 17:42:01 +0100 Subject: Checkpkg false positive References: <99584327-B060-4114-93BE-1DB4B0CE3467@baltic-online.de> Message-ID: <2C91C5F9-CCDF-4068-8B48-A22F1D675375@opencsw.org> Hi folks, looks like there are false positives on checkpkg when there are multiple symlink redirects. The errors look like this: CHECKPKG_OVERRIDES_CSWlibicu-dev += file-needed-but-no-package-satisfies-it|/opt/csw/lib/amd64/icu/current/Makefile.inc|CSWlibicu-dev|contains|symlink|/opt/csw/lib/amd64/icu/Makefile.inc|which|needs|the|target|file:|/opt/csw/lib/amd64/icu/current/Makefile.inc CHECKPKG_OVERRIDES_CSWlibicu-dev += file-needed-but-no-package-satisfies-it|/opt/csw/lib/amd64/icu/current/pkgdata.inc|CSWlibicu-dev|contains|symlink|/opt/csw/lib/amd64/icu/pkgdata.inc|which|needs|the|target|file:|/opt/csw/lib/amd64/icu/current/pkgdata.inc CHECKPKG_OVERRIDES_CSWlibicu-dev += file-needed-but-no-package-satisfies-it|/opt/csw/lib/icu/current/Makefile.inc|CSWlibicu-dev|contains|symlink|/opt/csw/lib/icu/Makefile.inc|which|needs|the|target|file:|/opt/csw/lib/icu/current/Makefile.inc CHECKPKG_OVERRIDES_CSWlibicu-dev += file-needed-but-no-package-satisfies-it|/opt/csw/lib/icu/current/pkgdata.inc|CSWlibicu-dev|contains|symlink|/opt/csw/lib/icu/pkgdata.inc|which|needs|the|target|file:|/opt/csw/lib/icu/current/pkgdata.inc This is the filetree at that position: dam at unstable10x [global]:/home/dam/mgar/pkg/libicu54/trunk > find work/solaris10-i386/pkgroot/opt/csw/lib/icu -ls 5318390 2 drwxr-xr-x 3 dam csw 6 Jan 21 16:56 work/solaris10-i386/pkgroot/opt/csw/lib/icu 5318391 1 lrwxrwxrwx 1 dam csw 20 Jan 21 16:56 work/solaris10-i386/pkgroot/opt/csw/lib/icu/Makefile.inc -> current/Makefile.inc 5318392 1 lrwxrwxrwx 1 dam csw 19 Jan 21 16:56 work/solaris10-i386/pkgroot/opt/csw/lib/icu/pkgdata.inc -> current/pkgdata.inc 5318394 2 drwxr-xr-x 2 dam csw 4 Jan 21 16:56 work/solaris10-i386/pkgroot/opt/csw/lib/icu/54.1 5318396 1 -rw-r--r-- 1 dam csw 548 Jan 21 16:55 work/solaris10-i386/pkgroot/opt/csw/lib/icu/54.1/pkgdata.inc 5318395 6 -rw-r--r-- 1 dam csw 9826 Jan 21 16:55 work/solaris10-i386/pkgroot/opt/csw/lib/icu/54.1/Makefile.inc 5318393 1 lrwxrwxrwx 1 dam csw 4 Jan 21 16:56 work/solaris10-i386/pkgroot/opt/csw/lib/icu/current -> 54.1 Best regards ? Dago -- "You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process." - xkcd #896 -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 2418 bytes Desc: not available URL: From dam at opencsw.org Thu Jan 22 11:46:16 2015 From: dam at opencsw.org (Dagobert Michelsen) Date: Thu, 22 Jan 2015 11:46:16 +0100 Subject: Checkpkg false positive In-Reply-To: <2C91C5F9-CCDF-4068-8B48-A22F1D675375@opencsw.org> References: <99584327-B060-4114-93BE-1DB4B0CE3467@baltic-online.de> <2C91C5F9-CCDF-4068-8B48-A22F1D675375@opencsw.org> Message-ID: <69C9626D-AB97-4989-BF84-01DB48480651@opencsw.org> Hi folks, > Am 21.01.2015 um 17:42 schrieb Dagobert Michelsen : > > looks like there are false positives on checkpkg when there are multiple > symlink redirects. The errors look like this: > > CHECKPKG_OVERRIDES_CSWlibicu-dev += file-needed-but-no-package-satisfies-it|/opt/csw/lib/amd64/icu/current/Makefile.inc|CSWlibicu-dev|contains|symlink|/opt/csw/lib/amd64/icu/Makefile.inc|which|needs|the|target|file:|/opt/csw/lib/amd64/icu/current/Makefile.inc > CHECKPKG_OVERRIDES_CSWlibicu-dev += file-needed-but-no-package-satisfies-it|/opt/csw/lib/amd64/icu/current/pkgdata.inc|CSWlibicu-dev|contains|symlink|/opt/csw/lib/amd64/icu/pkgdata.inc|which|needs|the|target|file:|/opt/csw/lib/amd64/icu/current/pkgdata.inc > CHECKPKG_OVERRIDES_CSWlibicu-dev += file-needed-but-no-package-satisfies-it|/opt/csw/lib/icu/current/Makefile.inc|CSWlibicu-dev|contains|symlink|/opt/csw/lib/icu/Makefile.inc|which|needs|the|target|file:|/opt/csw/lib/icu/current/Makefile.inc > CHECKPKG_OVERRIDES_CSWlibicu-dev += file-needed-but-no-package-satisfies-it|/opt/csw/lib/icu/current/pkgdata.inc|CSWlibicu-dev|contains|symlink|/opt/csw/lib/icu/pkgdata.inc|which|needs|the|target|file:|/opt/csw/lib/icu/current/pkgdata.inc > > This is the filetree at that position: > > dam at unstable10x [global]:/home/dam/mgar/pkg/libicu54/trunk > find work/solaris10-i386/pkgroot/opt/csw/lib/icu -ls > 5318390 2 drwxr-xr-x 3 dam csw 6 Jan 21 16:56 work/solaris10-i386/pkgroot/opt/csw/lib/icu > 5318391 1 lrwxrwxrwx 1 dam csw 20 Jan 21 16:56 work/solaris10-i386/pkgroot/opt/csw/lib/icu/Makefile.inc -> current/Makefile.inc > 5318392 1 lrwxrwxrwx 1 dam csw 19 Jan 21 16:56 work/solaris10-i386/pkgroot/opt/csw/lib/icu/pkgdata.inc -> current/pkgdata.inc > 5318394 2 drwxr-xr-x 2 dam csw 4 Jan 21 16:56 work/solaris10-i386/pkgroot/opt/csw/lib/icu/54.1 > 5318396 1 -rw-r--r-- 1 dam csw 548 Jan 21 16:55 work/solaris10-i386/pkgroot/opt/csw/lib/icu/54.1/pkgdata.inc > 5318395 6 -rw-r--r-- 1 dam csw 9826 Jan 21 16:55 work/solaris10-i386/pkgroot/opt/csw/lib/icu/54.1/Makefile.inc > 5318393 1 lrwxrwxrwx 1 dam csw 4 Jan 21 16:56 work/solaris10-i386/pkgroot/opt/csw/lib/icu/current -> 54.1 I overrode these for now. It looks like the error is systematic and also present in the csw-upload-pkg code: Checking 9 packages against catalog unstable i386 SunOS5.10 WARNING 2015-01-22 11:42:56,672 checkpkg_lib.py:658 Did not find packages for '/opt/csw/lib/amd64/icu/current/Makefile.inc' WARNING 2015-01-22 11:42:56,705 checkpkg_lib.py:658 Did not find packages for '/opt/csw/lib/amd64/icu/current/pkgdata.inc' WARNING 2015-01-22 11:42:56,725 checkpkg_lib.py:658 Did not find packages for '/opt/csw/lib/icu/current/Makefile.inc' WARNING 2015-01-22 11:42:56,761 checkpkg_lib.py:658 Did not find packages for '/opt/csw/lib/icu/current/pkgdata.inc? Best regards ? Dago -- "You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process." - xkcd #896 -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 2418 bytes Desc: not available URL: From maciej at opencsw.org Fri Jan 23 09:44:14 2015 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Fri, 23 Jan 2015 08:44:14 +0000 Subject: Checkpkg false positive In-Reply-To: <2C91C5F9-CCDF-4068-8B48-A22F1D675375@opencsw.org> References: <99584327-B060-4114-93BE-1DB4B0CE3467@baltic-online.de> <2C91C5F9-CCDF-4068-8B48-A22F1D675375@opencsw.org> Message-ID: It's interesting that it took us that long to encounter this! s foo ? bar/foo s bar ? baz f baz/foo This means that there is no regular file with path bar/foo. From skimming the code, checkpkg looks for packages containing pkgmap entries equal to the target of the symlink, which does not yield a correct result in this case. To fix this, checkpkg needs to follow directory symlinks. If anyone wants to take that on, the place in the code is checkpkg_lib.py:651, and there are existing unit tests you can copy and modify to expose the problem. Maciej -------------- next part -------------- An HTML attachment was scrubbed... URL: From rmottola at opencsw.org Fri Jan 23 14:35:30 2015 From: rmottola at opencsw.org (Riccardo Mottola) Date: Fri, 23 Jan 2015 14:35:30 +0100 Subject: EXTRA_RUNPATH_DIRS, LD_OPTIONS an B direct In-Reply-To: <2C4A66BC-EB65-4F7F-B013-21C361147270@opencsw.org> References: <549ADEBA.2050608@opencsw.org> <2C4A66BC-EB65-4F7F-B013-21C361147270@opencsw.org> Message-ID: <54C24E22.7040005@opencsw.org> Hi Dagobert, sorry for the delay.... Dagobert Michelsen wrote: >> as well as >> >> CHECKPKG_OVERRIDES_CSWgnustep-base += soname-not-found|libgnustep-base.so.1.24|is|needed|by|opt/csw/GNUstep/System/Tools/gdnc >> >> which is different since it is some of the tools against the library itself being built. >> Here Dago suggests setting