Re: [csw-devel] SF.net SVN: gar:[15616] csw/mgar/pkg/subversion/trunk/Makefile
Hi Rupert, 2011/9/11 : > +CHECKPKG_OVERRIDES_CSWap2svn += > license-missing|/opt/csw/share/doc/ap2_subversion/license Why don't you add the license file? Maciej ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[15656] csw/mgar/pkg/tokyocabinet/trunk/Makefile
2011/9/19 : > +CHECKPKG_OVERRIDES_CSWtokyocabinet += > shared-lib-package-contains-so-symlink|file=/opt/csw/lib/sparcv9/libtokyocabinet.so > +CHECKPKG_OVERRIDES_CSWtokyocabinet += > shared-lib-pkgname-mismatch|file=opt/csw/lib/sparcv9/libtokyocabinet.so.9.10.0|soname=libtokyocabinet.so.9|pkgname=CSWtokyocabinet|expected=CSWlibtokyocabinet9 > +CHECKPKG_OVERRIDES_CSWtokyocabinet += > shared-lib-package-contains-so-symlink|file=/opt/csw/lib/sparcv9/libtokyocabinet.so Please don't override these errors. If there is a shared library, it should be packaged separately. checkpkg prints out the GAR code that implements it, you can copy it and paste into the build recipe and it'll work. Maciej ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[15657] csw/mgar/pkg/tokyocabinet/trunk/Makefile
2011/9/19 : > -CHECKPKG_OVERRIDES_CSWtokyocabinet += > shared-lib-package-contains-so-symlink|file=/opt/csw/lib/sparcv9/libtokyocabinet.so > -CHECKPKG_OVERRIDES_CSWtokyocabinet += > shared-lib-pkgname-mismatch|file=opt/csw/lib/sparcv9/libtokyocabinet.so.9.10.0|soname=libtokyocabinet.so.9|pkgname=CSWtokyocabinet|expected=CSWlibtokyocabinet9 > -CHECKPKG_OVERRIDES_CSWtokyocabinet += > shared-lib-package-contains-so-symlink|file=/opt/csw/lib/sparcv9/libtokyocabinet.so > +PACKAGES += CSWlibtokyocabinet9 > +CATALOGNAME_CSWlibtokyocabinet9 = libtokyocabinet9 > +PKGFILES_CSWlibtokyocabinet9 += $(call > baseisadirs,$(libdir),libtokyocabinet\.so\.9\.10\.0) > +PKGFILES_CSWlibtokyocabinet9 += $(call > baseisadirs,$(libdir),libtokyocabinet\.so\.9(\.\d+)*) > +SPKG_DESC_CSWlibtokyocabinet9 += $(DESCRIPTION), libtokyocabinet.so.9 > +RUNTIME_DEP_PKGS_CSWtokyocabinet += CSWlibtokyocabinet9 > +PACKAGES += CSWtokyocabinet-dev > +CATALOGNAME_CSWtokyocabinet-dev = tokyocabinet_dev > +SPKG_DESC_CSWtokyocabinet-dev += $(DESCRIPTION), development files > +PKGFILES_CSWtokyocabinet-dev += /opt/csw/lib/sparcv9/libtokyocabinet.so Cool, as you can see I'm processing all the emails in order. :-) ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[15663] csw/mgar/pkg/squid/branches/squid3
Can we do it better than mass-pasting overrides? Some more comments below. 2011/9/19 : > +CHECKPKG_OVERRIDES_CSW$(NAME) += > file-with-bad-content|/usr/local|root/opt/csw/share/man/man8/squid_unix_group.8 That's okay to override. > +CHECKPKG_OVERRIDES_CSW$(NAME) += pkginfo-opencsw-repository-uncommitted Don't override this one. It will go away when you commit the code to subversion. > +CHECKPKG_OVERRIDES_CSW$(NAME) += bad-rpath-entry|/lib|opt/csw/bin/squidclient > +CHECKPKG_OVERRIDES_CSW$(NAME) += > bad-rpath-entry|/opt/sunstudio12.1/lib|opt/csw/bin/squidclient > +CHECKPKG_OVERRIDES_CSW$(NAME) += > bad-rpath-entry|/opt/sunstudio12.1/lib/rw7|opt/csw/bin/squidclient > +CHECKPKG_OVERRIDES_CSW$(NAME) += > bad-rpath-entry|/lib|opt/csw/libexec/cachemgr.cgi > +CHECKPKG_OVERRIDES_CSW$(NAME) += > bad-rpath-entry|/opt/sunstudio12.1/lib|opt/csw/libexec/cachemgr.cgi > +CHECKPKG_OVERRIDES_CSW$(NAME) += > bad-rpath-entry|/opt/sunstudio12.1/lib/rw7|opt/csw/libexec/cachemgr.cgi > +CHECKPKG_OVERRIDES_CSW$(NAME) += bad-rpath-entry|/lib|opt/csw/libexec/diskd > +CHECKPKG_OVERRIDES_CSW$(NAME) += > bad-rpath-entry|/opt/sunstudio12.1/lib|opt/csw/libexec/diskd > +CHECKPKG_OVERRIDES_CSW$(NAME) += > bad-rpath-entry|/opt/sunstudio12.1/lib/rw7|opt/csw/libexec/diskd > +CHECKPKG_OVERRIDES_CSW$(NAME) += bad-rpath-entry|/lib|opt/csw/libexec/unlinkd > +CHECKPKG_OVERRIDES_CSW$(NAME) += > bad-rpath-entry|/opt/sunstudio12.1/lib|opt/csw/libexec/unlinkd > +CHECKPKG_OVERRIDES_CSW$(NAME) += > bad-rpath-entry|/opt/sunstudio12.1/lib/rw7|opt/csw/libexec/unlinkd > +CHECKPKG_OVERRIDES_CSW$(NAME) += bad-rpath-entry|/lib|opt/csw/sbin/squid > +CHECKPKG_OVERRIDES_CSW$(NAME) += > bad-rpath-entry|/opt/sunstudio12.1/lib|opt/csw/sbin/squid > +CHECKPKG_OVERRIDES_CSW$(NAME) += > bad-rpath-entry|/opt/sunstudio12.1/lib/rw7|opt/csw/sbin/squid These can be easily fixed, see: http://wiki.opencsw.org/checkpkg-error-tags#toc8 > +CHECKPKG_OVERRIDES_CSW$(NAME) += unidentified-dependency|CSWcas-initsmf > +CHECKPKG_OVERRIDES_CSW$(NAME) += unidentified-dependency|CSWcas-migrateconf > +CHECKPKG_OVERRIDES_CSW$(NAME) += unidentified-dependency|CSWcas-preserveconf > +CHECKPKG_OVERRIDES_CSW$(NAME) += unidentified-dependency|CSWcommon > +CHECKPKG_OVERRIDES_CSW$(NAME) += unidentified-dependency|CSWoldaprt > +CHECKPKG_OVERRIDES_CSW$(NAME) += unidentified-dependency|CSWosslrt > +CHECKPKG_OVERRIDES_CSW$(NAME) += unidentified-dependency|CSWsasl > +CHECKPKG_OVERRIDES_CSW$(NAME) += > soname-not-found|libsasl2.so.2|is|needed|by|opt/csw/libexec/sasl_auth > +CHECKPKG_OVERRIDES_CSW$(NAME) += > soname-not-found|liblber-2.4.so.2|is|needed|by|opt/csw/libexec/squid_ldap_group > +CHECKPKG_OVERRIDES_CSW$(NAME) += > soname-not-found|libcrypto.so.0.9.8|is|needed|by|opt/csw/libexec/ncsa_auth > +CHECKPKG_OVERRIDES_CSW$(NAME) += > soname-not-found|libssl.so.0.9.8|is|needed|by|opt/csw/libexec/diskd > +CHECKPKG_OVERRIDES_CSW$(NAME) += > soname-not-found|libssl.so.0.9.8|is|needed|by|opt/csw/sbin/squid > +CHECKPKG_OVERRIDES_CSW$(NAME) += > soname-not-found|libldap-2.4.so.2|is|needed|by|opt/csw/libexec/squid_ldap_group > +CHECKPKG_OVERRIDES_CSW$(NAME) += > soname-not-found|libssl.so.0.9.8|is|needed|by|opt/csw/libexec/unlinkd > +CHECKPKG_OVERRIDES_CSW$(NAME) += > soname-not-found|libssl.so.0.9.8|is|needed|by|opt/csw/libexec/ncsa_auth > +CHECKPKG_OVERRIDES_CSW$(NAME) += > soname-not-found|libcrypto.so.0.9.8|is|needed|by|opt/csw/libexec/diskd > +CHECKPKG_OVERRIDES_CSW$(NAME) += > soname-not-found|libcrypto.so.0.9.8|is|needed|by|opt/csw/libexec/unlinkd > +CHECKPKG_OVERRIDES_CSW$(NAME) += > soname-not-found|libltdl.so.7|is|needed|by|opt/csw/sbin/squid > +CHECKPKG_OVERRIDES_CSW$(NAME) += > soname-not-found|libldap-2.4.so.2|is|needed|by|opt/csw/libexec/squid_ldap_auth > +CHECKPKG_OVERRIDES_CSW$(NAME) += > soname-not-found|libcrypto.so.0.9.8|is|needed|by|opt/csw/sbin/squid > +CHECKPKG_OVERRIDES_CSW$(NAME) += > soname-not-found|liblber-2.4.so.2|is|needed|by|opt/csw/libexec/squid_ldap_auth These indicate that the database doesn't know about the packages and files. Is it from a build outside the buildfarm? If so, I'd be interested in improving off-buildfarm support so that these false positives can go away. > +CHECKPKG_OVERRIDES_CSW$(NAME) += missing-dependency|CSWperl Can you look why checkpkg thinks that CSWperl is necessary? There is an explanation in the checkpkg output. > +CHECKPKG_OVERRIDES_CSW$(NAME) += surplus-dependency|CSWoldaprt > +CHECKPKG_OVERRIDES_CSW$(NAME) += surplus-dependency|CSWsasl > +CHECKPKG_OVERRIDES_CSW$(NAME) += surplus-dependency|CSWosslrt ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: opencsw:[558] utilities/pkgutil-deptree-to-dot.py
2011/9/26 : > Revision: 558 > http://opencsw.svn.sourceforge.net/opencsw/?rev=558&view=rev > Author: skayser > Date: 2011-09-26 15:50:56 + (Mon, 26 Sep 2011) > Log Message: > --- > utilities: add pkgutil deptree to graphviz converter > > Added Paths: > --- > utilities/pkgutil-deptree-to-dot.py In most cases, you want to write unit tests for your programs. To test classes, you need to import them. You cannot import a file with dashes init. Therefore, you always need to create *.py files with underscores. If you want to call the file using a name with dashes, make a symlink. You can see csw-upload-pkg for an example. ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[15805] csw/mgar/pkg/libldns/trunk/Makefile
2011/9/30 : > Revision: 15805 > http://gar.svn.sourceforge.net/gar/?rev=15805&view=rev > Author: idogan23 > Date: 2011-09-30 08:58:33 + (Fri, 30 Sep 2011) > Log Message: > --- > liblnds: fixed overrides > > Modified Paths: > -- > csw/mgar/pkg/libldns/trunk/Makefile > > Modified: csw/mgar/pkg/libldns/trunk/Makefile > === > --- csw/mgar/pkg/libldns/trunk/Makefile 2011-09-29 22:13:48 UTC (rev 15804) > +++ csw/mgar/pkg/libldns/trunk/Makefile 2011-09-30 08:58:33 UTC (rev 15805) > @@ -544,7 +544,7 @@ > CHECKPKG_OVERRIDES_CSWlibldns-devel += > file-collision|/opt/csw/lib/sparcv8plus+vis/libldns.so|CSWlibldns-devel|CSWlibldnsdevel > CHECKPKG_OVERRIDES_CSWlibldns-devel += > file-collision|/opt/csw/share/man/man3/ldns_rr_new_frm_type.3|CSWlibldns-devel|CSWlibldnsdevel > CHECKPKG_OVERRIDES_CSWlibldns-devel += > file-collision|/opt/csw/share/man/man3/ldns_dnssec_derive_trust_tree_normal_rrset.3|CSWlibldns-devel|CSWlibldnsdevel > -#CHECKPKG_OVERRIDES_CSWlibldns-devel += > file-collision|/opt/csw/lib/amd64/libldns.so|CSWlibldns-devel|CSWlibldnsdevel > -#CHECKPKG_OVERRIDES_CSWlibldns-devel += > file-collision|/opt/csw/lib/pentium_pro/libldns.so|CSWlibldns-devel|CSWlibldnsdevel > +CHECKPKG_OVERRIDES_CSWlibldns-devel += > file-collision|/opt/csw/lib/amd64/libldns.so|CSWlibldns-devel|CSWlibldnsdevel > +CHECKPKG_OVERRIDES_CSWlibldns-devel += > file-collision|/opt/csw/lib/pentium_pro/libldns.so|CSWlibldns-devel|CSWlibldnsdevel Is there a reason to keep conflicting files in the repository? This would probably fix the problem: OBSOLETED_BY_CSWlibldns-devel = CSWlibldnsdevel ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[15882] csw/mgar/pkg/sasl/branches/sasl-2.1.23-mGARv2
Why does it depend on libkrb5? ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[15900] csw/mgar/pkg/cmake/trunk/Makefile
Em 09/10/2011 11:40, escreveu: > > Revision: 15900 > http://gar.svn.sourceforge.net/gar/?rev=15900&view=rev > Author: rthurner > Date: 2011-10-09 10:40:09 + (Sun, 09 Oct 2011) > Log Message: > --- > cmake, checkpkg requires still > > Modified Paths: > -- >csw/mgar/pkg/cmake/trunk/Makefile > > Modified: csw/mgar/pkg/cmake/trunk/Makefile > === > --- csw/mgar/pkg/cmake/trunk/Makefile 2011-10-09 10:30:06 UTC (rev 15899) > +++ csw/mgar/pkg/cmake/trunk/Makefile 2011-10-09 10:40:09 UTC (rev 15900) > @@ -24,8 +24,8 @@ > CATALOGNAME_CSWcmake = cmake > SPKG_DESC_CSWcmake = $(DESCRIPTION) > > -RUNTIME_DEP_PKGS_CSWcmake = CSWcurlrt CSWexpat > -RUNTIME_DEP_PKGS_CSWcmake += CSWzlib > +RUNTIME_DEP_PKGS_CSWcmake = CSWlibcurl4 CSWlibexpat1 > +RUNTIME_DEP_PKGS_CSWcmake += CSWlibz1 > > > # We define upstream file regex so we can be notifed of > @@ -42,6 +42,81 @@ > # switch off tests as two of them fail > TEST_TARGET = > > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/bin/cmake Do you know that you can write one line instead? CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/bin/ctest > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/bin/ccmake > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/bin/cpack > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/share/cmake-2.8.6/Modules/UseJava.cmake > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/share/cmake-2.8.6/Modules/FindSDL_image.cmake > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/share/cmake-2.8.6/Modules/FindSDL_sound.cmake > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/share/cmake-2.8.6/Modules/FindSDL_ttf.cmake > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/share/cmake-2.8.6/Modules/CMakeGenericSystem.cmake > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/share/cmake-2.8.6/Modules/GetPrerequisites.cmake > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/share/cmake-2.8.6/Modules/FindLAPACK.cmake > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/share/cmake-2.8.6/Modules/FindSDL_net.cmake > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/share/cmake-2.8.6/Modules/FindAVIFile.cmake > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/share/cmake-2.8.6/Modules/FindJava.cmake > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/share/cmake-2.8.6/Modules/FindProducer.cmake > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/share/cmake-2.8.6/Modules/FindFLTK.cmake > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/share/cmake-2.8.6/Modules/CMakeDetermineJavaCompiler.cmake > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/share/cmake-2.8.6/Modules/FindSelfPackers.cmake > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/share/cmake-2.8.6/Modules/FindPNG.cmake > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/share/cmake-2.8.6/Modules/FindPostgreSQL.cmake > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/share/cmake-2.8.6/Modules/FindFreetype.cmake > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/share/cmake-2.8.6/Modules/FindUnixCommands.cmake > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/share/cmake-2.8.6/Modules/FindSDL.cmake > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/share/cmake-2.8.6/Modules/FindLua50.cmake > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/share/cmake-2.8.6/Modules/FindTCL.cmake > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/share/cmake-2.8.6/Modules/FindMatlab.cmake > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/share/cmake-2.8.6/Modules/FindPhysFS.cmake > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/share/cmake-2.8.6/Modules/FindQt3.cmake > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/share/cmake-2.8.6/Modules/FindPike.cmake > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/share/cmake-2.8.6/Modules/FindwxWidgets.cmake > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/share/cmake-2.8.6/Modules/FindOpenAL.cmake > +CHECKPKG_OVERRIDES_CSWcmake += file-with-bad-content|/usr/local|root/opt/csw/share/cmake-2.8.6/Modules/FindLua51.cmake > +CHECKPKG
Re: [csw-devel] SF.net SVN: gar:[15903] csw/mgar/gar/v2/gar.pkg.mk
2011/10/10 :
> Revision: 15903
> http://gar.svn.sourceforge.net/gar/?rev=15903&view=rev
> Author: skayser
> Date: 2011-10-10 09:07:57 + (Mon, 10 Oct 2011)
> Log Message:
> ---
> gar/v2: fix CATALOG_RELEASE regex
>
> Modified Paths:
> --
> csw/mgar/gar/v2/gar.pkg.mk
>
> Modified: csw/mgar/gar/v2/gar.pkg.mk
> ===
> --- csw/mgar/gar/v2/gar.pkg.mk 2011-10-10 07:44:44 UTC (rev 15902)
> +++ csw/mgar/gar/v2/gar.pkg.mk 2011-10-10 09:07:57 UTC (rev 15903)
> @@ -62,7 +62,7 @@
> # set CATALOG_RELEASE to 'current'. Used by checkpkg to query the right
> branch
> # in the buildfarm pkgdb. For off-site usage this can default to unstable.
> HOSTNAME := $(shell hostname)
> -CATALOG_RELEASE ?= $(shell echo $(HOSTNAME) | gsed -e 's/[0-9][sx]*$$//')
> +CATALOG_RELEASE ?= $(shell echo $(HOSTNAME) | gsed -re
> 's/[0-9]{1,2}[sx]$$//')
Oh, the missing count spec after [0-9], sorry for not catching that!
Maciej
___
devel mailing list
devel@lists.opencsw.org
https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[15913] csw/mgar/pkg/lang-python/python3/trunk
2011/10/16 : > Revision: 15913 > http://gar.svn.sourceforge.net/gar/?rev=15913&view=rev > Author: rthurner > Date: 2011-10-16 11:14:52 + (Sun, 16 Oct 2011) > Log Message: > --- > python3: upgrade to python-3.2.2 Hi Rupert, In case you're interested, Python 3.1 packages are available from unstable. The build recipe is here: http://sourceforge.net/apps/trac/gar/browser/csw/mgar/pkg/lang-python/python/branches/python-3.1 Maciej ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[15916] csw/mgar/pkg/lang-python/python3/trunk/Makefile
Hi Rupert, I've commented on this commit below in this email. 2011/10/16 : > Revision: 15916 > http://gar.svn.sourceforge.net/gar/?rev=15916&view=rev > Author: rthurner > Date: 2011-10-16 15:48:41 + (Sun, 16 Oct 2011) > Log Message: > --- > python3, try to get rid of checkpkg errors ... > > Modified Paths: > -- > csw/mgar/pkg/lang-python/python3/trunk/Makefile > > Modified: csw/mgar/pkg/lang-python/python3/trunk/Makefile > === > --- csw/mgar/pkg/lang-python/python3/trunk/Makefile 2011-10-16 12:19:05 > UTC (rev 15915) > +++ csw/mgar/pkg/lang-python/python3/trunk/Makefile 2011-10-16 15:48:41 > UTC (rev 15916) > @@ -98,7 +98,6 @@ > SPKG_DESC_CSWlibpython$(DVER)-1-0 = Python shared library ($(VER)) > > INCOMPATIBLE_PKGS_CSWpython$(CVER) = CSWpydistutils > -OBSOLETED_BY_CSWpython$(CVER)-dev = CSWpython$(CVER)-devel > > PKGFILES_CSWidle$(CVER) = $(libdir)/.*/idlelib/.* > PKGFILES_CSWidle$(CVER) += $(bindir)/idle > @@ -126,18 +125,15 @@ > > RUNTIME_DEP_PKGS_CSWidle$(CVER) += CSWpython$(CVER) > RUNTIME_DEP_PKGS_CSWidle$(CVER) += CSWpython$(CVER)-tk > -RUNTIME_DEP_PKGS_CSWpython$(CVER) += CSWbdb47 > -RUNTIME_DEP_PKGS_CSWpython$(CVER) += CSWbzip2 > +RUNTIME_DEP_PKGS_CSWpython$(CVER) += CSWlibbz2-1-0 > RUNTIME_DEP_PKGS_CSWpython$(CVER) += CSWlibintl8 > -RUNTIME_DEP_PKGS_CSWpython$(CVER) += CSWlibffi5 > RUNTIME_DEP_PKGS_CSWpython$(CVER) += CSWlibgdbm3 > -RUNTIME_DEP_PKGS_CSWpython$(CVER) += CSWlibpython$(DVER)-1-0 > -RUNTIME_DEP_PKGS_CSWpython$(CVER) += CSWlibncursesw5 > +RUNTIME_DEP_PKGS_CSWpython$(CVER) += CSWlibncurses5 > RUNTIME_DEP_PKGS_CSWpython$(CVER) += CSWlibpanel5 > RUNTIME_DEP_PKGS_CSWpython$(CVER) += CSWosslrt > -RUNTIME_DEP_PKGS_CSWpython$(CVER) += CSWreadline > -RUNTIME_DEP_PKGS_CSWpython$(CVER) += CSWsqlite3rt > -RUNTIME_DEP_PKGS_CSWpython$(CVER) += CSWzlib > +RUNTIME_DEP_PKGS_CSWpython$(CVER) += CSWlibreadline6 > +RUNTIME_DEP_PKGS_CSWpython$(CVER) += CSWlibsqlite3-0 > +RUNTIME_DEP_PKGS_CSWpython$(CVER) += CSWlibz1 > RUNTIME_DEP_PKGS_CSWpython$(CVER)-dev += CSWlibpython$(DVER)-1-0 > RUNTIME_DEP_PKGS_CSWpython$(CVER)-dev += CSWpython$(CVER) > RUNTIME_DEP_PKGS_CSWpython$(CVER)-test += CSWpython$(CVER) > @@ -150,15 +146,12 @@ > > # There are packages depending on libpython2.5, and listing CSWpython as the > # dependency. > -RUNTIME_DEP_PKGS_CSWpython += CSWlibpython2-5-1-0 > -CHECKPKG_OVERRIDES_CSWpython += surplus-dependency|CSWlibpython2-5-1-0 > +RUNTIME_DEP_PKGS_CSWpython += CSWlibpython3-2-1-0 > +CHECKPKG_OVERRIDES_CSWpython += surplus-dependency|CSWlibpython3-2-1-0 > > # This could be handled by automatic dependency detection > CHECKPKG_OVERRIDES_CSWidle$(CVER) += surplus-dependency|CSWpython$(CVER)-tk > > -# This package is not meant to be released > -CHECKPKG_OVERRIDES_CSWpython-test += file-with-bad-content > - > CATALOGNAME_CSWidle$(CVER) = idle > CATALOGNAME_CSWpython$(CVER) = python > CATALOGNAME_CSWpython$(CVER)-dev = python$(CVER)_dev > @@ -176,15 +169,43 @@ > # This is an exception, since these are original Python packages. > CHECKPKG_OVERRIDES_CSWidle$(CVER) += > pkgname-does-not-start-with-CSWpy- > CHECKPKG_OVERRIDES_CSWidle$(CVER) += > catalogname-does-not-start-with-py_ > -CHECKPKG_OVERRIDES_CSWpython$(CVER)-test += > pkgname-does-not-start-with-CSWpy- > -CHECKPKG_OVERRIDES_CSWpython$(CVER)-test += > catalogname-does-not-start-with-py_ > -CHECKPKG_OVERRIDES_CSWpython$(CVER)-dev += pkgname-does-not-start-with-CSWpy- > -CHECKPKG_OVERRIDES_CSWpython$(CVER)-dev += > catalogname-does-not-start-with-py_ > +#CHECKPKG_OVERRIDES_CSWpython$(CVER)-test += > pkgname-does-not-start-with-CSWpy- > +#CHECKPKG_OVERRIDES_CSWpython$(CVER)-test += > catalogname-does-not-start-with-py_ > +#CHECKPKG_OVERRIDES_CSWpython$(CVER)-dev += > pkgname-does-not-start-with-CSWpy- > +#CHECKPKG_OVERRIDES_CSWpython$(CVER)-dev += > catalogname-does-not-start-with-py_ > CHECKPKG_OVERRIDES_CSWpython$(CVER)-tk += > pkgname-does-not-start-with-CSWpy- > CHECKPKG_OVERRIDES_CSWpython$(CVER)-tk += > catalogname-does-not-start-with-py_ > CHECKPKG_OVERRIDES_CSWpython$(CVER) += pkgname-does-not-start-with-CSWpy- > CHECKPKG_OVERRIDES_CSWpython$(CVER) += catalogname-does-not-start-with-py_ > +CHECKPKG_OVERRIDES_CSWpython += > file-collision|/opt/csw/lib/pkgconfig/python3.pc|CSWpython|CSWpython31-dev > +CHECKPKG_OVERRIDES_CSWpython += > file-collision|/opt/csw/bin/python3|CSWpython|CSWpython31 > > +CHECKPKG_OVERRIDES_CSWpython-tk += surplus-dependency|CSWlibpython3-2-1-0 > +CHECKPKG_OVERRIDES_CSWpython-tk += surplus-dependency|CSWpython > +CHECKPKG_OVERRIDES_CSWidle += > file-with-bad-content|/usr/share|root/opt/csw/lib/python3.2/idlelib/EditorWindow.py > +CHECKPKG_OVERRIDES_CSWidle += > file-with-bad-content|/usr/share|root/opt/csw/lib/python3.2/idlelib/config-main.def > +CHECKPKG_OVERRIDES_CSW
Re: [csw-devel] SF.net SVN: opencsw:[568] web/wordpresscsw/wp-content/themes/opencsw/ packages.php
2011/10/21 Sebastian Kayser : > This is an example for cleanup work mentioned at summercamp. Okay if I > push this? Looks good overall. A comment about the pushes to current: There are some pushes. current is a symlink to dublin, and dublin does get occasional updates. They are done by me, manually. I plan a bigger push when I sort out the Ada frontend to GCC (timeline unknown). Maciej ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[15979] csw/mgar/pkg/postfix/trunk
2011/10/24 : > Log Message: > --- > Add svn:keywords, just for sanity. This log message doesn't say, which package has been updated. You can either say type it in manually, or say "mgar ci -m ...". mgar can add the recipe path automatically. ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[15981] csw/mgar/pkg/lighttpd/trunk/Makefile
2011/10/25 : > CONFIGURE_ARGS = $(DIRPATHS) > > +CONFIGURE_ARGS += --with-openssl=$(prefix) Does it actually work with the 64-bit build? ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[16002] csw/mgar/pkg/libmemcached/trunk/Makefile
2011/10/26 : > Revision: 16002 > http://gar.svn.sourceforge.net/gar/?rev=16002&view=rev > Author: wilbury > Date: 2011-10-26 16:22:22 + (Wed, 26 Oct 2011) > Log Message: > --- > libmemcached/trunk: Build both SPARC and i386 platforms for release purposes. > > Modified Paths: > -- > csw/mgar/pkg/libmemcached/trunk/Makefile > > Modified: csw/mgar/pkg/libmemcached/trunk/Makefile > === > --- csw/mgar/pkg/libmemcached/trunk/Makefile 2011-10-26 16:14:51 UTC (rev > 16001) > +++ csw/mgar/pkg/libmemcached/trunk/Makefile 2011-10-26 16:22:22 UTC (rev > 16002) > @@ -32,7 +32,7 @@ > GARCOMPILER = GCC4 GARCOMPILER = GNU The GNU value is better, because it'll progress to the future versions of GCC instead of sticking to one. ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[16012] csw/mgar/pkg/subversion/trunk/Makefile
2011/10/26 : > +CHECKPKG_OVERRIDES_CSWsvn += > discouraged-path-in-pkgmap|/opt/csw/lib/svn/python/site-packages/libsvn/__init__.pyc Hi Rupert, These errors aren't about the location, they are about the file type: *.pyc. You don't need to include the *.pyc files in the package. They are best excluded. You can exclude them by setting the EXTRA_MERGE_EXCLUDE_FILES to for example .*\.pyc Maciej ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[16010] csw/mgar/pkg/subversion/trunk/Makefile
2011/10/26 : > - @$(INSTALL_ENV) gmake -C $(WORKSRC) install-swig-py \ > - swig_pydir=$(libdir)/python2.3/libsvn \ > - swig_pydir_extra=$(libdir)/python2.3/svn > + @$(INSTALL_ENV) gmake -C $(WORKSRC) install-swig-py Maybe instead of removing the settings entirely, you could hint the svn bindings where the python library directory is? It's in /opt/csw/lib/python/site-packages, and the files would be placed under /opt/csw/lib/python/site-packages/svn. Something like: swig_pydir_extra=$(libdir)/python/site-packages Maciej ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[16011] csw/mgar/pkg/libmemcached/trunk/Makefile
2011/10/26 : > +GARTYPE = 2 I think this needs to match the gar branch string, e.g. "v2". ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[16010] csw/mgar/pkg/subversion/trunk/Makefile
2011/10/27 Dagobert Michelsen : >> swig_pydir_extra=$(libdir)/python/site-packages > > It is correct by default, so no need for hinting :-) Are you sure Dago? Also, what do you mean by correct? For example, our /opt/csw/lib/python/site-packages path is wrong, it should be /opt/csw/lib/python2.6/site-packages. Subversion could put files into /opt/csw/lib/python/site-packages, which would be right given our wrong setup, but wrong according to the Python installation standard. I've seen that there are overrides for *.pyc files pointing at wrong paths. Maciej ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[16010] csw/mgar/pkg/subversion/trunk/Makefile
2011/10/28 Dagobert Michelsen : > Hi Maciej, > > Am 27.10.2011 um 21:26 schrieb Maciej (Matchek) Bliziński: >> 2011/10/27 Dagobert Michelsen : >>>> swig_pydir_extra=$(libdir)/python/site-packages >>> >>> It is correct by default, so no need for hinting :-) >> >> Are you sure Dago? Also, what do you mean by correct? For example, >> our /opt/csw/lib/python/site-packages path is wrong, it should be >> /opt/csw/lib/python2.6/site-packages. > > Than it is wrong, I saw the previous path. The way to go is to move > the args to build stage instead of install stage. Sounds good. >> Subversion could put files into >> /opt/csw/lib/python/site-packages, which would be right given our >> wrong setup, but wrong according to the Python installation standard. >> >> I've seen that there are overrides for *.pyc files pointing at wrong paths. > > But we have lots of packages with that, right? I suggest sparing this for > "kiel" and just release svn 1.7.1 with the current pathes. I was planning to use versioned library directory for Python 2.7 and/or Python 3.1/3.2. Maciej ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[16010] csw/mgar/pkg/subversion/trunk/Makefile
2011/10/29 rupert THURNER : > i am not fully sure how you plan to implement this with gar, could you have > a look please? for the moment i just dropped the concerning lines, and it > built. i was also not sure about a quick way to test that the python > bindings. maybe try to clone a svn repository using (python based) mercurial > ... hmm. The first and simplest way is look at the paths. Here's an old pythonsvn package and its files: http://www.opencsw.org/search/pythonsvn/ If you get the same paths in the new package, it's quite likely good. You can then release it into unstable and ask the original bug reporter to test it. Maciej ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[16032] csw/mgar/gar
2011/10/28 : > Revision: 16032 > http://gar.svn.sourceforge.net/gar/?rev=16032&view=rev > Author: dmichelsen > Date: 2011-10-28 12:29:35 + (Fri, 28 Oct 2011) > Log Message: > --- > Make branch v2-build64only > > Modified Paths: > -- > csw/mgar/gar/v2-build64only/gar.conf.mk > > Added Paths: > --- > csw/mgar/gar/v2-build64only/ > csw/mgar/gar/v2-build64only/lib/python/package_checks.py > > Removed Paths: > - > csw/mgar/gar/v2-build64only/lib/python/package_checks.py Why does this change remove and add this file? It should not touch package_checks at all. Maciej ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[16047] csw/mgar/pkg/php5/trunk/Makefile
2011/10/30 : > Revision: 16047 > http://gar.svn.sourceforge.net/gar/?rev=16047&view=rev > Author: wilbury > Date: 2011-10-30 21:00:55 + (Sun, 30 Oct 2011) > Log Message: > --- > php5/trunk: Fix build deps. > > Modified Paths: > -- > csw/mgar/pkg/php5/trunk/Makefile > > Modified: csw/mgar/pkg/php5/trunk/Makefile > === > --- csw/mgar/pkg/php5/trunk/Makefile 2011-10-30 18:03:32 UTC (rev 16046) > +++ csw/mgar/pkg/php5/trunk/Makefile 2011-10-30 21:00:55 UTC (rev 16047) > @@ -53,7 +53,7 @@ > > > BUILD_DEP_PKGS += CSWlibgdbm-dev CSWlibidn-dev CSWlibgd-dev CSWcoreutils > -BUILD_DEP_PKGS += CSWlibxml2-dev CSWlibexpat1 CSWlibz-dev CSWmysql-dev > CSWjpeg > +BUILD_DEP_PKGS += CSWlibxml2-dev CSWlibexpat1 CSWlibz-dev CSWlibcurl-dev > CSWjpeg Why remove CSWmysql-dev? ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[16261] csw/mgar/pkg/ganglia/trunk/Makefile