EXTRA_RUNPATH_DIRS. >> >> I added in my Makefile >> EXTRA_RUNPATH_DIRS = /opt/csw/GNUstep/System/Library/Libraries >> >> but I still get the warning! > Try ldd -r on libgnustep-base.so.1.24 and see if all libraries are found. If not inspect > with dump -Lv to see the RPATH and make sure all directories with linked libs are in the > RPATH. This is also explained at > http://wiki.opencsw.org/checkpkg-error-tags#toc54 > It looks they are! I checked the object in the build directory, see: rmottola at unstable10s [unstable10s]:~/opencsw/gnustep-base/trunk > ldd -r work/solaris10-sparc/build-isa-sparcv8plus/gnustep-base-1.24.7/ Source/obj/libgnustep-base.so libobjc.so.4 => /opt/csw/lib/libobjc.so.4 libgmp.so.10 => /opt/csw/lib/libgmp.so.10 libgnutls.so.28 => /opt/csw/lib/libgnutls.so.28 libxslt.so.1 => /opt/csw/lib/libxslt.so.1 libxml2.so.2 => /opt/csw/lib/libxml2.so.2 libz.so.1 => /opt/csw/lib/libz.so.1 libpthread.so.1 => /lib/libpthread.so.1 libiconv.so.2 => /opt/csw/lib/libiconv.so.2 libm.so.2 => /lib/libm.so.2 libsocket.so.1 => /lib/libsocket.so.1 libnsl.so.1 => /lib/libnsl.so.1 libffi.so.5 => /opt/csw/lib/libffi.so.5 librt.so.1 => /lib/librt.so.1 libdl.so.1 => /lib/libdl.so.1 libicui18n.so.54 => /opt/csw/lib/libicui18n.so.54 libicuuc.so.54 => /opt/csw/lib/libicuuc.so.54 libicudata.so.54 => /opt/csw/lib/libicudata.so.54 libgcc_s.so.1 => /opt/csw/lib/libgcc_s.so.1 libc.so.1 => /lib/libc.so.1 libp11-kit.so.0 => /opt/csw/lib/libp11-kit.so.0 libnettle.so.4 => /opt/csw/lib/libnettle.so.4 libhogweed.so.2 => /opt/csw/lib/libhogweed.so.2 libintl.so.8 => /opt/csw/lib/libintl.so.8 libz.so.1 => /opt/csw/lib/sparcv8plus+vis/libz.so.1 liblzma.so.5 => /opt/csw/lib/sparcv8/liblzma.so.5 libmp.so.2 => /lib/libmp.so.2 libmd.so.1 => /lib/libmd.so.1 libscf.so.1 => /lib/libscf.so.1 libaio.so.1 => /lib/libaio.so.1 libCrun.so.1 => /usr/lib/libCrun.so.1 libCstd.so.1 => /usr/lib/libCstd.so.1 libm.so.1 => /lib/libm.so.1 libdoor.so.1 => /lib/libdoor.so.1 libuutil.so.1 => /lib/libuutil.so.1 libgen.so.1 => /lib/libgen.so.1 /platform/SUNW,SPARC-Enterprise-T5220/lib/libc_psr.so.1 /platform/SUNW,SPARC-Enterprise-T5220/lib/libmd_psr.so.1 /usr/lib/cpu/sparcv8plus/libCstd_isa.so.1 Does it means it is a false warning? Riccardo From dam at opencsw.org Fri Jan 23 16:54:54 2015 From: dam at opencsw.org (Dagobert Michelsen) Date: Fri, 23 Jan 2015 16:54:54 +0100 Subject: EXTRA_RUNPATH_DIRS, LD_OPTIONS an B direct In-Reply-To: <54C24E22.7040005@opencsw.org> References: <549ADEBA.2050608@opencsw.org> <2C4A66BC-EB65-4F7F-B013-21C361147270@opencsw.org> <54C24E22.7040005@opencsw.org> Message-ID: Hi Riccardo, > Am 23.01.2015 um 14:35 schrieb Riccardo Mottola : > > sorry for the delay?. *I* don?t want this released fast ;-) > Dagobert Michelsen wrote: >>> as well as >>> >>> CHECKPKG_OVERRIDES_CSWgnustep-base += soname-not-found|libgnustep-base.so.1.24|is|needed|by|opt/csw/GNUstep/System/Tools/gdnc >>> >>> which is different since it is some of the tools against the library itself being built. >>> Here Dago suggests setting EXTRA_RUNPATH_DIRS. >>> >>> I added in my Makefile >>> EXTRA_RUNPATH_DIRS = /opt/csw/GNUstep/System/Library/Libraries >>> >>> but I still get the warning! >> Try ldd -r on libgnustep-base.so.1.24 and see if all libraries are found. If not inspect >> with dump -Lv to see the RPATH and make sure all directories with linked libs are in the >> RPATH. This is also explained at >> http://wiki.opencsw.org/checkpkg-error-tags#toc54 >> > It looks they are! I checked the object in the build directory, see: > rmottola at unstable10s [unstable10s]:~/opencsw/gnustep-base/trunk > ldd -r work/solaris10-sparc/build-isa-sparcv8plus/gnustep-base-1.24.7/ > Source/obj/libgnustep-base.so > libobjc.so.4 => /opt/csw/lib/libobjc.so.4 > libgmp.so.10 => /opt/csw/lib/libgmp.so.10 > libgnutls.so.28 => /opt/csw/lib/libgnutls.so.28 > libxslt.so.1 => /opt/csw/lib/libxslt.so.1 > libxml2.so.2 => /opt/csw/lib/libxml2.so.2 > libz.so.1 => /opt/csw/lib/libz.so.1 > libpthread.so.1 => /lib/libpthread.so.1 > libiconv.so.2 => /opt/csw/lib/libiconv.so.2 > libm.so.2 => /lib/libm.so.2 > libsocket.so.1 => /lib/libsocket.so.1 > libnsl.so.1 => /lib/libnsl.so.1 > libffi.so.5 => /opt/csw/lib/libffi.so.5 > librt.so.1 => /lib/librt.so.1 > libdl.so.1 => /lib/libdl.so.1 > libicui18n.so.54 => /opt/csw/lib/libicui18n.so.54 > libicuuc.so.54 => /opt/csw/lib/libicuuc.so.54 > libicudata.so.54 => /opt/csw/lib/libicudata.so.54 > libgcc_s.so.1 => /opt/csw/lib/libgcc_s.so.1 > libc.so.1 => /lib/libc.so.1 > libp11-kit.so.0 => /opt/csw/lib/libp11-kit.so.0 > libnettle.so.4 => /opt/csw/lib/libnettle.so.4 > libhogweed.so.2 => /opt/csw/lib/libhogweed.so.2 > libintl.so.8 => /opt/csw/lib/libintl.so.8 > libz.so.1 => /opt/csw/lib/sparcv8plus+vis/libz.so.1 > liblzma.so.5 => /opt/csw/lib/sparcv8/liblzma.so.5 > libmp.so.2 => /lib/libmp.so.2 > libmd.so.1 => /lib/libmd.so.1 > libscf.so.1 => /lib/libscf.so.1 > libaio.so.1 => /lib/libaio.so.1 > libCrun.so.1 => /usr/lib/libCrun.so.1 > libCstd.so.1 => /usr/lib/libCstd.so.1 > libm.so.1 => /lib/libm.so.1 > libdoor.so.1 => /lib/libdoor.so.1 > libuutil.so.1 => /lib/libuutil.so.1 > libgen.so.1 => /lib/libgen.so.1 > /platform/SUNW,SPARC-Enterprise-T5220/lib/libc_psr.so.1 > /platform/SUNW,SPARC-Enterprise-T5220/lib/libmd_psr.so.1 > /usr/lib/cpu/sparcv8plus/libCstd_isa.so.1 > > Does it means it is a false warning? No, it means you are checking the wrong thing. Please reread: > CHECKPKG_OVERRIDES_CSWgnustep-base += soname-not-found|libgnustep-base.so.1.24|is|needed|by|opt/csw/GNUstep/System/Tools/gdnc Try ldd -r on gdnc, then libgnustep-base.so.1.24 can not be found. Best regards ? Dago -- "You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process." - xkcd #896 -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 2418 bytes Desc: not available URL: From rmottola at opencsw.org Fri Jan 23 23:21:38 2015 From: rmottola at opencsw.org (Riccardo Mottola) Date: Fri, 23 Jan 2015 23:21:38 +0100 Subject: EXTRA_RUNPATH_DIRS, LD_OPTIONS an B direct In-Reply-To: References: <549ADEBA.2050608@opencsw.org> <2C4A66BC-EB65-4F7F-B013-21C361147270@opencsw.org> <54C24E22.7040005@opencsw.org> Message-ID: <54C2C972.2030709@opencsw.org> Hi, Dagobert Michelsen wrote: >> It looks they are! I checked the object in the build directory, see: >> rmottola at unstable10s [unstable10s]:~/opencsw/gnustep-base/trunk > ldd -r work/solaris10-sparc/build-isa-sparcv8plus/gnustep-base-1.24.7/ >> Source/obj/libgnustep-base.so >> libobjc.so.4 => /opt/csw/lib/libobjc.so.4 >> libgmp.so.10 => /opt/csw/lib/libgmp.so.10 >> libgnutls.so.28 => /opt/csw/lib/libgnutls.so.28 >> libxslt.so.1 => /opt/csw/lib/libxslt.so.1 >> libxml2.so.2 => /opt/csw/lib/libxml2.so.2 >> libz.so.1 => /opt/csw/lib/libz.so.1 >> libpthread.so.1 => /lib/libpthread.so.1 >> libiconv.so.2 => /opt/csw/lib/libiconv.so.2 >> libm.so.2 => /lib/libm.so.2 >> libsocket.so.1 => /lib/libsocket.so.1 >> libnsl.so.1 => /lib/libnsl.so.1 >> libffi.so.5 => /opt/csw/lib/libffi.so.5 >> librt.so.1 => /lib/librt.so.1 >> libdl.so.1 => /lib/libdl.so.1 >> libicui18n.so.54 => /opt/csw/lib/libicui18n.so.54 >> libicuuc.so.54 => /opt/csw/lib/libicuuc.so.54 >> libicudata.so.54 => /opt/csw/lib/libicudata.so.54 >> libgcc_s.so.1 => /opt/csw/lib/libgcc_s.so.1 >> libc.so.1 => /lib/libc.so.1 >> libp11-kit.so.0 => /opt/csw/lib/libp11-kit.so.0 >> libnettle.so.4 => /opt/csw/lib/libnettle.so.4 >> libhogweed.so.2 => /opt/csw/lib/libhogweed.so.2 >> libintl.so.8 => /opt/csw/lib/libintl.so.8 >> libz.so.1 => /opt/csw/lib/sparcv8plus+vis/libz.so.1 >> liblzma.so.5 => /opt/csw/lib/sparcv8/liblzma.so.5 >> libmp.so.2 => /lib/libmp.so.2 >> libmd.so.1 => /lib/libmd.so.1 >> libscf.so.1 => /lib/libscf.so.1 >> libaio.so.1 => /lib/libaio.so.1 >> libCrun.so.1 => /usr/lib/libCrun.so.1 >> libCstd.so.1 => /usr/lib/libCstd.so.1 >> libm.so.1 => /lib/libm.so.1 >> libdoor.so.1 => /lib/libdoor.so.1 >> libuutil.so.1 => /lib/libuutil.so.1 >> libgen.so.1 => /lib/libgen.so.1 >> /platform/SUNW,SPARC-Enterprise-T5220/lib/libc_psr.so.1 >> /platform/SUNW,SPARC-Enterprise-T5220/lib/libmd_psr.so.1 >> /usr/lib/cpu/sparcv8plus/libCstd_isa.so.1 >> >> Does it means it is a false warning? > No, it means you are checking the wrong thing. Please reread: > >> CHECKPKG_OVERRIDES_CSWgnustep-base += soname-not-found|libgnustep-base.so.1.24|is|needed|by|opt/csw/GNUstep/System/Tools/gdnc I did that for this warning: * No symbol of binary /opt/csw/GNUstep/System/Library/Libraries/libgnustep- base.so.1.24.7 is directly bound against the following libraries: libffi.so.5, libgnutls.so.28, libiconv.so.2, libicui18n.so.54, libicuuc.so.54, libobjc.so.4, libxml2.so.2, libxslt.so.1, libz.so.1. Please make sure the binaries are compiled using the "-Bdirect" linker option. > > Try ldd -r on gdnc, then libgnustep-base.so.1.24 can not be found. > libgnustep-base.so.1.24 => (file not found) that indeed shows a lot of missing stuff! However, it is custom for GNUstep stuff that you first need to source a shell script. GNUstep has bundles like MacOS where libraries are in, so locaiton is a little more dynamic. in fact if I do ". /opt/csw/GNUstep/System/Library/Makefiles/GNUstep.sh " it should work. I will try t add all the exception regarding this problem and see if once installed everything works or not. I do wonder about the above stuff instead which are opencsw libraries. Riccardo From rmottola at opencsw.org Sat Jan 24 00:15:17 2015 From: rmottola at opencsw.org (Riccardo Mottola) Date: Sat, 24 Jan 2015 00:15:17 +0100 Subject: EXTRA_RUNPATH_DIRS, LD_OPTIONS an B direct In-Reply-To: References: <549ADEBA.2050608@opencsw.org> <2C4A66BC-EB65-4F7F-B013-21C361147270@opencsw.org> <54C24E22.7040005@opencsw.org> Message-ID: <54C2D605.1020808@opencsw.org> Hi, Dagobert Michelsen wrote: > Try ldd -r on gdnc, then libgnustep-base.so.1.24 can not be found. I came to the conclusion that for teh standard setup this is fine and should work. However, other warnings do worry me: * No symbol of binary /opt/csw/GNUstep/System/Tools/autogsdoc is directly bound against the following libraries: libgcc_s.so.1, libgnustep- base.so.1.24, libobjc.so.4. Please make sure the binaries are compiled using the "-Bdirect" linker option. I explained libgnustep-base. But libgcc and libobjc are "standard" stuff from opencsw, it should just work? In fact: rmottola at unstable10s [unstable10s]:~/opencsw/gnustep-base/trunk > ldd ./work/solaris10-sparc/install-isa-sparcv8plus/opt/csw/GNUstep/System/Tools/autogsdoc libgnustep-base.so.1.24 => (file not found) libobjc.so.4 => /opt/csw/lib/libobjc.so.4 libsocket.so.1 => /lib/libsocket.so.1 libnsl.so.1 => /lib/libnsl.so.1 libm.so.2 => /lib/libm.so.2 libgcc_s.so.1 => /opt/csw/lib/libgcc_s.so.1 libpthread.so.1 => /lib/libpthread.so.1 libc.so.1 => /lib/libc.so.1 libmp.so.2 => /lib/libmp.so.2 libmd.so.1 => /lib/libmd.so.1 libscf.so.1 => /lib/libscf.so.1 libdoor.so.1 => /lib/libdoor.so.1 libuutil.so.1 => /lib/libuutil.so.1 libgen.so.1 => /lib/libgen.so.1 /platform/SUNW,SPARC-Enterprise-T5220/lib/libc_psr.so.1 /platform/SUNW,SPARC-Enterprise-T5220/lib/libmd_psr.so.1 Looks fine except for gnustep-base? And what about these warnigs? CHECKPKG_OVERRIDES_CSWgnustep-base += binary-architecture-does-not-match-placement|file=opt/csw/GNUstep/System/Tools/defaults|arch_id=18|arch_name=sparcv8+ Thank you Riccardo From maciej at opencsw.org Sat Jan 24 13:46:42 2015 