2011/11/24 : > Revision: 16261 > http://gar.svn.sourceforge.net/gar/?rev=16261&view=rev > Author: d_pocock > Date: 2011-11-24 17:08:05 + (Thu, 24 Nov 2011) > Log Message: > --- > disable BUILD64 again Please prefix every commit message with the name of the package, e.g. "bash: enabling feature foo". You can say: mgar ci -m "enabling feature foo", and mgar will prefix the message for you automatically. Maciej ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] PostgreSQL
[+devel] This time with the correct address of the devel mailing list. 2011/12/5 Maciej Bliziński : > Rafael Ostertag wrote: >> Hi Maciej >> >> You may want to have a look at the latest postgresql commit. The recipe does >> now basically the following: >> >> * Installs libraries into $(libdir). >> >> * Binaries got into $(libexecdir)/postgresql/XY/ and will be linked to >> $(bindir) using alternatives. >> >> * Some other stuff making parallel installs possible. >> >> This recipe should make it possible to install different PostreSQL versions, >> e.g. 8.3, 8.4 etc, in parallel. >> >> Please let me know what you think about it. > > The recipe looks good overall, good job! I pasted it below for some > detailed comments. > > I included Dago, maybe he'll have some comments about alternatives. > I also Cc'd the devel mailing list, where the reviews usually happen. > >> # Copyright 2009 OpenCSW >> # Distributed under the terms of the GNU General Public License v2 >> # $Id$ >> # >> # Note 1 >> # -- >> # >> # A word of caution: PostgreSQL does use autoconf, but it does *not* use >> # automake. That said, it `misses' many features found in pure >> # autoconf/automake builds, e.g. --program-suffix is not available. This also >> # explains some quirks of this recipe. > > Good clarification. > >> # >> # Note 2 >> # -- >> # >> # The idea is to make it possible to install different major/minor versions >> of >> # PostgreSQL at the same time but without installing them to distinct >> --prefix >> # directories, which in turn makes linking against PostgreSQL easier. This >> has >> # been achieved as follows: >> # >> # * libraries go into $(libdir) >> # >> # * binaries go into $(libexec)/postgres/XY/ and alternatives are used to >> # bring the binaries into $(prefix)/bin >> # >> # Generally, putting libraries in $(libdir) should not pose a problem, but in >> # the case of PostgresSQL, there is also a directory $(libdir)/postgres which >> # is unversioned and cannot be changed by passing arguments to configure. >> This >> # shortcoming can be lifted by adjusting $(pkglibdir) in >> # $(WORKSRC)/src/Makefile.global.in which is what we do. >> # >> # >> # Note 3: Alternatives >> # -- >> # >> # I was unable to get the alternatives system to work with using >> # >> # ALTERNATIVES_CSWpostgresql83-server = server_amd64 server_pentium_pro >> # >> # and then defining those alternatives. /opt/csw/sbin/alternatives choked on >> # the one file generated for server containing two distinct alternatives. So >> I >> # cooked my own soup: >> # >> # 1. Create for server, client, contrib, devel separate files containing the >> # alternatives for the ISAs >> # >> # 2. Adjust the prototype to add include the class 'cswalternatives' for the >> # files in question. >> # >> # The problem appeared to be having two lines in the same file. So my >> approach >> # is to generate one file per line of alternatives. > > Maybe Dago can have a look? > >> # >> # Note 4: New upstream releases >> # -- >> # >> # 1. Adjust BASE_VERSION and/or PATCHLEVEL. >> # >> # 2. Make sure the SONAME patch for >> # $(WORKDIR)/postgresql-x.y.z/src/Makefile.shlib still works >> # >> # 3. Comment out 'post-install-modulated:'. So we can identify changes in the >> # server/client environment. >> # >> # 4. `mgar package` >> # >> # 5. If step 4 succeeded, check the prototype of >> # CSW$(NAME)$(BASE_VERSION_NODOT) for files that have slipped by and put >> # them in the proper place. >> # >> # 6. Uncomment 'post-install-modulated:'. This will shed light on changes in >> # the contrib environment. >> # >> # 7. `mgar spotless package` >> # >> # 8. If step 6 succeeded, check the prototype of >> # CSW$(NAME)$(BASE_VERSION_NODOT) for files that have slipped by and put >> # them in the proper place. >> # >> # 9. Make sure the proper version numbers are used for the library packages. >> >> >> # This BUILD is NOT READY! It needs work before it's usable. >> >> # TODO: Initialize the database cluster during installation if not already >> # there. >> # >> # TODO: Migration plan from 8.3: Installing 8.3 and 8.4 at the same time, and >> # migrating >> # >> # TODO: The isaexec part in GAR doesn't work very well. Let's write a >> minimal >> # example of it. (gmake repackage may produce the right output) > > These are my comments; you can remove them if they're no longer > relevant; or you can merge them with your notes. > >> NAME = postgresql >> BASE_VERSION = 8.3 >> VERSION_NODOT = $(subst .,_,$(BASE_VERSION)) >> BASE_VERSION_NODOT = $(subst .,,$(BASE_VERSION)) >> PATCHLEVEL = 16 >> VERSION = $(BASE_VERSION).$(PATCHLEVEL) >> CATEGORIES = apps >> >> # Useful when making a series of builds on the same day >> # GARFLAVOR ?= DBG >> >> VENDOR_URL = http://www.postgresql.org/ >> >> define BLURB >> endef >> >> # Here we will fetch various variables that are used in this recipe. I >> decided >> # to place those in a separate file fo
Re: [csw-devel] SF.net SVN: gar:[16568] csw/mgar/pkg
2011/12/19 > > Revision: 16568 > http://gar.svn.sourceforge.net/gar/?rev=16568&view=rev > Author: cgrzemba > Date: 2011-12-19 09:55:53 + (Mon, 19 Dec 2011) > Log Message: > --- > svrcore/trunk: Mozilla Svrcore, first build > > Added Paths: > --- > csw/mgar/pkg/svrcore/ > csw/mgar/pkg/svrcore/Makefile > csw/mgar/pkg/svrcore/branches/ > csw/mgar/pkg/svrcore/tags/ > csw/mgar/pkg/svrcore/trunk/ > csw/mgar/pkg/svrcore/trunk/Makefile > csw/mgar/pkg/svrcore/trunk/checksums > csw/mgar/pkg/svrcore/trunk/files/ > > Added: csw/mgar/pkg/svrcore/Makefile > === > --- csw/mgar/pkg/svrcore/Makefile (rev 0) > +++ csw/mgar/pkg/svrcore/Makefile 2011-12-19 09:55:53 UTC (rev 16568) > @@ -0,0 +1,2 @@ > +%: > + $(MAKE) -C trunk $* > > > Property changes on: csw/mgar/pkg/svrcore/trunk > ___ > Added: svn:ignore > + cookies > download > work > > > Added: svn:externals > + gar https://gar.svn.sourceforge.net/svnroot/gar/csw/mgar/gar/v2 > > > Added: csw/mgar/pkg/svrcore/trunk/Makefile > === > --- csw/mgar/pkg/svrcore/trunk/Makefile (rev 0) > +++ csw/mgar/pkg/svrcore/trunk/Makefile 2011-12-19 09:55:53 UTC (rev 16568) > @@ -0,0 +1,72 @@ > +# $Id$ > +# > +NAME = svrcore > +MAJOR_VERSION = 4 > +MIN_VERSION = 0.4 > +VERSION = $(MAJOR_VERSION).$(MIN_VERSION) > +GARTYPE = v2 > +CATEGORIES = lib > + > +DESCRIPTION = Mozilla LDAP SVR Core > +define BLURB > + This is the Mozilla SVR Core needed by LDAP SDK > +endef > + > +MASTER_SITES = > http://ftp.mozilla.org/pub/directory/svrcore/releases/4.0.4/src/ > +DISTFILES = $(NAME)-$(VERSION).tar.bz2 > + > +# File name regex to get notifications about upstream software releases > +# NOTE: Use this only if the automatic regex creation > +# does not work for your package > +# UFILES_REGEX = $(NAME)-(\d+(?:\.\d+)*).tar.gz > + > +# If the url used to check for software update is different of MASTER_SITES, > then > +# uncomment the next line. Otherwise it is set by default to the value of > MASTER_SITES > +# UPSTREAM_MASTER_SITES = You can remove the unused lines above. > +PACKAGES += CSWlibsvrcore0 > +CATALOGNAME_CSWlibsvrcore0 = libsvrcore0 > +SPKG_DESC_CSWlibsvrcore0 += $(DESCRIPTION) > +PKGFILES_CSWlibsvrcore0 = $(call pkgfiles_lib,libsvrcore.so.0) > +RUNTIME_DEP_PKGS_CSWlibsvrcore0 += CSWlibnspr4 > +RUNTIME_DEP_PKGS_CSWlibsvrcore0 += CSWlibnss3 > +RUNTIME_DEP_PKGS_CSWlibsvrcore0 += CSWlibplc4 > +RUNTIME_DEP_PKGS_CSWlibsvrcore0 += CSWlibplds4 > +PACKAGING_PLATFORMS += solaris10-sparc > +PACKAGING_PLATFORMS += solaris10-i386 You can move the PACKAGING_PLATFORMS lines closer to the top of the file; they affect the whole recipe. > +PACKAGES += CSWlibsvrcore0-dev > +CATALOGNAME_CSWlibsvrcore0-dev = libsvrcore0_dev > +SPKG_DESC_CSWlibsvrcore0-dev += $(DESCRIPTION), development files > +PKGFILES_CSWlibsvrcore0-dev = $(PKGFILES_DEVEL) > +PKGFILES_CSWlibsvrcore0-dev += /opt/csw/lib/libsvrcore.so > +PKGFILES_CSWlibsvrcore0-dev += /opt/csw/lib/sparcv9/libsvrcore.so > +ARCHALL_CSWgnutlsdevel = 1 > + > +# Move the include files to include/nspr to prevent polluting the main > +# header file namespace. > +includedir = $(prefix)/include/nspr > + > +CONFIGURE_ARGS = $(DIRPATHS) > +CONFIGURE_ARGS_isa-sparcv9 = --enable-64bit > +CONFIGURE_ARGS_isa-amd64 = --enable-64bit > +CONFIGURE_ARGS += $(CONFIGURE_ARGS_$(MODULATION)) > +CONFIGURE_ARGS_OPT += --enable-optimize="-xO3" > +CONFIGURE_ARGS_OPT += --disable-debug > +CONFIGURE_ARGS_OPT += --enable-strip > +CONFIGURE_ARGS_DBG += --enable-debug > +CONFIGURE_ARGS_DBG += --disable-strip > +CONFIGURE_ARGS_OPT += --with-nss-inc=/opt/csw/include/nss > +CONFIGURE_ARGS_OPT += --with-nss-lib=/opt/csw/lib > +CONFIGURE_ARGS_OPT += --with-nspr-inc=/opt/csw/include/nspr > +CONFIGURE_ARGS_OPT += --with-nspr-lib=/opt/csw/lib Two things: the library switches are probably not dependent on GARFLAVOR; instead of using "/opt/csw/lib", use $(libdir), the same goes for $(includedir). > +CONFIGURE_ARGS += $(CONFIGURE_ARGS_$(GARFLAVOR)) > + > + > +RUNTIME_DEP_PKGS_CSWlibsvrcore0-dev += CSWlibsvrcore0 You can move that to the CSWlibsvrcore0-dev definition. > +BUILD64 = 1 > + > +include gar/category.mk Looks good overall! Good job! ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[16750] csw/mgar/pkg/gcc4/branches/private-lib/
2012/1/12 Peter FELECAN : > What's the idea? The idea is this: if libstdc++.so.6 is in /opt/csw/lib and we automatically add -R/opt/csw/lib/$ISALIST to all binaries, there's a new failure mode: a binary might link against a SONAME traditionally from /usr/lib. If / when we introduce the an incompatible library with the same SONAME to /opt/csw/lib, we break that binary. State 1: /opt/csw/bin/foo NEEDED libbar.so.1, RPATH /opt/csw/lib /usr/lib/libbar.so.1 is available State 2: /opt/csw/bin/foo NEEDED libbar.so.1, RPATH /opt/csw/lib /usr/lib/libbar.so.1 is available /opt/csw/lib/libbar.so.1 is available too, incompatible with /opt/csw/bin/foo I'm not sure if this is a scenario we have to worry about, because the solution is to keep GCC libraries in a subdirectory, which has its own disadvantages - library space fragmentation, sometimes harder to debug linking problems, etc. Maciej ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[16773] csw/mgar/pkg/nspr/trunk
2012/1/16 : > Revision: 16773 > http://gar.svn.sourceforge.net/gar/?rev=16773&view=rev > Author: cgrzemba > Date: 2012-01-16 10:14:31 + (Mon, 16 Jan 2012) > Log Message: > --- > nspr/trunk: update version 4.8.9 > > Modified Paths: > -- > csw/mgar/pkg/nspr/trunk/Makefile > csw/mgar/pkg/nspr/trunk/checksums > > Modified: csw/mgar/pkg/nspr/trunk/Makefile > === > --- csw/mgar/pkg/nspr/trunk/Makefile 2012-01-16 09:10:42 UTC (rev 16772) > +++ csw/mgar/pkg/nspr/trunk/Makefile 2012-01-16 10:14:31 UTC (rev 16773) > @@ -8,7 +8,7 @@ > > NAME = nspr > MAJOR_VERSION = 4 > -MIN_VERSION = 8.7 > +MIN_VERSION = 8.9 > VERSION = $(MAJOR_VERSION).$(MIN_VERSION) > CATEGORIES = lib > > @@ -18,11 +18,13 @@ > many of Red Hat's, Sun's, and other software offerings. > endef > > -MASTER_SITES = > https://ftp.mozilla.org/pub/mozilla.org/nspr/releases/v$(VERSION)/src/ > +MASTER_SITES = > http://ftp.mozilla.org/pub/mozilla.org/nspr/releases/v$(VERSION)/src/ Why this change? > DISTFILES = $(NAME)-$(VERSION).tar.gz > DISTFILES += LICENSE > PATCHFILES = nspr-4.6.1-config.patch > > +PACKAGING_PLATFORMS = solaris10-sparc solaris10-i386 > + This should compile on Solaris 9, is there any trouble with that? > PACKAGES = CSWnspr-dev > CATALOGNAME_CSWnspr-dev = nspr_dev > SPKG_DESC_CSWnspr-dev = Netscape Portable Runtime, header files > @@ -104,6 +106,9 @@ > # # three libraries to be in one package. > # CHECKPKG_OVERRIDES_CSWlibnspr4 += shared-lib-pkgname-mismatch > > +CPPFLAGS = > +PREFIX = $(prefix) > + Did you mean to export these variables? > include gar/category.mk > > configure-nspr: > @@ -135,8 +140,7 @@ > ginstall -m 755 $(WORKSRC)/build/config/nspr-config \ > $(DESTDIR)$(bindir) > ginstall -m 755 -d $(DESTDIR)$(libdir)/pkgconfig > - ginstall -m 644 $(WORKSRC)/build/config/nspr.pc \ > - $(DESTDIR)$(libdir)/pkgconfig > + ginstall -m 644 $(WORKSRC)/build/config/nspr.pc > $(DESTDIR)$(libdir)/pkgconfig Please try to keep lines at up to 80 columns wide. > if [ $(MEMORYMODEL) = 64 ]; then \ > gsed -i -e 's+/lib$$+lib/64+' > $(DESTDIR)$(libdir)/pkgconfig/nspr.pc ; \ > fi > > Modified: csw/mgar/pkg/nspr/trunk/checksums > === > --- csw/mgar/pkg/nspr/trunk/checksums 2012-01-16 09:10:42 UTC (rev 16772) > +++ csw/mgar/pkg/nspr/trunk/checksums 2012-01-16 10:14:31 UTC (rev 16773) > @@ -1 +1 @@ > -97e30989a56ab813453b71261849c200 nspr-4.8.7.tar.gz > +60770d45dc08c0f181b22cdfce5be3e8 nspr-4.8.9.tar.gz > > This was sent by the SourceForge.net collaborative development platform, the > world's largest Open Source development site. > > ___ > devel mailing list > devel@lists.opencsw.org > https://lists.opencsw.org/mailman/listinfo/devel ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[16750] csw/mgar/pkg/gcc4/branches/private-lib/
2012/1/15 Peter FELECAN : >> /opt/csw/bin/foo NEEDED libbar.so.1, RPATH /opt/csw/lib >> /usr/lib/libbar.so.1 is available >> /opt/csw/lib/libbar.so.1 is available too, incompatible with /opt/csw/bin/foo > > Is a question of order, isn't it? Consequently, if /opt/csw/lib is > before /usr/lib, everything is dandy. If /opt/csw/lib is first, then /opt/csw/lib/libbar.so.1 will be loaded first, and /opt/csw/bin/foo will crash (it was linked against /usr/lib/libbar.so.1). >> I'm not sure if this is a scenario we have to worry about, because the >> solution is to keep GCC libraries in a subdirectory, which has its own >> disadvantages - library space fragmentation, sometimes harder to debug >> linking problems, etc. > > The 3rd branch as the 4th, in the beginning, when I was the maintainer, > had this strategy and, as far as I remember, we hadn't an issue with. Am > I wrong? My latest thoughts are that the above failure scenario is possible, but is quite contrived and unlikely. It also assumes that we need to support old binaries indefinitely. I understand that Solaris has the backward binary compatibility policy, but this is not one our goals, if I'm not mistaken. If we really run into a linking problem, we'll rebuild the failing binaries. If we can't rebuild a package, then library location is not the worst of our troubles. So, I'm leaning towards keeping all libraries in /opt/csw/lib, including libgcc_s and libstdc++. Maciej ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[16816] csw/mgar/pkg/nss/trunk/Makefile
2012/1/19 : > Revision: 16816 > http://gar.svn.sourceforge.net/gar/?rev=16816&view=rev > Author: cgrzemba > Date: 2012-01-19 13:12:44 + (Thu, 19 Jan 2012) > Log Message: > --- > nss/trunk: knows checkpkg what it wants? > > Modified Paths: > -- > csw/mgar/pkg/nss/trunk/Makefile > > Modified: csw/mgar/pkg/nss/trunk/Makefile > === > --- csw/mgar/pkg/nss/trunk/Makefile 2012-01-19 13:01:39 UTC (rev 16815) > +++ csw/mgar/pkg/nss/trunk/Makefile 2012-01-19 13:12:44 UTC (rev 16816) > @@ -250,6 +250,7 @@ > LD_OPTIONS = > # NSS_LD_OPTIONS = -R'$$ORIGIN' > > +# CHECKPKG_OVERRIDES_CSWlibfreebl3 += > shared-lib-pkgname-mismatch|sonames=libfreebl_32fpu_3.so,libfreebl_32int64_3.so,libfreebl_32int_3.so,libfreebl_64fpu_3.so,libfreebl_64int_3.so|pkgname=CSWlibfreebl3|expected=CSWlibfreebl| Make sure to run "gmake remerge" after you modify overrides. Otherwise override changes will not be taken into account. Maciej ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[16942] csw/mgar/pkg/gettext/trunk/Makefile
2012/1/31 Dagobert Michelsen : > ggettest? Really?? ;-) It's ggettest on Soalris! ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[17064] csw/mgar/pkg/mysql5/branches/mysql-6.0/
2012/2/16 Peter FELECAN : >> csw/mgar/pkg/mysql5/branches/mysql-6.0/ > > in the end csw/mgar/pkg/mysql6 isn't it? Yes, but it's still at alpha, and we're nowhere near releasing it, so it's more like a free fire area for me to experiment. If/when it builds and works, we'll move to to pkg/mysql6. ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] [vlc-devel] [PATCH 2/2] Work around the lack of dirfd() (corrected)
2012/2/20 Brad : > - Original message - >> Detect whether the system supports dirfd(). There was a Solaris-specific >> workaround in one of the files, but it's a better idea to test for the >> function and not rely on the OS detection. > > VLC alreay does this. From the looks of it just remove the redundant Solaris > ifdef'd function. I don't see the compat/dirfd.c file in the tree at all. I was building from the source referenced from http://wiki.videolan.org/Solaris - http://repo.or.cz/w/vlc/solaris.git - my goal is to build VLC on Solaris 10. It seems like the two trees have diverged significantly. I'm not sure which tree should I base my changes on. Any suggestions? Maciej ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] [vlc-devel] [PATCH 1/2] Bootstrap on Solaris
2012/2/27 Rémi Denis-Courmont > So 'if ! test COND' fails. Shouldn't 'if test ! COND' work in any POSIX shell? That seems to work. > cat test.sh #!/bin/sh if test ! -z "foo"; then echo True else echo False fi > ./test.sh True ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] [vlc-devel] [PATCH 2/2] Work around the lack of dirfd()
2012/2/27 Rémi Denis-Courmont : > IIRC this problem was already fixed earlier. Are you using up-to-date vlc.git? Apparently not. I was building from the Solaris branch[1] which seems to be behind master. Maybe I should build from master, so I can actually get any fixes I make back into the source tree? But on the other hand, there's already work that has been put into the Solaris tree, so I'd be probably duplicating it. I'm not sure what's the best way to go about it. My current build error is this: http://buildfarm.opencsw.org/~maciej/vlc-2012-02-28-00-51.log Maciej [1] git://repo.or.cz/vlc/solaris.git ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[17391] csw/mgar/pkg/postfix/trunk/Makefile
2012/3/14 Peter Bonivart : > On Tue, Mar 13, 2012 at 9:06 AM, wrote: >> Revision: 17391 >> http://gar.svn.sourceforge.net/gar/?rev=17391&view=rev >> Author: wahwah >> Date: 2012-03-13 08:06:08 + (Tue, 13 Mar 2012) >> Log Message: >> --- >> postfix/trunk: switch to bdb48, update deps > > Didn't you get collisions with Sendmail? I got collisions with Postfix > when I updated Sendmail and implemented alternatives, I also contacted > Sebastian about doing the same in Postfix. No, I didn't get collisions. Maybe it's because the files have been removed from sendmail. > # Some files collide with CSWpostfix > ALTERNATIVES_CSWsendmail = sendmail > ALTERNATIVE_sendmail = $(bindir)/mailq sendmail $(bindir)/mailq.sendmail 100 > ALTERNATIVE_sendmail += $(bindir)/newaliases newaliases > $(bindir)/newaliases.sendmail > ALTERNATIVE_sendmail += $(mandir)/man1/mailq.1 mailq.1 > $(mandir)/man1/mailq1.1.sendmail > ALTERNATIVE_sendmail += $(mandir)/man1/newaliases.1 newaliases.1 > $(mandir)/man1/newaliases.1.sendmail > ALTERNATIVE_sendmail += $(mandir)/man5/aliases.5 aliases.5 > $(mandir)/man5/aliases.5.sendmail If you know how, just make the change to postfix and push it out. ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[17418] csw/mgar/pkg/postfix/trunk/Makefile
2012/3/15 : > Revision: 17418 > http://gar.svn.sourceforge.net/gar/?rev=17418&view=rev > Author: wilbury > Date: 2012-03-15 12:10:46 + (Thu, 15 Mar 2012) > Log Message: > --- > postfix/trunk: Update PACKAGING_PLATFORMS > > Modified Paths: > -- > csw/mgar/pkg/postfix/trunk/Makefile > > Modified: csw/mgar/pkg/postfix/trunk/Makefile > === > --- csw/mgar/pkg/postfix/trunk/Makefile 2012-03-15 12:09:31 UTC (rev 17417) > +++ csw/mgar/pkg/postfix/trunk/Makefile 2012-03-15 12:10:46 UTC (rev 17418) > @@ -42,6 +42,8 @@ > UFILES_REGEX = $(NAME)-(\d+(?:\.\d+)*).tar.gz > LICENSE = LICENSE TLS_LICENSE > > +PACKAGING_PLATFORMS = solaris10-i386 solaris10-sparc Does the compilation fail in Solaris 9? If not delete these lines and run "gmake replatforms". Chances are, it'll just work. If not, mail maintainers@ with the error message. ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[17675] csw/mgar/gar/v2/lib/python/build_tier_list.py
Hi William,
A quick style scan. Comments inline.
Thanks for writing this!
No dia 14 de Abril de 2012 17:35, escreveu:
> Revision: 17675
> http://gar.svn.sourceforge.net/gar/?rev=17675&view=rev
> Author: wbonnet
> Date: 2012-04-14 16:35:28 + (Sat, 14 Apr 2012)
> Log Message:
> ---
> Initial commit
>
> Added Paths:
> ---
> csw/mgar/gar/v2/lib/python/build_tier_list.py
>
> Added: csw/mgar/gar/v2/lib/python/build_tier_list.py
> ===
> --- csw/mgar/gar/v2/lib/python/build_tier_list.py
> (rev 0)
> +++ csw/mgar/gar/v2/lib/python/build_tier_list.py 2012-04-14 16:35:28
> UTC (rev 17675)
> @@ -0,0 +1,158 @@
> +#!/opt/csw/bin/python
> +
> +import sys
> +import os
> +from optparse import OptionParser
> +
> +catalog = {};
> +
> +
> +#
> -
> +#
> +#
> +class CommandLineParser(object):
> + """Command line parser. This class is a helper used to retrive options
> from command line
> + """
> +
> + def __init__(self):
> + # Create the option parser
> + self.parser = OptionParser()
> +
> + # Add the different command line options to the parser
> + self.parser.add_option("-c", "--catalog", help="Defines the catalog
> to parse. Default is ./catalog",
> + action="store", dest="catalog",
> type="string")
> + self.parser.add_option("-1", "--tier1-list", help="List of tier 1
> packages",
> + action="store", dest="tier1", type="string")
> + self.parser.add_option("-2", "--tier2-list", help="List of tier 2
> packages",
> + action="store", dest="tier2", type="string")
> + self.parser.add_option("-3", "--tier3-list", help="List of tier 3
> packages",
> + action="store", dest="tier3", type="string")
> + self.parser.add_option("-V", "--verbose", help="Activate verbose
> mode", action="store_true", dest="verbose")
> +
> + def parse(self):
> + (self.options, self.args) = self.parser.parse_args()
> + return self.options, self.args
> +
> +#
> -
> +#
> +#
> +class ConfigurationParser(object):
> + """This class is a helper providing getter and setter on the option from
> command line
> + """
> +
> + def __init__(self, args):
> +
> + if args.verbose != None:
> + self.verbose = args.verbose
> + else:
> + self.verbose = False
You can use "default=..." option when defining the flag, and avoid
this boilerplate code.
> + if args.catalog != None:
> + self.catalog = args.catalog
> + else:
> + self.catalog = "./catalog"
> +
> + if args.tier1 != None:
> + self.tier1 = args.tier1
> + else:
> + self.tier1 = "./tier1"
> +
> + if args.tier2 != None:
> + self.tier2 = args.tier2
> + else:
> + self.tier2 = "./tier2"
> +
> + if args.tier3 != None:
> + self.tier3 = args.tier3
> + else:
> + self.tier3 = "./tier3"
> +
> + #
> -
> +
> + def getCatalog(self):
> + return self.catalog
> +
> + def getTierInputFile(self, tier):
> + if (tier == 1):
> + return self.tier1
This could be replaced with a dictionary:
tiers = {
1: self.tier1,
...
}
return tiers[tier] if tier in tiers else None
(or maybe something less compressed, but still data driven)
> + if (tier == 2):
> + return self.tier2
> +
> + if (tier == 3):
> + return self.tier3
> +
> + return None
> +
> + def getTierOutputFile(self, tier):
> + return "%(filename)s.out" % { 'filename' :
> self.getTierInputFile(tier) }
> +
> + def getVerbose(self):
> + return self.verbose
Looks like a javaism. Just let clients use youobject.verbose. If at
any stage you want to control it, you can use @property. This is the
beauty of Python!
> +#
> -
> +#
> +#
> +class Package:
> + """ Defines a package. A package has a name (CSWfoo), a list of
> dependencies (same syntax as
> + catalog (CSWfoo|CSWbar) and a tier (1, 2, 3 or None if undefined)
> + """
> + def __init__(self, name=None, depends=None, tier=None):
> + self.name = name
> + if (depends != "none"):
> + self.depends = depends
> + else:
> + self.depends = None
>
Re: [csw-devel] SF.net SVN: gar:[17678] csw/mgar/gar/v2/lib/python/build_tier_list.py
No dia 14 de Abril de 2012 19:31, escreveu:
> + print "tier %(tier)d\t%(1)d\t%(2)d\t%(3)d\t%(4)d" % {
> 'tier' : tier + 1 , '1' : countPkg[0][tier], '2' : countPkg[1][tier],'3' :
> countPkg[2][tier] , '4' : countPkg[2][tier] - countPkg[1][tier] }
It's probably a better idea to build a data structure at this point.
You can then either serialize this data structure to JSON, or use a
templating language to render it (I suggest jinja2 if you're choosing
a template. I started disliking Cheetah.).
Maciej
___
devel mailing list
devel@lists.opencsw.org
https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[17983] csw/mgar/pkg/libotf/trunk/Makefile
2012/5/10 > +# this is private and not available publicly yet > +mydependencies: > + $(HOME)/bin/ocswdeplist --package $(NAME) > That's a problem. Anyone who tries to build it, will get an error. If you checked it into GAR sources, it wouldn't fail. ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[17983] csw/mgar/pkg/libotf/trunk/Makefile
2012/5/13 Peter FELECAN > > BTW, the advantage of this tool is to extract the dependencies using the > system on which it runs and making reduction of the dependency tree, > mainly by transitivity... Transitivity, in what sense? As far as dependencies go, the OpenCSW policy is to always include direct dependencies, and never to rely on transitive dependencies. > also, more precise than pkgcheck; you can see > the sources in my home bin/python directory. It's interesting. Do you think I'll be able to tell how more precise it is by just looking a the sources? I wouldn't mind looking at some output. Maciej ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18015] csw/mgar/pkg/m17n/trunk/Makefile
2012/5/14 : > override incorrectly evaluated, by checkpkg, surplus dependencies. Could you elaborate, what's the problem exactly? If checkpkg does something incorrectly, we should fix it, no? ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18013] csw/mgar/pkg/m17n/trunk/Makefile
2012/5/14 : > # added this overrides because checkpkg evaluates them incorrectly, > # e.g. libthai0 doesn't exist and how the heck it determines that it > # depends on libgs2? checkpkg always tells exactly why it introduces dependencies. Did you look at the output? ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18012] csw/mgar/pkg/qt4-gcc/trunk/Makefile
2012/5/14 : > +# CONFIGURE_ARGS_sparc += -big-endian -host-big-endian Does it still build on sparc? ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18012] csw/mgar/pkg/qt4-gcc/trunk/Makefile
2012/5/14 Carsten Grzemba : > > Am 14.05.12, schrieb Maciej (Matchek) Bliziński : > > 2012/5/14 : >> +# CONFIGURE_ARGS_sparc += -big-endian -host-big-endian > > Does it still build on sparc? > > I have not tried it yet. I think you'll find that -big-endian and -host-big-endian is necessary on sparc. ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18012] csw/mgar/pkg/qt4-gcc/trunk/Makefile
>> Does it still build on sparc? >> >> I have not tried it yet. > > I think you'll find that -big-endian and -host-big-endian is necessary on > sparc. > > Ok, I add them again, but why has i386: -little-endian -host-big-endian ? Not sure abut -host-big-endian, it might be unnecessary. ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18019] csw/mgar/pkg/qt4-gcc/trunk/Makefile
2012/5/15 > Revision: 18019 > http://gar.svn.sourceforge.net/gar/?rev=18019&view=rev > Author: cgrzemba > Date: 2012-05-15 06:08:53 + (Tue, 15 May 2012) > Log Message: > --- > qt4-gcc/trunk: readd sparc configure args > > Modified Paths: > -- >csw/mgar/pkg/qt4-gcc/trunk/Makefile > > Modified: csw/mgar/pkg/qt4-gcc/trunk/Makefile > === > --- csw/mgar/pkg/qt4-gcc/trunk/Makefile 2012-05-14 13:59:27 UTC (rev 18018) > +++ csw/mgar/pkg/qt4-gcc/trunk/Makefile 2012-05-15 06:08:53 UTC (rev 18019) > @@ -35,7 +35,7 @@ > CONFIGURE_ARGS += -qt-libjpeg -qt-libpng -qt-libmng -qt-zlib > CONFIGURE_ARGS += -I/usr/include -I/usr/X11/include > CONFIGURE_ARGS += -I/usr/X11/share/include -I/usr/sfw/include > -# CONFIGURE_ARGS_sparc += -big-endian -host-big-endian > +CONFIGURE_ARGS_sparc += -big-endian -host-big-endian > Adding this line above... > # CONFIGURE_ARGS_i386 += -little-endian -host-big-endian > # CONFIGURE_ARGS += $(CONFIGURE_ARGS_$(GARCH)) > ...without adding this line, will not do anything. ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18118] csw/mgar/pkg/emacs/trunk/Makefile
2012/5/25 > Revision: 18118 > http://gar.svn.sourceforge.net/gar/?rev=18118&view=rev > Author: pfelecan > Date: 2012-05-25 12:50:58 + (Fri, 25 May 2012) > Log Message: > --- > new override refinements (it never ends?) > Maybe the overrides you put in are too specific? If you put in specific overrides in place, and anything changes, they will stop matching. It could be what you want, but in this case it sounds like it isn't. > > Modified Paths: > -- >csw/mgar/pkg/emacs/trunk/Makefile > > Modified: csw/mgar/pkg/emacs/trunk/Makefile > === > --- csw/mgar/pkg/emacs/trunk/Makefile 2012-05-25 09:40:49 UTC (rev 18117) > +++ csw/mgar/pkg/emacs/trunk/Makefile 2012-05-25 12:50:58 UTC (rev 18118) > @@ -115,6 +115,72 @@ > SPKG_DESC_CSWemacs-common = $(DESCRIPTION) - > architecture neutral common components > CATALOGNAME_CSWemacs-common= emacs_common > CHECKPKG_OVERRIDES_CSWemacs-common += > missing-dependency|CSWemacscommon > +CHECKPKG_OVERRIDES_CSWemacs-common += > file-with-bad-content|/usr/share|root/opt/csw/share/info/dbus > You could say: CHECKPKG_OVERRIDES_CSWemacs-common += file-with-bad-content Which would override all of the file-with-bad-content error tags. ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18119] csw/mgar/pkg/emacs/trunk/Makefile
2012/5/25 > Revision: 18119 > http://gar.svn.sourceforge.net/gar/?rev=18119&view=rev > Author: pfelecan > Date: 2012-05-25 13:24:03 + (Fri, 25 May 2012) > Log Message: > --- > new refinement for overrides... probably 1 step from getting them right! > One other thought is - if the references to /usr/share and /usr/local aren't necessary, you could just remove them, e.g. sed the files. ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18119] csw/mgar/pkg/emacs/trunk/Makefile
And one more thought - I just want to make sure that you're aware that you need to run "mgar remerge" after you change any overrides. ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[17983] csw/mgar/pkg/libotf/trunk/Makefile
2012/5/26 Peter FELECAN : > "Maciej (Matchek) Bliziński" writes: > >> 2012/5/13 Peter FELECAN >>> >>> BTW, the advantage of this tool is to extract the dependencies using the >>> system on which it runs and making reduction of the dependency tree, >>> mainly by transitivity... >> >> Transitivity, in what sense? As far as dependencies go, the OpenCSW >> policy is to always include direct dependencies, and never to rely on >> transitive dependencies. > > Is this policy documented? I had an exchange with Phil once about it. I scanned the archived website content and didn't find it mentioned. I'll need to look more to find the email thread. So, I can't provide you with a link right now. The argument is that if you have CSWfoo → CSWbar → CSWbaz when CSWfoo really depends on files in CSWbaz, then at some point CSWbar can stop depending on CSWbaz, and drop the dependency. This would break CSWfoo. So, if there is a direct actual dependency, it should be declared directly. >>> also, more precise than pkgcheck; you can see >>> the sources in my home bin/python directory. >> >> It's interesting. Do you think I'll be able to tell how more precise >> it is by just looking a the sources? I wouldn't mind looking at some >> output. > > Frankly I thought so... I'll send you something when I find a simple and > usefull output. Cool. ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18162] csw/mgar/gar/v2-solaris11/lib/python/system_pkgmap. py
2012/5/29
> Revision: 18162
> http://gar.svn.sourceforge.net/gar/?rev=18162&view=rev
> Author: chninkel
> Date: 2012-05-29 21:59:21 + (Tue, 29 May 2012)
> Log Message:
> ---
> v2-solaris11: create a fake SVR4 pkg name for IPS package
>
> Modified Paths:
> --
>csw/mgar/gar/v2-solaris11/lib/python/system_pkgmap.py
>
> Modified: csw/mgar/gar/v2-solaris11/lib/python/system_pkgmap.py
> ===
> --- csw/mgar/gar/v2-solaris11/lib/python/system_pkgmap.py 2012-05-29
> 19:29:11 UTC (rev 18161)
> +++ csw/mgar/gar/v2-solaris11/lib/python/system_pkgmap.py 2012-05-29
> 21:59:21 UTC (rev 18162)
> @@ -111,7 +111,7 @@
>
> def _ParsePkgListLine(self, line):
> fields = re.split(c.WS_RE, line)
> -pkgname = fields[0]
> +pkgname = self._IpsNameToPkgname(fields[0])
> desc_field_start = 1
> # The optional publisher field is always between
> # parenthesis, we skip it if necessary
> @@ -144,7 +144,7 @@
> f_owner = None
> f_group = None
> f_pkgname = None
> -pkgnames = [ parts[2] ]
> +pkgnames = [ self._IpsNameToPkgname(parts[2]) ]
> if f_type == 's' or f_type == 'l':
> f_target = parts[3]
> else:
> @@ -385,8 +385,11 @@
> packages_by_pkgname.setdefault(pkgname, pkg_desc)
> logging.debug("<- _ParsePkginfoOutput()")
> return packages_by_pkgname
> +
> + def _IpsNameToPkgname(self, ips_name):
> +"""Create a fake Svr4 pkgname from an ips pkgname"""
> +return "SUNW" + ips_name.replace("/", "_")
>
Underscores aren't legal in pkgnames. I'd suggest doing something more
strict: finding all the alphanumeric string segments and joining them with
dashes.
___
devel mailing list
devel@lists.opencsw.org
https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18158] csw/mgar/pkg/t1lib/trunk/Makefile
2012/5/29 > +# why is this? .so are just symbolic links: > +CHECKPKG_OVERRIDES_CSWt1lib-dev+= archall-devel-package > Discussion on maintainers: http://lists.opencsw.org/pipermail/maintainers/2010-February/011234.html Are you sure that you get identical header results on both architectures? ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18112] csw/mgar/pkg/gsl/trunk/Makefile
2012/5/25 > Revision: 18112 > http://gar.svn.sourceforge.net/gar/?rev=18112&view=rev > Author: pfelecan > Date: 2012-05-25 07:13:06 + (Fri, 25 May 2012) > Log Message: > --- > acted upon checkpkg suggestions > > Modified Paths: > -- >csw/mgar/pkg/gsl/trunk/Makefile > > Modified: csw/mgar/pkg/gsl/trunk/Makefile > === > --- csw/mgar/pkg/gsl/trunk/Makefile 2012-05-24 20:15:49 UTC (rev 18111) > +++ csw/mgar/pkg/gsl/trunk/Makefile 2012-05-25 07:13:06 UTC (rev 18112) > @@ -35,6 +35,7 @@ > PKGFILES_CSWlibgslcblas0 += $(call > baseisadirs,$(libdir),libgslcblas\.so\.0\.0\.0) > PKGFILES_CSWlibgslcblas0 += $(call > baseisadirs,$(libdir),libgslcblas\.so\.0(\.\d+)*) > RUNTIME_DEP_PKGS_CSWlibgslcblas0 += CSWlibgcc-s1 > +ARCHALL_CSWgslrt = 1 > This ARCHALL can be probably done automatically by OBSOLETED_BY_. > +# note that CSWgsl-doc is not an architectural neutral > +ARCHALL_CSWgsldoc = 1 > If not, why do you mark it ARCHALL? > +CHECKPKG_OVERRIDES_CSWgsl-doc += > file-with-bad-content|/usr/local|root/opt/csw/share/info/gsl-ref.info-3 > +CHECKPKG_OVERRIDES_CSWgsl-doc += > file-with-bad-content|/usr/local|root/opt/csw/share/info/gsl-ref.info-1 > +CHECKPKG_OVERRIDES_CSWgsl-doc += > file-collision|/opt/csw/share/info/gsl-ref.info-6|CSWgsl-doc|CSWgsldoc > +CHECKPKG_OVERRIDES_CSWgsl-doc += file-collision|/opt/csw/share/info/ > gsl-ref.info|CSWgsl-doc|CSWgsldoc > +CHECKPKG_OVERRIDES_CSWgsl-doc += > file-collision|/opt/csw/share/info/gsl-ref.info-3|CSWgsl-doc|CSWgsldoc > +CHECKPKG_OVERRIDES_CSWgsl-doc += > file-collision|/opt/csw/share/info/gsl-ref.info-2|CSWgsl-doc|CSWgsldoc > +CHECKPKG_OVERRIDES_CSWgsl-doc += > file-collision|/opt/csw/share/info/gsl-ref.info-5|CSWgsl-doc|CSWgsldoc > +CHECKPKG_OVERRIDES_CSWgsl-doc += > file-collision|/opt/csw/share/info/gsl-ref.info-1|CSWgsl-doc|CSWgsldoc > +CHECKPKG_OVERRIDES_CSWgsl-doc += > file-collision|/opt/csw/share/info/gsl-ref.info-4|CSWgsl-doc|CSWgsldoc > Are you sure you want to introduce file collisions? ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18088] csw/mgar/pkg/squid/branches/squid3/Makefile
2012/5/23 > Revision: 18088 > http://gar.svn.sourceforge.net/gar/?rev=18088&view=rev > Author: wilbury > Date: 2012-05-23 11:32:17 + (Wed, 23 May 2012) > Log Message: > --- > squid/branches/squid3: Adjust CONFIGURE_ARGS > > Modified Paths: > -- >csw/mgar/pkg/squid/branches/squid3/Makefile > > Modified: csw/mgar/pkg/squid/branches/squid3/Makefile > === > --- csw/mgar/pkg/squid/branches/squid3/Makefile 2012-05-23 11:30:52 UTC > (rev 18087) > +++ csw/mgar/pkg/squid/branches/squid3/Makefile 2012-05-23 11:32:17 UTC > (rev 18088) > @@ -58,7 +58,7 @@ > CONFIGURE_ARGS += --docdir=$(datadir)/$(NAME) > CONFIGURE_ARGS += --localstatedir=/var/opt/csw/$(NAME) > CONFIGURE_ARGS += --enable-arp-acl > -CONFIGURE_ARGS += --enable-auth=basic,ntlm > +CONFIGURE_ARGS += --enable-auth=basic,ntlm,digest,negotiate > CONFIGURE_ARGS += --enable-basic-auth-helpers=LDAP,SMB,YP,PAM,SASL,NCSA > CONFIGURE_ARGS += --enable-cache-digests > CONFIGURE_ARGS += --enable-carp > @@ -72,6 +72,7 @@ > CONFIGURE_ARGS += --enable-snmp > CONFIGURE_ARGS += --enable-ssl > CONFIGURE_ARGS += --enable-useragent-log > +CONFIGURE_ARGS += --enable-ipf-transparent > CONFIGURE_ARGS += --disable-ident-lookups > CONFIGURE_ARGS += --disable-icmp > CONFIGURE_ARGS += --with-dl > @@ -80,7 +81,7 @@ > CONFIGURE_ARGS += --with-default-user=$(NAME) > > GARCOMPILER = GNU > -EXTRA_LDFLAGS += $(prefix)/bdb48/lib > +EXTRA_LDFLAGS += -L$(prefix)/bdb48/lib > I think you meang EXTRA_LIB. http://sourceforge.net/apps/trac/gar/wiki/GAR%20Variable%20Reference It would look like: EXTRA_LIB = $(prefix)/bdb68/lib > EXTRA_INC += $(prefix)/bdb48/include > > TEST_TARGET = > > This was sent by the SourceForge.net collaborative development platform, > the world's largest Open Source development site. > > ___ > devel mailing list > devel@lists.opencsw.org > https://lists.opencsw.org/mailman/listinfo/devel > ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18140] csw/mgar/gar/v2-solaris11/lib/python/system_pkgmap. py
2012/5/27
> Revision: 18140
> http://gar.svn.sourceforge.net/gar/?rev=18140&view=rev
> Author: chninkel
> Date: 2012-05-27 21:50:40 + (Sun, 27 May 2012)
> Log Message:
> ---
> no, None is not iterable...
>
> Modified Paths:
> --
>csw/mgar/gar/v2-solaris11/lib/python/system_pkgmap.py
>
> Modified: csw/mgar/gar/v2-solaris11/lib/python/system_pkgmap.py
> ===
> --- csw/mgar/gar/v2-solaris11/lib/python/system_pkgmap.py 2012-05-27
> 21:47:40 UTC (rev 18139)
> +++ csw/mgar/gar/v2-solaris11/lib/python/system_pkgmap.py 2012-05-27
> 21:50:40 UTC (rev 18140)
> @@ -268,6 +268,8 @@
> streams_and_parsers = zip(streams, (self._ParsePkgmapLine,
> self._ParsePkgContentsLine))
> for stream_info in streams_and_parsers:
> parseMethod = stream_info[1]
> + if stream_info[0] == None:
> +continue
>
The better Python idiom is:
if foo is None:
...
but in this case, it's enough to test for the boolean value:
if not stream_info[0]:
continue
> for line in stream_info[0]:
> if show_progress:
> if not c.next() % 1000:
> @@ -377,9 +379,10 @@
> for line in streams[0]:
> pkgname, pkg_desc = self._ParsePkginfoLine(line)
> packages_by_pkgname.setdefault(pkgname, pkg_desc)
> -for line in streams[1]:
> - pkgname, pkg_desc = self._ParsePkgListLine(line)
> - packages_by_pkgname.setdefault(pkgname, pkg_desc)
> +if streams[1] != None:
>
Same deal:
if streams[1]:
for line in streams[1]:
(...)
> + for line in streams[1]:
> +pkgname, pkg_desc = self._ParsePkgListLine(line)
> +packages_by_pkgname.setdefault(pkgname, pkg_desc)
> logging.debug("<- _ParsePkginfoOutput()")
> return packages_by_pkgname
>
>
> This was sent by the SourceForge.net collaborative development platform,
> the world's largest Open Source development site.
>
> ___
> devel mailing list
> devel@lists.opencsw.org
> https://lists.opencsw.org/mailman/listinfo/devel
>
___
devel mailing list
devel@lists.opencsw.org
https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18162] csw/mgar/gar/v2-solaris11/lib/python/system_pkgmap. py
2012/5/29 > Revision: 18162 > http://gar.svn.sourceforge.net/gar/?rev=18162&view=rev > Author: chninkel > Date: 2012-05-29 21:59:21 + (Tue, 29 May 2012) > Log Message: > --- > v2-solaris11: create a fake SVR4 pkg name for IPS package > > Modified Paths: > -- >csw/mgar/gar/v2-solaris11/lib/python/system_pkgmap.py > > Modified: csw/mgar/gar/v2-solaris11/lib/python/system_pkgmap.py > === > --- csw/mgar/gar/v2-solaris11/lib/python/system_pkgmap.py 2012-05-29 > 19:29:11 UTC (rev 18161) > +++ csw/mgar/gar/v2-solaris11/lib/python/system_pkgmap.py 2012-05-29 > 21:59:21 UTC (rev 18162) > @@ -111,7 +111,7 @@ > > def _ParsePkgListLine(self, line): > fields = re.split(c.WS_RE, line) > -pkgname = fields[0] > +pkgname = self._IpsNameToPkgname(fields[0]) > Does this code still work as expected for Solaris 9 and 10? ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18177] csw/mgar/gar/v2-solaris11/lib/python/system_pkgmap. py
2012/5/30 > Revision: 18177 > http://gar.svn.sourceforge.net/gar/?rev=18177&view=rev > Author: chninkel > Date: 2012-05-30 11:17:31 + (Wed, 30 May 2012) > Log Message: > --- > v2-solaris11: improve ips name to svr4 name conversion > > Modified Paths: > -- >csw/mgar/gar/v2-solaris11/lib/python/system_pkgmap.py > > Modified: csw/mgar/gar/v2-solaris11/lib/python/system_pkgmap.py > === > --- csw/mgar/gar/v2-solaris11/lib/python/system_pkgmap.py 2012-05-30 > 08:37:18 UTC (rev 18176) > +++ csw/mgar/gar/v2-solaris11/lib/python/system_pkgmap.py 2012-05-30 > 11:17:31 UTC (rev 18177) > @@ -21,6 +21,7 @@ > import sys > > CONTENT_PKG_RE = r"^\*?(CSW|SUNW)[0-9a-zA-Z\-]?[0-9a-z\-]+$" > +ALPHANUMERIC_RE = r"[0-9a-zA-Z]+" > The special \w sequence will do the same, so: r"\w+" ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18158] csw/mgar/pkg/t1lib/trunk/Makefile
2012/5/31 Peter FELECAN : > > "Maciej (Matchek) Bliziński" writes: > >> 2012/5/29 >> >>> +# why is this? .so are just symbolic links: >>> +CHECKPKG_OVERRIDES_CSWt1lib-dev += archall-devel-package >>> >> >> Discussion on maintainers: >> http://lists.opencsw.org/pipermail/maintainers/2010-February/011234.html >> >> Are you sure that you get identical header results on both architectures? > > Quite. BTW, the headers are not generated but supplied as is from > upstream. TBT[1] You can override the error and package arch-independent headers, but for the checking purposes we decided that a cost of a mistake is less if you create arch-dependent packages by default. The only downside is a tiny increase in catalog size (completely dwarfed by other packages we have). If you make a mistake by creating an arch-independent package, the costs are much higher, because it will cause weird compilation problems, draining hours of people's time. For simplicity, we always create arch-dependent development packages. If you create arch-independent package, you'll need to add overrides, that's all. The choice is of course, yours. Maciej ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18112] csw/mgar/pkg/gsl/trunk/Makefile
2012/5/31 Peter FELECAN : >>> +CHECKPKG_OVERRIDES_CSWgsl-doc += >>> file-collision|/opt/csw/share/info/gsl-ref.info-4|CSWgsl-doc|CSWgsldoc >>> >> >> Are you sure you want to introduce file collisions? > > gsl_doc obsoletes gsldoc, consequently checkpkg is confused, isn't it? Hm... why should it be confused? It just follows what will be the state of the catalog when you insert the packages under test. > this are not real collision but I must shut up the dear > checker... I'm not sure that you must. Did you empty the previous package and push it as well? > wondering if this is not a side effect from the time when I > didn't provide the stub for gsldoc. TBC[1] I would call it a direct effect. One way or another, you need to remove the old colliding files. You either need to remove the old package (csw-upload-pkg --remove), or provide a stub for it. Maciej ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18307] csw/mgar/pkg/grip/trunk/Makefile
2012/6/10 : > -GCC4_VERSION = 4.7 > +#GCC4_VERSION = 4.7 You can set this in ~/.garrc on your private buildfarm. This way it won't leak to build recipes, which will work in both places (provided that the upstream compiles with both 4.6 and 4.7). ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18299] csw/mgar/pkg/irssi/trunk/Makefile
2012/6/9 : > -PACKAING_PLATFORMS = solaris10-i386 solaris10-sparc > +PACKAGING_PLATFORMS = solaris10-i386 solaris10-sparc I always suggest building first, committing later. ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18275] csw/mgar/pkg/irssi/trunk/Makefile
2012/6/9 : > -RUNTIME_DEP_PKGS += CSWlibglib2-0-0 > -RUNTIME_DEP_PKGS += CSWlibgmodule2-0-0 > +RUNTIME_DEP_PKGS += CSWglib2 Try this: RUNTIME_DEP_PKGS_CSWsvn-5.9 += CSWglib2 RUNTIME_DEP_PKGS_CSWsvn-5.10 += CSWlibglib2-0-0 RUNTIME_DEP_PKGS_CSWsvn-5.10 += CSWlibgmodule2-0-0 RUNTIME_DEP_PKGS_CSWsvn += $(RUNTIME_DEP_PKGS_CSWsvn-$(GAROSREL)) ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[16652] csw/mgar/pkg/subversion/trunk/Makefile
Hi Rupert, On Mon, Jan 02, 2012 at 08:17:25PM +, rthur...@users.sourceforge.net wrote: > +CHECKPKG_OVERRIDES_CSWpythonsvn += > discouraged-path-in-pkgmap|/opt/csw/lib/svn/python/site-packages/svn/repos.pyc > +CHECKPKG_OVERRIDES_CSWpythonsvn += > discouraged-path-in-pkgmap|/opt/csw/lib/svn/python/site-packages/svn/wc.pyc > +CHECKPKG_OVERRIDES_CSWpythonsvn += > discouraged-path-in-pkgmap|/opt/csw/lib/svn/python/site-packages/libsvn/__init__.pyc > (...) We don't ship *.pyc files in Python packages. We use the pycompile class instead, and generate the *.pyc files locally. Instead of overriding the error tags, you can remove the files in a post-install-modulated step and set the pycompile class on them. ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18514] csw/mgar/pkg/amanda/trunk
2012/6/24 Peter Bonivart : > On Sun, Jun 24, 2012 at 3:35 PM, Peter FELECAN wrote: >> wah...@users.sourceforge.net writes: >> >>> +# DEP_PKGS += CSWcurlrt CSWosslrt CSWlibnet >> >> What's the role of DEP_PKGS (viz. RUNTIME_DEP_PKGS or BUILD_DEP_PKGS)? > > It's been replaced by RUNTIME_DEP_PKGS. I think that DEP_PKGS adds a package name to both runtime and build time deps. But in practice it's of little use, because with most libraries you want the *_dev package as the build dep, and lib* as the runtime dependency. ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18518] csw/mgar/pkg/pstoedit/trunk/Makefile
2012/6/25 : > Revision: 18518 > http://gar.svn.sourceforge.net/gar/?rev=18518&view=rev > Author: pfelecan > Date: 2012-06-25 09:55:11 + (Mon, 25 Jun 2012) > Log Message: > --- > availability limited to Solaris 10 When committing code, please add the package name as the first word of the commit message. It makes it easier to read the timeline and commit logs. You can use mgar to do it for you: mgar ci -m "short message" If you do this, mgar will automatically prepend the commit message with the package name (and branch), and call svn. ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18513] csw/mgar/pkg/pilotlink/trunk/Makefile
2012/6/23 : > Revision: 18513 > http://gar.svn.sourceforge.net/gar/?rev=18513&view=rev > Author: pfelecan > Date: 2012-06-23 09:59:22 + (Sat, 23 Jun 2012) > Log Message: > --- > some additional overrides > > Modified Paths: > -- > csw/mgar/pkg/pilotlink/trunk/Makefile > > Modified: csw/mgar/pkg/pilotlink/trunk/Makefile > === > --- csw/mgar/pkg/pilotlink/trunk/Makefile 2012-06-23 09:37:54 UTC (rev > 18512) > +++ csw/mgar/pkg/pilotlink/trunk/Makefile 2012-06-23 09:59:22 UTC (rev > 18513) > @@ -150,7 +150,7 @@ > CHECKPKG_OVERRIDES_CSWpilot-link+= > bad-rpath-entry|/usr/sfw/lib|opt/csw/bin/pilot-dlpsh > CHECKPKG_OVERRIDES_CSWpilot-link+= > bad-rpath-entry|/usr/sfw/lib|opt/csw/bin/pilot-file > CHECKPKG_OVERRIDES_CSWpilot-link+= > bad-rpath-entry|/usr/sfw/lib|opt/csw/bin/pilot-foto > -CHECKPKG_OVERRIDES_CSWpilot-link+= > bad-rpath-entry|/usr/sfw/lib|opt/csw/bin/piot-foto-treo600 > +CHECKPKG_OVERRIDES_CSWpilot-link+= > bad-rpath-entry|/usr/sfw/lib|opt/csw/bin/pilot-foto-treo600 > CHECKPKG_OVERRIDES_CSWpilot-link+= > bad-rpath-entry|/usr/sfw/lib|opt/csw/bin/pilot-foto-treo650 > CHECKPKG_OVERRIDES_CSWpilot-link+= > bad-rpath-entry|/usr/sfw/lib|opt/csw/bin/pilot-getram > CHECKPKG_OVERRIDES_CSWpilot-link+= > bad-rpath-entry|/usr/sfw/lib|opt/csw/bin/pilot-getrom > @@ -236,6 +236,8 @@ > RUNTIME_DEP_PKGS_CSWpilot-link-perl += CSWlibpisock9 > RUNTIME_DEP_PKGS_CSWpilot-link-perl += CSWperl > RUNTIME_DEP_PKGS_CSWpilot-link-perl += CSWlibz1 > +CHECKPKG_OVERRIDES_CSWpilot-link-perl+= > perllocal-pod-in-pkgmap|/opt/csw/lib/perl/5.10.1/perllocal.pod Why override this and not remove the file? I don't know what was the original motivation, but AFAIK, we shouldn't and don't ship the perllocal.pod file. > +CHECKPKG_OVERRIDES_CSWpilot-link-perl+= surplus-dependency|CSWlibz1 Why this override? Is libz.so.1 loaded dynamically? ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18512] csw/mgar/pkg/pilotlink/trunk/Makefile
2012/6/23 : > Revision: 18512 > http://gar.svn.sourceforge.net/gar/?rev=18512&view=rev > Author: pfelecan > Date: 2012-06-23 09:37:54 + (Sat, 23 Jun 2012) > Log Message: > --- > corrected end of line issues after copy&paste... I would suggest only committing after you've verified that the change actually works. It's less code and better quality code to review for others. Maciej ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18511] csw/mgar/pkg/pilotlink/trunk/Makefile
2012/6/23 : > PACKAGES += CSWlibpitcl0 > CATALOGNAME_CSWlibpitcl0 = libpitcl0 > @@ -170,6 +225,10 @@ > RUNTIME_DEP_PKGS_CSWlibpitcl0 += CSWlibpisock9 > RUNTIME_DEP_PKGS_CSWlibpitcl0 += CSWtcl > LICENCE_CSWlibpitcl0 = COPYING.LIB > +CHECKPKG_OVERRIDES_CSWlibpitcl0 += > bad-rpath-entry|/usr/sfw/lib|opt/csw/lib/libpitcl.so.0.0.0 > +CHECKPKG_OVERRIDES_CSWlibpitcl0 += surplus-dependency|CSWlibz1 > +CHECKPKG_OVERRIDES_CSWlibpitcl0 += > surplus-dependency|CSWlibiconv2 > +CHECKPKG_OVERRIDES_CSWlibpitcl0 += surplus-dependency|CSWlibintl8 How do you know that these dependencies are actually required? In most packages, you can rely of the NEEDED field, and there's no need for additional dependencies. Maciej ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18463] csw/mgar/pkg/plotutils/trunk/Makefile
2012/6/19 : > RUNTIME_DEP_PKGS_CSWplotutils += CSWlibplotter2 > RUNTIME_DEP_PKGS_CSWplotutils += CSWlibxmi0 > OBSOLETED_BY_CSWplotutils = CSWplotutil > +CHECKPKG_OVERRIDES_CSWplotutils += > missing-dependency|CSWlibgcc-s1 > +CHECKPKG_OVERRIDES_CSWplotutils += > missing-dependency|CSWlibstdc++6 > +CHECKPKG_OVERRIDES_CSWplotutils += > missing-dependency|CSWlibpng12-0 > +CHECKPKG_OVERRIDES_CSWplotutils += missing-dependency|CSWlibz1 > +CHECKPKG_OVERRIDES_CSWplotutils += surplus-dependency|CSWlibxmi0 > +CHECKPKG_OVERRIDES_CSWplotutils += > surplus-dependency|CSWlibplotter2 This is suspicious, why do you not add dependencies to a package containing binaries with specific SONAMEs in NEEDED? If CSWplotutils contains a binary foo, and foo requires the "libz.so.1" string in NEEDED, it does make sense to make CSWplotutils depend on CSWlibz1, no? Maciej ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18440] csw/mgar/pkg/libgnet/trunk/Makefile
2012/6/16 : > RUNTIME_DEP_PKGS_CSWlibgnet2-0-0 += CSWlibintl8 > +CHECKPKG_OVERRIDES_CSWlibgnet2-0-0 += surplus-dependency|CSWlibiconv2 A similar issue as in other packages - CSWlibiconv2 only adds a single library, usually when a library is used, it's in the NEEDED of its dependent binaries, and checkpkg detects that. The only case when checkpkg doesn't detect that, is dynamic loading. Is it the case here? Maciej ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18439] csw/mgar/pkg/mule/trunk/Makefile
2012/6/16 : > +CHECKPKG_OVERRIDES_CSWmule += > binary-architecture-does-not-match-placement|file=opt/csw/mule/lib/boot/libwrapper-solaris-sparc-64.so|arch_id=43|arch_name=sparcv9 > +CHECKPKG_OVERRIDES_CSWmule += > binary-architecture-does-not-match-placement|file=opt/csw/mule/lib/boot/exec/wrapper-solaris-sparc-64|arch_id=43|arch_name=sparcv9 > +CHECKPKG_OVERRIDES_CSWmule += > archall-with-binaries|opt/csw/mule/lib/boot/exec/wrapper-solaris-sparc-32 > +CHECKPKG_OVERRIDES_CSWmule += > archall-with-binaries|opt/csw/mule/lib/boot/exec/wrapper-solaris-sparc-64 > +CHECKPKG_OVERRIDES_CSWmule += > archall-with-binaries|opt/csw/mule/lib/boot/exec/wrapper-solaris-x86-32 > +CHECKPKG_OVERRIDES_CSWmule += > archall-with-binaries|opt/csw/mule/lib/boot/libwrapper-solaris-sparc-32.so > +CHECKPKG_OVERRIDES_CSWmule += > archall-with-binaries|opt/csw/mule/lib/boot/libwrapper-solaris-sparc-64.so > +CHECKPKG_OVERRIDES_CSWmule += > archall-with-binaries|opt/csw/mule/lib/boot/libwrapper-solaris-x86-32.so > +CHECKPKG_OVERRIDES_CSWmule += > file-with-bad-content|/usr/share|root/var/opt/csw/svc/manifest/application/cswmule.xml > +CHECKPKG_OVERRIDES_CSWmule += > binary-wrong-architecture|file=opt/csw/mule/lib/boot/libwrapper-solaris-sparc-32.so|pkginfo-says=all|actual-binary=sparc > +CHECKPKG_OVERRIDES_CSWmule += > binary-wrong-architecture|file=opt/csw/mule/lib/boot/libwrapper-solaris-sparc-64.so|pkginfo-says=all|actual-binary=sparc > +CHECKPKG_OVERRIDES_CSWmule += > binary-wrong-architecture|file=opt/csw/mule/lib/boot/libwrapper-solaris-x86-32.so|pkginfo-says=all|actual-binary=i386 > +CHECKPKG_OVERRIDES_CSWmule += > binary-wrong-architecture|file=opt/csw/mule/lib/boot/exec/wrapper-solaris-sparc-32|pkginfo-says=all|actual-binary=sparc > +CHECKPKG_OVERRIDES_CSWmule += > binary-wrong-architecture|file=opt/csw/mule/lib/boot/exec/wrapper-solaris-sparc-64|pkginfo-says=all|actual-binary=sparc > +CHECKPKG_OVERRIDES_CSWmule += > binary-wrong-architecture|file=opt/csw/mule/lib/boot/exec/wrapper-solaris-x86-32|pkginfo-says=all|actual-binary=i386 Binaries in an arch=all package, why? ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18436] csw/mgar/pkg/tokyocabinet/trunk/Makefile
2012/6/16 : > CHECKPKG_OVERRIDES_CSWlibtokyocabinet9 += > file-with-bad-content|/usr/local|root/opt/csw/lib/libtokyocabinet.so.9.10.0 > CHECKPKG_OVERRIDES_CSWlibtokyocabinet9 += > file-with-bad-content|/usr/local|root/opt/csw/lib/sparcv9/libtokyocabinet.so.9.10.0 > +CHECKPKG_OVERRIDES_CSWlibtokyocabinet9 += > file-with-bad-content|/usr/local|root/opt/csw/lib/amd64/libtokyocabinet.so.9.10.0 You can compact the above 3 lines into: CHECKPKG_OVERRIDES_CSWlibtokyocabinet9 += file-with-bad-content ...because you probably mean to override all the file-with-bad-content errors for the CSWlibtokyocabinet9 package. ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18518] csw/mgar/pkg/pstoedit/trunk/Makefile
2012/6/27 Peter FELECAN : > "Maciej (Matchek) Bliziński" writes: > >> 2012/6/25 : >>> Revision: 18518 >>> http://gar.svn.sourceforge.net/gar/?rev=18518&view=rev >>> Author: pfelecan >>> Date: 2012-06-25 09:55:11 + (Mon, 25 Jun 2012) >>> Log Message: >>> --- >>> availability limited to Solaris 10 >> >> When committing code, please add the package name as the first word >> of the commit message. It makes it easier to read the timeline and >> commit logs. > > I'll do but I don't see the difference when browsing the repository; > maybe I'm not looking at the relevant place. BTW, I wasn't aware about > this convention; is this published somewhere? I look at it in the Trac timeline view: http://sourceforge.net/apps/trac/gar/timeline ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18512] csw/mgar/pkg/pilotlink/trunk/Makefile
2012/6/27 Peter FELECAN : > "Maciej (Matchek) Bliziński" writes: > >> 2012/6/23 : >>> Revision: 18512 >>> http://gar.svn.sourceforge.net/gar/?rev=18512&view=rev >>> Author: pfelecan >>> Date: 2012-06-23 09:37:54 + (Sat, 23 Jun 2012) >>> Log Message: >>> --- >>> corrected end of line issues after copy&paste... >> >> I would suggest only committing after you've verified that the change >> actually works. It's less code and better quality code to review for >> others. > > Sure, but I can run checkpkg only on the build farm but my modification > takes place on my local system, on which I test but this wasn't detected > without checkpkg. Remember that I asked to run this beast also outside > the build farm? Does it not run outside the buildfarm? There are instructions how to set up a db locally: http://wiki.opencsw.org/checkpkg This is how I run a dev instance. ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18511] csw/mgar/pkg/pilotlink/trunk/Makefile
2012/6/27 Peter FELECAN : > *here* libintl.so.8 => /opt/csw/lib/libintl.so.8 > *here* libiconv.so.2 => /opt/csw/lib/libiconv.so.2 These libraries are loaded, but they immediate dependencies. If you have something like this: foo → libfoo.so.1 → libbar.so.2 Then you need to declare the dependency from CSWfoo to CSWlibfoo1, and CSwlibfoo1 to CSWlibbar2, but not from CSWfoo to CSWlibbar2. In other words, checkpkg tells you the right thing, and you shouldn't need to add any overrides. ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18440] csw/mgar/pkg/libgnet/trunk/Makefile
2012/6/27 Peter FELECAN : > *here* libiconv.so.2 => /opt/csw/lib/libiconv.so.2 The transitive dependency again. Let's discuss this on maintainers@. ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18579] csw/mgar/pkg
2012/6/28 : > +PACKAGING_PLATFORMS = solarsi10-sparc solarsi10-i386 Typo spotted. ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18584] csw/mgar/pkg/grip/trunk/Makefile
2012/6/28 : > Revision: 18584 > http://gar.svn.sourceforge.net/gar/?rev=18584&view=rev > Author: pfelecan > Date: 2012-06-28 11:43:25 + (Thu, 28 Jun 2012) > Log Message: > --- > grip/trunk: remove explicit dependency on old SSL library > > Modified Paths: > -- > csw/mgar/pkg/grip/trunk/Makefile > > Modified: csw/mgar/pkg/grip/trunk/Makefile > === > --- csw/mgar/pkg/grip/trunk/Makefile 2012-06-28 10:01:07 UTC (rev 18583) > +++ csw/mgar/pkg/grip/trunk/Makefile 2012-06-28 11:43:25 UTC (rev 18584) > @@ -95,7 +95,6 @@ > RUNTIME_DEP_PKGS += CSWlibpixman1-0 > RUNTIME_DEP_PKGS += CSWlibpng12-0 > RUNTIME_DEP_PKGS += CSWlibpopt0 > -RUNTIME_DEP_PKGS += CSWlibssl0-9-8 > RUNTIME_DEP_PKGS += CSWlibssl1-0-0 > RUNTIME_DEP_PKGS += CSWlibstdc++6 > RUNTIME_DEP_PKGS += CSWlibxft2 > @@ -123,7 +122,6 @@ > CHECKPKG_OVERRIDES_CSWgrip += surplus-dependency|CSWlibiconv2 > CHECKPKG_OVERRIDES_CSWgrip += surplus-dependency|CSWlibgailutil18 > CHECKPKG_OVERRIDES_CSWgrip += surplus-dependency|CSWlibintl3 > -CHECKPKG_OVERRIDES_CSWgrip += surplus-dependency|CSWlibssl0-9-8 > CHECKPKG_OVERRIDES_CSWgrip += surplus-dependency|CSWlibxml2-2 > CHECKPKG_OVERRIDES_CSWgrip += surplus-dependency|CSWlibpixman1-0 > CHECKPKG_OVERRIDES_CSWgrip += > surplus-dependency|CSWlibgnome-keyring0 These overrides, is this the transitive dependency thing? ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18584] csw/mgar/pkg/grip/trunk/Makefile
2012/6/29 Peter FELECAN > > "Maciej (Matchek) Bliziński" writes: > > > These overrides, is this the transitive dependency thing? > > yes and NEEDED from readelf and elfdump point of view So you disagree with this post? http://lists.opencsw.org/pipermail/maintainers/2012-June/016921.html ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[16652] csw/mgar/pkg/subversion/trunk/Makefile
2012/6/24 Peter FELECAN > > Maciej Bliziński writes: > > On Mon, Jan 02, 2012 at 08:17:25PM +, rthur...@users.sourceforge.net > > wrote: > >> +CHECKPKG_OVERRIDES_CSWpythonsvn += > >> discouraged-path-in-pkgmap|/opt/csw/lib/svn/python/site-packages/svn/repos.pyc > >> +CHECKPKG_OVERRIDES_CSWpythonsvn += > >> discouraged-path-in-pkgmap|/opt/csw/lib/svn/python/site-packages/svn/wc.pyc > >> +CHECKPKG_OVERRIDES_CSWpythonsvn += > >> discouraged-path-in-pkgmap|/opt/csw/lib/svn/python/site-packages/libsvn/__init__.pyc > >> (...) > > > > We don't ship *.pyc files in Python packages. We use the pycompile class > > instead, and generate the *.pyc files locally. Instead of overriding the > > error tags, you can remove the files in a post-install-modulated step > > and set the pycompile class on them. > > Sincerely, I don't get the rationale of this: pyc as elc (e-lisp > compiled) are architecture neutral. The only reason that I can imagine > is that compiling on the target, i.e., on which the package containing > the source is installed, uses environmental, i.e., specific, > information, e.g. Python's current version. > > As our Python reference, can you enlighten us on this? Yes, the *.pyc and *.pyo files are platform-independent. I found the corresponding Debian policy document: http://www.debian.org/doc/packaging-manuals/python-policy/ch-module_packages.html#s-byte_compilation "(...) the corresponding byte-compiled modules (foo.pyc files) and optimized modules (foo.pyo files) must not ship in the package." "Pure Python modules in private installation directories that are byte-compiled with the default Python version must be forcefully byte-compiled again when the default Python version changes." It's not stated directly, but it sounds like the byte-compiled files are dependent on the version of Python runtime. Compiling them in post-install allows to not update the packaged modules if the runtime changes and the module is source-compatible with the new runtime. Maciej ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18843] csw/mgar/pkg/lang-python/xpra/trunk/Makefile
2012/7/24 : > +BUILD_ARGS += --rpath=$(libdir):$(prefix)/X11/lib I think we are going away from the custom X11 libraries. When we build on Solaris 10, the X headers are complete enough to work. ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[18841] csw/mgar/pkg/lang-python/xpra/trunk/Makefile
2012/7/24 : > Revision: 18841 > http://gar.svn.sourceforge.net/gar/?rev=18841&view=rev > Author: cgrzemba > Date: 2012-07-24 11:58:36 + (Tue, 24 Jul 2012) > Log Message: > --- > lang-python/xpra/trunk: add runtime dependencies > > Modified Paths: > -- > csw/mgar/pkg/lang-python/xpra/trunk/Makefile > > Modified: csw/mgar/pkg/lang-python/xpra/trunk/Makefile > === > --- csw/mgar/pkg/lang-python/xpra/trunk/Makefile2012-07-24 11:12:28 > UTC (rev 18840) > +++ csw/mgar/pkg/lang-python/xpra/trunk/Makefile2012-07-24 11:58:36 > UTC (rev 18841) > @@ -26,13 +26,33 @@ > # uncomment the next line. Otherwise it is set by default to the value of > MASTER_SITES > # UPSTREAM_MASTER_SITES = > > -ARCHALL = 1 > -BUILD_DEP_PKGS += CSWpy-gobject-dev CSWlibgtk2-dev CSWlibxtstdevel > +BUILD_DEP_PKGS += CSWpy-gobject-dev CSWlibgtk2-dev CSWlibxtstdevel > CSWpangodevel I would generally keep one dependency per line. This way the diffs are easier to review. > BUILD_DEP_PKGS += CSWlibxfixesdevel CSWlibxcompositedev CSWlibxdamagedevel > CSWlibxrandrdevel > BUILD_DEP_PKGS += CSWx11recordproto CSWlibxextdevel CSWx11fixesproto > CSWx11compositeproto > BUILD_DEP_PKGS += CSWx11damageproto CSWx11randrproto CSWcython CSWpython-dev > BUILD_DEP_PKGS += CSWlibx264-dev CSWffmpeg-dev > > +REINPLACE_USRSHARE += $(libdir)/python/site-packages/xpra/xposix/gui.py > +REINPLACE_USRSHARE += > $(libdir)/python/site-packages/xpra/platform/client_extras_base.py > +REINPLACE_WHEN_USRSHARE = postinstall > +REINPLACE_USRLOCAL += $(libdir)/python/site-packages/xpra/xposix/gui.py > +REINPLACE_WHEN_USRLOCAL = postinstall > + > +PACKAGE += CSWxpra > + > +RUNTIME_DEP_PKGS_CSWxpra += CSWpython > +RUNTIME_DEP_PKGS_CSWxpra += CSWlibswscale2 CSWlibavcodec54 Same goes here. If you keep one dep per line, it's easier to review. > +RUNTIME_DEP_PKGS_CSWxpra += CSWlibx11 I think that we don't want to depend on this library. > CSWlibgio2-0-0 > +RUNTIME_DEP_PKGS_CSWxpra += CSWpango CSWlibcairo2 CSWlibfreetype6 > +RUNTIME_DEP_PKGS_CSWxpra += CSWlibxdamage CSWlibxcomposite > +RUNTIME_DEP_PKGS_CSWxpra += CSWlibx264-125 > +RUNTIME_DEP_PKGS_CSWxpra += CSWlibxfixes > + > +CHECKPKG_OVERRIDES_CSWxpra += > pkginfo-description-not-starting-with-uppercase|screen|for|X > +CHECKPKG_OVERRIDES_CSWxpra += pkgname-does-not-start-with-CSWpy- > +CHECKPKG_OVERRIDES_CSWxpra += catalogname-does-not-start-with-py_ > + > + > EXTRA_MERGE_EXCLUDE_FILES = .*\.pyo .*\.pyc > > CONFIGURE_SCRIPTS = > > This was sent by the SourceForge.net collaborative development platform, the > world's largest Open Source development site. > > ___ > devel mailing list > devel@lists.opencsw.org > https://lists.opencsw.org/mailman/listinfo/devel ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[19140] csw/mgar/pkg/lang-python/salt/trunk/Makefile
2012/9/4 : > Revision: 19140 > http://gar.svn.sourceforge.net/gar/?rev=19140&view=rev > Author: romeotheriault > Date: 2012-09-04 08:47:12 + (Tue, 04 Sep 2012) > Log Message: > --- > lang-python/salt/trunk: Comment out uncommited line > > Modified Paths: > -- > csw/mgar/pkg/lang-python/salt/trunk/Makefile > > Modified: csw/mgar/pkg/lang-python/salt/trunk/Makefile > === > --- csw/mgar/pkg/lang-python/salt/trunk/Makefile2012-09-04 08:45:10 > UTC (rev 19139) > +++ csw/mgar/pkg/lang-python/salt/trunk/Makefile2012-09-04 08:47:12 > UTC (rev 19140) > @@ -20,7 +20,7 @@ > RUNTIME_DEP_PKGS_CSWpy-salt += CSWpy-pyzmq > RUNTIME_DEP_PKGS_CSWpy-salt += CSWpy-msgpack-python > > -CHECKPKG_OVERRIDES_CSWpy-salt += pkginfo-opencsw-repository-uncommitted > +#CHECKPKG_OVERRIDES_CSWpy-salt += pkginfo-opencsw-repository-uncommitted What was the reason you were adding this override? ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[19097] csw/mgar/pkg/gnome-base/gnomevfs2/trunk/Makefile
2012/8/26 : > # For GNU gettext > -PATH := /opt/csw/gnu:$(PATH) > +CONFIGURE_ENV += XGETTEXT=/opt/csw/gnu/xgettext > +CONFIGURE_ENV += MSGMERGE=/opt/csw/gnu/msgmerge > +CONFIGURE_ENV += MSGFMT=/opt/csw/gnu/msgfmt I was curious if there wasn't a more specific GAR abstraction for this, e.g. extra env flags that GAR would put together for you. ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[19095] csw/mgar/pkg/lang-python/salt/trunk
2012/8/26 : > +CHECKPKG_OVERRIDES_CSWpy-salt += > file-with-bad-content|/usr/local|root/opt/csw/lib/python/site-packages/salt/states/pkgng.py > +CHECKPKG_OVERRIDES_CSWpy-salt += > file-with-bad-content|/usr/local|root/opt/csw/lib/python/site-packages/salt/modules/pkgng.py > +CHECKPKG_OVERRIDES_CSWpy-salt += > file-with-bad-content|/usr/local|root/opt/csw/lib/python/site-packages/salt/modules/poudriere.py > +CHECKPKG_OVERRIDES_CSWpy-salt += > file-with-bad-content|/usr/local|root/opt/csw/lib/python/site-packages/salt/modules/cassandra.py Instead of all these specific overrides, you can add one general one: CHECKPKG_OVERRIDES_CSWpy-salt += file-with-bad-content It will apply to all files in the package. ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[19140] csw/mgar/pkg/lang-python/salt/trunk/Makefile
2012/9/4 Romeo Theriault > On Tue, Sep 4, 2012 at 6:57 AM, Maciej (Matchek) Bliziński < > mac...@opencsw.org> wrote: > >> 2012/9/4 : >> > -CHECKPKG_OVERRIDES_CSWpy-salt += pkginfo-opencsw-repository-uncommitted >> > +#CHECKPKG_OVERRIDES_CSWpy-salt += >> pkginfo-opencsw-repository-uncommitted >> >> What was the reason you were adding this override? >> > > Basically, so I can test a full build process before I commit to make sure > I have everything working correctly before I commit. > This error tag informs you that you haven't committed the recipe (which you know anyway most of the time, but humanum est errare). You can work until the pkginfo-opencsw-repository-uncommitted is the only error tag you see. Then you commit, run "mgar repackage" once more and all the error tags are gone ‒ which is what you want. This is the safest way, highly recommended. Maciej ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[19218] csw/mgar/gar/v2/lib/python
2012/9/17 > Modified: csw/mgar/gar/v2/lib/python/checkpkg2.py > === > --- csw/mgar/gar/v2/lib/python/checkpkg2.py 2012-09-17 12:37:40 UTC > (rev 19217) > +++ csw/mgar/gar/v2/lib/python/checkpkg2.py 2012-09-17 13:31:31 UTC > (rev 19218) > @@ -1,4 +1,4 @@ > -#!/usr/bin/env python2.6 > +#!/opt/csw/bin/python2.6 > There are also the unit tests. You're supposed to run them before you commit code. ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] [csw-maintainers] Checkpkg test for direct binding
2012/9/6 Yann Rouillard
> --- dependency_checks.py (revision 19130)
> +++ dependency_checks.py (working copy)
> @@ -39,6 +39,24 @@
>
> PREFERRED_DIRECTORY_PROVIDERS = set([u"CSWcommon"])
>
> [...]
> +
> +ALLOWED_VERSION_DEPENDENCIES = { "libc.so.1": [ 'SYSVABI_1.3',
> 'SUNWprivate_1.1', 'SUNW_1.22.6', 'SUNW_1.22.5',
>
If you can, limit line width to 80 chars.
> +'SUNW_1.22.4',
> 'SUNW_1.22.3', 'SUNW_1.22.2', 'SUNW_1.22.1',
> +'SUNW_1.22',
> 'SUNW_1.21.3', 'SUNW_1.21.2', 'SUNW_1.21.1',
> +'SUNW_1.21',
> 'SUNW_1.20.4', 'SUNW_1.20.1', 'SUNW_1.20',
> +'SUNW_1.19',
> 'SUNW_1.18.1', 'SUNW_1.18', 'SUNW_1.17',
> +'SUNW_1.16', 'SUNW_1.15',
> 'SUNW_1.14', 'SUNW_1.13',
> +'SUNW_1.12', 'SUNW_1.11',
> 'SUNW_1.10', 'SUNW_1.9',
> +'SUNW_1.8', 'SUNW_1.7',
> 'SUNW_1.6', 'SUNW_1.5',
> +'SUNW_1.4', 'SUNW_1.3',
> 'SUNW_1.2', 'SUNW_1.1',
> +'SUNW_0.9', 'SUNW_0.8',
> 'SUNW_0.7', 'SISCD_2.3' ] }
> +
> +
>
> @@ -147,6 +165,57 @@
>error_mgr,
>pkgname, messenger)
>orphan_sonames.extend(orphan_sonames_tmp)
> +
> +
> +ldd_info = pkg_data['ldd_info'][binary_info["path"]]
> +for ldd_response in ldd_info:
> + if ldd_response['state'] == 'soname-unused' and
> ldd_response['soname'] not in BASE_SOLARIS_LIBRARIES:
> +messenger.Message(
> + "Binary %s links to library %s but doesn't seem to use any of
> its symbols. "
> + "It usually happens because superfluous libraries were added to
> the linker options, "
> + "either because of the configure script itself or because of the
> \"pkg-config --libs\""
> + " output of one the dependency."
> + % ("/" + binary_info["path"], ldd_response['soname']))
> +error_mgr.ReportError(
> +pkgname, "soname-unused",
> +"%s is needed by %s but never used" %
> (ldd_response['soname'], "/" + binary_info["path"]))
> +
> +# Even when direct binding is enabled, some symbols might not be
> +# directly bound because the library explicitely requested the symbol
> +# not to be drectly bound to.
> +# For example, libc.so.1 does it for symbol sigaction, free, malloc
> and realloc
> +# So we consider that direct binding is enabled if at least one
> symbol is directly
> +# bound to because that definitely means that -B direct or -z direct
> was used.
> +directly_binded = set()
>
'To bind' is an irregular verb, past tense for bind is bound.
http://forum.wordreference.com/showthread.php?t=600594
+for syminfo in
> pkg_data["binaries_elf_info"][binary_info["path"]]['symbol table']:
> + if syminfo['external'] and syminfo['flags'] and 'D' in
> syminfo['flags'] and 'B' in syminfo['flags']:
> + directly_binded.add(syminfo['soname'])
> +not_directly_binded =
> directly_binded.symmetric_difference(binary_info["needed sonames"])
> +
> +if not_directly_binded:
> + messenger.Message(
> + "No symbol of binary %s is directly binded against the following
> libraries: %s. "
> +"Please make sure the binaries are compiled using the
> \"-Bdirect\" linker option."
> +% ("/" + binary_info["path"], ", ".join(not_directly_binded)))
> + for soname in not_directly_binded:
> +error_mgr.ReportError(
> + pkgname, "no-direct-binding",
> + "%s is not directly binded to soname %s" % ("/" +
> binary_info["path"], soname))
> +
> +
> +for version_dep in
> pkg_data["binaries_elf_info"][binary_info["path"]]['version needed']:
> + if (version_dep['soname'] in ALLOWED_VERSION_DEPENDENCIES and
> +not version_dep['version'] in
> ALLOWED_VERSION_DEPENDENCIES[version_dep['soname']]):
> +messenger.Message(
> + "Binary %s requires interface version %s in library %s which is
> only available "
> + "in recent Solaris releases."
> + % ("/" + binary_info["path"], version_dep['version'],
> version_dep['soname']))
> +error_mgr.ReportError(
> + pkgname, "forbidden-version-interface-dependencies",
> + "%s requires forbidden interface version %s in library %s"
> + % ("/" + binary_info["path"], version_dep['version'],
> version_dep['soname']))
> +
> +
>orphan_sonames = set(orphan_sonames)
>for soname, binary_path in orphan_sonames:
> if soname not in ALLOWED_ORPHAN_SONAMES:
>
>
> --- inspective_package.py (revision 19130)
> +++ inspective_package.py (working copy)
> @@ -229,17 +229,132 @@
>
> return defined_symbols
>
> +
> + def GetBinaryElfInfo(self):
> +"""Returns various informations symbol and version present in elf
> header
> +
>
Re: [csw-devel] SF.net SVN: gar:[19258] csw/mgar/pkg/libsoup2/trunk/Makefile
Hi Carsten, Overall, it looks like you've submitted the change without testing if the build works and produces the output you want. If you can, try to build and review first, and submit second, so that you don't submit things like silly typos. More comments below. 2012/9/21 > Revision: 19258 > http://gar.svn.sourceforge.net/gar/?rev=19258&view=rev > Author: cgrzemba > Date: 2012-09-21 10:53:43 + (Fri, 21 Sep 2012) > Log Message: > --- > libsoup2/trunk: rework packaging > > Modified Paths: > -- > csw/mgar/pkg/libsoup2/trunk/Makefile > > Modified: csw/mgar/pkg/libsoup2/trunk/Makefile > === > --- csw/mgar/pkg/libsoup2/trunk/Makefile2012-09-21 10:21:43 UTC > (rev 19257) > +++ csw/mgar/pkg/libsoup2/trunk/Makefile2012-09-21 10:53:43 UTC > (rev 19258) > @@ -10,34 +10,56 @@ > MASTER_SITES = $(GNOME_MIRROR) > DISTFILES = $(NAME)-$(VERSION).tar.bz2 > > +PACKAGING_PLATFORMS = Solaris10-sparc Solaris10-i386 > > # We define upstream file regex so we can be notifed of new upstream > software release > UFILES_REGEX = -(\d+(?:\.\d+)*).tar.bz2 > The above line is not necessary any more, there's a function that constructs these regexes on the fly. > > -PACKAGES = CSWlibsoup2 > -PACKAGES += CSWlibsoup2dev > -SPKG_DESC_CSWlibsoup2 = HTTP client/server library for GNOME > -SPKG_DESC_CSWlibsoup2dev = HTTP client/server library for GNOME, dev files > +PACKAGES += CSWlibsoup2-4-1 > +CATALOGNAME_CSWlibsoup2-4-1 = libsoup2_4_1 > +PKGFILES_CSWlibsoup2-4-1 += $(call > baseisadirs,$(libdir),libsoup-2\.4\.so\.1\.3\.0) > +PKGFILES_CSWlibsoup2-4-1 += $(call > baseisadirs,$(libdir),libsoup-2\.4\.so\.1(\.\d+)*) > You only need one line of the two lines above. Checkpkg shows both, and it's your job to choose which one you really need. I'd say, throw away the first one. > +SPKG_DESC_CSWlibsoup2-4-1 += $(DESCRIPTION), libsoup-2.4.so.1 > +RUNTIME_DEP_PKGS_CSWlibsoup2-4-1 += CSWlibgthread2-0-0 > +RUNTIME_DEP_PKGS_CSWlibsoup2-4-1 += CSWlibgmodule2-0-0 > +RUNTIME_DEP_PKGS_CSWlibsoup2-4-1 += CSWlibz1 > +RUNTIME_DEP_PKGS_CSWlibsoup2-4-1 += CSWlibintl8 > +RUNTIME_DEP_PKGS_CSWlibsoup2-4-1 += CSWlibgobject2-0-0 > +RUNTIME_DEP_PKGS_CSWlibsoup2-4-1 += CSWlibglib2-0-0 > +RUNTIME_DEP_PKGS_CSWlibsoup2-4-1 += CSWlibxml2-2 > +RUNTIME_DEP_PKGS_CSWlibsoup2-4-1 += CSWlibgio2-0-0 > +OBSOLTED_BY_CSWlibsoup2-4-1 = CSWlibsoup > Typo, should be OBSOLETED_BY. > > -RUNTIME_DEP_PKGS_CSWlibsoup2 = CSWgconf2 > -RUNTIME_DEP_PKGS_CSWlibsoup2 += CSWgcrypt > -RUNTIME_DEP_PKGS_CSWlibsoup2 += CSWggettextrt > -RUNTIME_DEP_PKGS_CSWlibsoup2 += CSWglib2 > -RUNTIME_DEP_PKGS_CSWlibsoup2 += CSWgnutls > -RUNTIME_DEP_PKGS_CSWlibsoup2 += CSWgpgerr > -RUNTIME_DEP_PKGS_CSWlibsoup2 += CSWlibproxy > -RUNTIME_DEP_PKGS_CSWlibsoup2 += CSWlibxml2 > -RUNTIME_DEP_PKGS_CSWlibsoup2 += CSWsqlite3rt > +PACKAGES += CSWlibsoup-gnome2-4-1 > +CATALOGNAME_CSWlibsoup-gnome2-4-1 = libsoup_gnome2_4_1 > The CATALOGNAME line is not necessary, gar guesses that dashes should be substituted by underscores. > +PKGFILES_CSWlibsoup-gnome2-4-1 += $(call > baseisadirs,$(libdir),libsoup-gnome-2\.4\.so\.1\.3\.0) > +PKGFILES_CSWlibsoup-gnome2-4-1 += $(call > baseisadirs,$(libdir),libsoup-gnome-2\.4\.so\.1(\.\d+)*) > Same comment, only one line needed. > +SPKG_DESC_CSWlibsoup-gnome2-4-1 += $(DESCRIPTION), libsoup-gnome-2.4.so.1 > +RUNTIME_DEP_PKGS_CSWlibsoup-gnome2-4-1 += CSWlibgthread2-0-0 > +RUNTIME_DEP_PKGS_CSWlibsoup-gnome2-4-1 += CSWlibsqlite3-0 > +RUNTIME_DEP_PKGS_CSWlibsoup-gnome2-4-1 += CSWlibgmodule2-0-0 > +RUNTIME_DEP_PKGS_CSWlibsoup-gnome2-4-1 += CSWlibz1 > +RUNTIME_DEP_PKGS_CSWlibsoup-gnome2-4-1 += CSWlibintl8 > +RUNTIME_DEP_PKGS_CSWlibsoup-gnome2-4-1 += CSWlibgnome-keyring0 > +RUNTIME_DEP_PKGS_CSWlibsoup-gnome2-4-1 += CSWlibsoup2-4-1 > +RUNTIME_DEP_PKGS_CSWlibsoup-gnome2-4-1 += CSWlibgobject2-0-0 > +RUNTIME_DEP_PKGS_CSWlibsoup-gnome2-4-1 += CSWlibgio2-0-0 > +RUNTIME_DEP_PKGS_CSWlibsoup-gnome2-4-1 += CSWlibxml2-2 > +RUNTIME_DEP_PKGS_CSWlibsoup-gnome2-4-1 += CSWlibglib2-0-0 > +OBSOLTED_BY_CSWlibsoup-gnome2-4-1 = CSWlibsoup > The same typo. > > -RUNTIME_DEP_PKGS_CSWlibsoup2dev += CSWlibsoup2 > +PACKAGES += CSWlibsoup2-dev > +CATALOGNAME_CSWlibsoup2-dev = libsoup2_dev > Not necessary. > +SPKG_DESC_CSWlibsoup2-dev += $(DESCRIPTION), development files > +# PKGFILES_CSWlibsoup2-dev += /opt/csw/lib/libsoup-2.4.so > +PKGFILES_CSWlibsoup2-dev += $(PKGFILES_DEVEL) > +RUNTIME_DEP_PKGS_CSWlibsoup2-dev += CSWlibsoup2-4-1 > +RUNTIME_DEP_PKGS_CSWlibsoup2-dev += CSWlibsoup-gnome2-4-1 > +OBSOLETED_BY_CSWlibsoup2-dev = CSWlibsoupdevel > +CATALOGNAME_CSWlibsoupdevel = libsoup_devel_stub > > - > -PKGFILES_CSWlibsoup2dev = /opt/csw/lib/pkgconfig/.* > -PKGFILES_CSWlibsoup2dev += /opt/csw/include/.* > - > - > -EXTRA_INC = /opt/csw/X11/include > -EXTRA_LIB = /opt/csw/X11/lib > +# EXTRA_INC = /opt/csw/X11/include > +# EXTRA_LIB = /opt/csw/X11/lib > EXTRA_PKG_CONFIG_DIRS = $(pre
Re: [csw-devel] SF.net SVN: gar:[19301] csw/mgar/pkg/gimp/trunk/Makefile
2012/9/27 > Revision: 19301 > http://gar.svn.sourceforge.net/gar/?rev=19301&view=rev > Author: cgrzemba > Date: 2012-09-27 10:08:15 + (Thu, 27 Sep 2012) > Log Message: > --- > gimp/trunk: add checkpkg overrides > I would suggest outlining in the commit description, what was it exactly that was overridden. For example "binaries contain the /usr/share string, but it's okay". ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[19689] csw/mgar/pkg/libdvdread/trunk/Makefile
[+pfelecan] In case Peter doesn't subscribe the devel list. 2012/11/14 Igor Galić > > +++ csw/mgar/pkg/libdvdread/trunk/Makefile2012-11-14 09:17:53 UTC > > (rev 19689) > > @@ -3,6 +3,7 @@ > > NAME = libdvdread > > VERSION = 4.2.0 > > GARTYPE = v2 > > +CATEGORIES = lib > > All categories (that were basically empty) have been removed. > This has been announced on the list. Please run mgar up > > i > > -- > Igor Galić > > Tel: +43 (0) 664 886 22 883 > Mail: i.ga...@brainsware.org > URL: http://brainsware.org/ > GPG: 6880 4155 74BD FD7C B515 2EA5 4B1D 9E08 A097 C9AE > > ___ > devel mailing list > devel@lists.opencsw.org > https://lists.opencsw.org/mailman/listinfo/devel > ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[19925] csw/mgar/pkg/perl/trunk/Makefile_5.16. 2_32+64b_privlib
Hi Claudio, Why are you overriding file collisions? You're also overriding filename-version-does-not-match-pkginfo-version - this means that you have an inconsistency between the pkginfo file and the file name. This is not good. Maciej 2012/12/20 : > Revision: 19925 > http://gar.svn.sourceforge.net/gar/?rev=19925&view=rev > Author: claudio_sf > Date: 2012-12-20 13:01:41 + (Thu, 20 Dec 2012) > Log Message: > --- > /home/claudio/opencsw/perl/trunk: overrides > > Modified Paths: > -- > csw/mgar/pkg/perl/trunk/Makefile_5.16.2_32+64b_privlib > > Modified: csw/mgar/pkg/perl/trunk/Makefile_5.16.2_32+64b_privlib > === > --- csw/mgar/pkg/perl/trunk/Makefile_5.16.2_32+64b_privlib 2012-12-20 > 08:19:37 UTC (rev 19924) > +++ csw/mgar/pkg/perl/trunk/Makefile_5.16.2_32+64b_privlib 2012-12-20 > 13:01:41 UTC (rev 19925) > @@ -51,8 +51,57 @@ > #OBSOLETED_BY_CSWperldoc += CSWpm-module-metadata > #OBSOLETED_BY_CSWperldoc += CSWpm-cpan-meta-requirements > #OBSOLETED_BY_CSWperldoc += CSWpm-perl-ostype > +CHECKPKG_OVERRIDES_CSWperl += > file-collision|/opt/csw/bin/config_data|CSWperl|CSWpm-module-build > +CHECKPKG_OVERRIDES_CSWperl += > file-collision|/opt/csw/bin/json_pp|CSWperl|CSWpm-json-pp > +CHECKPKG_OVERRIDES_CSWperl += > filename-version-does-not-match-pkginfo-version|filename=5.16.2,REV=2012.12.19|pkginfo=5.16.2,REV=2012.12.20 > +CHECKPKG_OVERRIDES_CSWperldoc += > file-collision|/opt/csw/share/man/man3/CPAN::Meta::Spec.3perl|CSWperldoc|CSWpm-cpan-meta > +CHECKPKG_OVERRIDES_CSWperldoc += > file-collision|/opt/csw/share/man/man3/CPAN::Meta::Validator.3perl|CSWperldoc|CSWpm-cpan-meta > +CHECKPKG_OVERRIDES_CSWperldoc += > file-collision|/opt/csw/share/man/man3/Module::Build::Notes.3perl|CSWperldoc|CSWpm-module-build > +CHECKPKG_OVERRIDES_CSWperldoc += > file-collision|/opt/csw/share/man/man1/config_data.1|CSWperldoc|CSWpm-module-build > +CHECKPKG_OVERRIDES_CSWperldoc += > file-collision|/opt/csw/share/man/man3/Module::Build::Platform::darwin.3perl|CSWperldoc|CSWpm-module-build > +CHECKPKG_OVERRIDES_CSWperldoc += > file-collision|/opt/csw/share/man/man3/Module::Build::Platform::cygwin.3perl|CSWperldoc|CSWpm-module-build > +CHECKPKG_OVERRIDES_CSWperldoc += > file-collision|/opt/csw/share/man/man3/CPAN::Meta.3perl|CSWperldoc|CSWpm-cpan-meta > +CHECKPKG_OVERRIDES_CSWperldoc += > file-collision|/opt/csw/share/man/man3/JSON::PP.3perl|CSWperldoc|CSWpm-json-pp > +CHECKPKG_OVERRIDES_CSWperldoc += > file-collision|/opt/csw/share/man/man3/CPAN::Meta::Prereqs.3perl|CSWperldoc|CSWpm-cpan-meta > +CHECKPKG_OVERRIDES_CSWperldoc += > file-collision|/opt/csw/share/man/man3/Module::Build::Platform::VOS.3perl|CSWperldoc|CSWpm-module-build > +CHECKPKG_OVERRIDES_CSWperldoc += > file-collision|/opt/csw/share/man/man3/Module::Build::YAML.3perl|CSWperldoc|CSWpm-module-build > +CHECKPKG_OVERRIDES_CSWperldoc += > file-collision|/opt/csw/share/man/man3/Module::Build::Base.3perl|CSWperldoc|CSWpm-module-build > +CHECKPKG_OVERRIDES_CSWperldoc += > file-collision|/opt/csw/share/man/man3/Module::Build::Platform::MPEiX.3perl|CSWperldoc|CSWpm