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Sat, 24 Jan 2015 12:46:42 +0000 Subject: EXTRA_RUNPATH_DIRS, LD_OPTIONS an B direct References: <549ADEBA.2050608@opencsw.org> <2C4A66BC-EB65-4F7F-B013-21C361147270@opencsw.org> <54C24E22.7040005@opencsw.org> <54C2D605.1020808@opencsw.org> Message-ID: Riccardo Mottola escreveu no dia Fri Jan 23 2015 at 23:15:36: > And what about these warnigs? > CHECKPKG_OVERRIDES_CSWgnustep-base += > binary-architecture-does-not-match-placement|file=opt/csw/ > GNUstep/System/Tools/defaults|arch_id=18|arch_name=sparcv8+ > This looks harmless, we don't constrain the lowest ISA to sparcv8 anyway. We could update the check to accept sparcv8+ binaries in random places. -------------- next part -------------- An HTML attachment was scrubbed... URL: From maciej at opencsw.org Sat Jan 24 14:09:14 2015 From: maciej at opencsw.org (=?UTF-8?Q?Maciej_=28Matchek=29_Blizi=C5=84ski?=) Date: Sat, 24 Jan 2015 13:09:14 +0000 Subject: FOSDEM anyone? In-Reply-To: References: <52D9917C.6080607@opencsw.org> Message-ID: I'm not going :( -- I didn't plan early enough and now there are things in my calendar which are hard to move. -------------- next part -------------- An HTML attachment was scrubbed... URL: From rmottola at opencsw.org Sun Jan 25 18:45:23 2015 From: rmottola at opencsw.org (Riccardo Mottola) Date: Sun, 25 Jan 2015 18:45:23 +0100 Subject: CHECKPKG_OVERRIDES trivial question Message-ID: <54C52BB3.5060503@opencsw.org> Hi, when I run "mgar package"and I get override suggestions, is this list always "full" or does it take in account those which remain? In other works, if I have added certain overrides which I consider checked, should I still see them (like it is happening for me) or should I not? Riccardo From dam at opencsw.org Mon Jan 26 08:42:46 2015 From: dam at opencsw.org (Dagobert Michelsen) Date: Mon, 26 Jan 2015 08:42:46 +0100 Subject: EXTRA_RUNPATH_DIRS, LD_OPTIONS an B direct In-Reply-To: <54C2D605.1020808@opencsw.org> References: <549ADEBA.2050608@opencsw.org> <2C4A66BC-EB65-4F7F-B013-21C361147270@opencsw.org> <54C24E22.7040005@opencsw.org> <54C2D605.1020808@opencsw.org> Message-ID: <38200DA6-05A9-43BA-A947-E5DA194C26F6@opencsw.org> Hi Riccardo, > Am 24.01.2015 um 00:15 schrieb Riccardo Mottola : > > Hi, > > Dagobert Michelsen wrote: >> Try ldd -r on gdnc, then libgnustep-base.so.1.24 can not be found. > > I came to the conclusion that for teh standard setup this is fine and should work. Ok. > However, other warnings do worry me: > > * No symbol of binary /opt/csw/GNUstep/System/Tools/autogsdoc is directly > bound against the following libraries: libgcc_s.so.1, libgnustep- > base.so.1.24, libobjc.so.4. Please make sure the binaries are compiled > using the "-Bdirect" linker option. > > I explained libgnustep-base. But libgcc and libobjc are "standard" stuff from opencsw, it should just work? > In fact: > rmottola at unstable10s [unstable10s]:~/opencsw/gnustep-base/trunk > ldd ./work/solaris10-sparc/install-isa-sparcv8plus/opt/csw/GNUstep/System/Tools/autogsdoc > libgnustep-base.so.1.24 => (file not found) > libobjc.so.4 => /opt/csw/lib/libobjc.so.4 > libsocket.so.1 => /lib/libsocket.so.1 > libnsl.so.1 => /lib/libnsl.so.1 > libm.so.2 => /lib/libm.so.2 > libgcc_s.so.1 => /opt/csw/lib/libgcc_s.so.1 > libpthread.so.1 => /lib/libpthread.so.1 > libc.so.1 => /lib/libc.so.1 > libmp.so.2 => /lib/libmp.so.2 > libmd.so.1 => /lib/libmd.so.1 > libscf.so.1 => /lib/libscf.so.1 > libdoor.so.1 => /lib/libdoor.so.1 > libuutil.so.1 => /lib/libuutil.so.1 > libgen.so.1 => /lib/libgen.so.1 > /platform/SUNW,SPARC-Enterprise-T5220/lib/libc_psr.so.1 > /platform/SUNW,SPARC-Enterprise-T5220/lib/libmd_psr.so.1 > > Looks fine except for gnustep-base? Direct Binding is somewhat complicated and helps in bumping library versions. For details please see http://wiki.opencsw.org/checkpkg-error-tags#toc32 and the linked http://wiki.opencsw.org/proposal:direct-binding-and-as-needed-linking-as-a-default Best regards ? Dago -- "You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process." - xkcd #896 -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 2418 bytes Desc: not available URL: From dam at opencsw.org Mon Jan 26 08:53:54 2015 From: dam at opencsw.org (Dagobert Michelsen) Date: Mon, 26 Jan 2015 08:53:54 +0100 Subject: CHECKPKG_OVERRIDES trivial question In-Reply-To: <54C52BB3.5060503@opencsw.org> References: <54C52BB3.5060503@opencsw.org> Message-ID: <90484FBC-F63B-45C5-8F1D-B31EDC8B9ABF@opencsw.org> Hi Riccardo, > Am 25.01.2015 um 18:45 schrieb Riccardo Mottola : > > when I run "mgar package"and I get override suggestions, is this list always "full" or does it take in account those which remain? > > In other works, if I have added certain overrides which I consider checked, should I still see them (like it is happening for me) or should I not? Unfortunately changing overrides requires a ?mgar remerge && mgar repackage?. Best regards ? Dago -- "You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process." - xkcd #896 From claudio at opencsw.org Mon Jan 26 23:17:39 2015 From: claudio at opencsw.org (Claudio) Date: Mon, 26 Jan 2015 23:17:39 +0100 Subject: FOSDEM anyone? In-Reply-To: References: <52D9917C.6080607@opencsw.org> Message-ID: I object!: :) C. On Sat, Jan 24, 2015 at 2:09 PM, Maciej (Matchek) Blizi?ski wrote: > I'm not going :( -- I didn't plan early enough and now there are things in > my calendar which are hard to move. From rmottola at opencsw.org Tue Jan 27 12:56:13 2015 From: rmottola at opencsw.org (Riccardo Mottola) Date: Tue, 27 Jan 2015 12:56:13 +0100 Subject: EXTRA_RUNPATH_DIRS, LD_OPTIONS an B direct In-Reply-To: <38200DA6-05A9-43BA-A947-E5DA194C26F6@opencsw.org> References: <549ADEBA.2050608@opencsw.org> <2C4A66BC-EB65-4F7F-B013-21C361147270@opencsw.org> <54C24E22.7040005@opencsw.org> <54C2D605.1020808@opencsw.org> <38200DA6-05A9-43BA-A947-E5DA194C26F6@opencsw.org> Message-ID: <54C77CDD.3090904@opencsw.org> Hi, Dagobert Michelsen wrote: > Direct Binding is somewhat complicated and helps in bumping library versions. For details > please see > http://wiki.opencsw.org/checkpkg-error-tags#toc32 > and the linked > http://wiki.opencsw.org/proposal:direct-binding-and-as-needed-linking-as-a-default I understand. For internal libraries it is thus not a problem: if I upgrade a tool inside gnustep-base, it will be relinked. However, linking to other packages should follow this rule, if possible: So an upgrade of a dependency will be detected. Do you usually link programs using gcc or the linker? Perhaps this is part of the problem. I made make messages verbose and see this: /opt/csw/bin/gcc-4.9 -mcpu=v9 -Wa,-xarch=v8plus -L/opt/csw/GNUstep/lib -L/opt/csw/lib -shared-libgcc -pthread -fexceptions -fgnu-runtime -o obj/make_strings \ ./obj/make_strings.obj/make_strings.m.o ./obj/make_strings.obj/SourceEntry.m.o ./obj/make_strings.obj/StringsEntry.m.o ./obj/make_strings.obj/StringsFile.m.o \ -L../../Source/./obj -L/home/rmottola/GNUstep/Library/Libraries -L/opt/csw/GNUstep/Local/Library/Libraries -L/opt/csw/GNUstep/System/Library/Libraries -L/opt/csw/GNUstep/Local/Library/Libraries -L/opt/csw/GNUstep/Local/Library/Libraries -L/opt/csw/GNUstep/System/Library/Libraries -L/opt/csw/lib -lgnustep-base -lobjc -lsocket -lnsl -lm Is this usual? How do other package link? Riccardo From rmottola at opencsw.org Sat Jan 31 11:02:21 2015 From: rmottola at opencsw.org (Riccardo Mottola) Date: Sat, 31 Jan 2015 11:02:21 +0100 Subject: EXTRA_RUNPATH_DIRS, LD_OPTIONS an B direct In-Reply-To: <54C77CDD.3090904@opencsw.org> References: <549ADEBA.2050608@opencsw.org> <2C4A66BC-EB65-4F7F-B013-21C361147270@opencsw.org> <54C24E22.7040005@opencsw.org> <54C2D605.1020808@opencsw.org> <38200DA6-05A9-43BA-A947-E5DA194C26F6@opencsw.org> <54C77CDD.3090904@opencsw.org> Message-ID: <54CCA82D.6000708@opencsw.org> Hi, might I ping this question? Riccardo Mottola wrote: > > Do you usually link programs using gcc or the linker? Perhaps this is > part of the problem. I made make messages verbose and see this: > /opt/csw/bin/gcc-4.9 -mcpu=v9 -Wa,-xarch=v8plus > -L/opt/csw/GNUstep/lib -L/opt/csw/lib -shared-libgcc -pthread > -fexceptions -fgnu-runtime -o obj/make_strings \ > ./obj/make_strings.obj/make_strings.m.o > ./obj/make_strings.obj/SourceEntry.m.o > ./obj/make_strings.obj/StringsEntry.m.o > ./obj/make_strings.obj/StringsFile.m.o \ > -L../../Source/./obj > -L/home/rmottola/GNUstep/Library/Libraries > -L/opt/csw/GNUstep/Local/Library/Libraries > -L/opt/csw/GNUstep/System/Library/Libraries > -L/opt/csw/GNUstep/Local/Library/Libraries > -L/opt/csw/GNUstep/Local/Library/Libraries > -L/opt/csw/GNUstep/System/Library/Libraries -L/opt/csw/lib > -lgnustep-base -lobjc -lsocket -lnsl -lm > > > Is this usual? How do other package link? My hypothesis is that gnustep-make leaves LL_OPTIONS untouched, however since "ld" is not invoked directly it doesn't work. Since the compiler itself (gcc) is used to link, there needs an other way to pass it. Ideas? Riccardo From dam at opencsw.org Sat Jan 31 11:23:07 2015 From: dam at opencsw.org (Dagobert Michelsen) Date: Sat, 31 Jan 2015 11:23:07 +0100 Subject: EXTRA_RUNPATH_DIRS, LD_OPTIONS an B direct In-Reply-To: <54CCA82D.6000708@opencsw.org> References: <549ADEBA.2050608@opencsw.org> <2C4A66BC-EB65-4F7F-B013-21C361147270@opencsw.org> <54C24E22.7040005@opencsw.org> <54C2D605.1020808@opencsw.org> <38200DA6-05A9-43BA-A947-E5DA194C26F6@opencsw.org> <54C77CDD.3090904@opencsw.org> <54CCA82D.6000708@opencsw.org> Message-ID: <7F14E6EF-13BD-4F38-9246-9A7832157D14@opencsw.org> Hi Riccardo, > Am 31.01.2015 um 11:02 schrieb Riccardo Mottola : > > Hi, > > might I ping this question? > > Riccardo Mottola wrote: >> >> Do you usually link programs using gcc or the linker? Perhaps this is part of the problem. I made make messages verbose and see this: >> /opt/csw/bin/gcc-4.9 -mcpu=v9 -Wa,-xarch=v8plus -L/opt/csw/GNUstep/lib -L/opt/csw/lib -shared-libgcc -pthread -fexceptions -fgnu-runtime -o obj/make_strings \ >> ./obj/make_strings.obj/make_strings.m.o ./obj/make_strings.obj/SourceEntry.m.o ./obj/make_strings.obj/StringsEntry.m.o ./obj/make_strings.obj/StringsFile.m.o \ >> -L../../Source/./obj -L/home/rmottola/GNUstep/Library/Libraries -L/opt/csw/GNUstep/Local/Library/Libraries -L/opt/csw/GNUstep/System/Library/Libraries -L/opt/csw/GNUstep/Local/Library/Libraries -L/opt/csw/GNUstep/Local/Library/Libraries -L/opt/csw/GNUstep/System/Library/Libraries -L/opt/csw/lib -lgnustep-base -lobjc -lsocket -lnsl -lm >> >> >> Is this usual? How do other package link? > > My hypothesis is that gnustep-make leaves LL_OPTIONS untouched, however since "ld" is not invoked directly it doesn't work. Since the compiler itself (gcc) is used to link, there needs an other way to pass it. No, when you link with gcc also ld is invoked: https://buildfarm.opencsw.org/source/xref/opencsw/csw/mgar/pkg/gcc4/trunk/Makefile#74 I wouldn?t mind too much with direct binding. libssl is important here, the other libs are by far not that relevant for direct binding. Best regards ? Dago -- "You don't become great by trying to be great, you become great by wanting to do something, and then doing it so hard that you become great in the process." - xkcd #896 -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 2418 bytes Desc: not available URL: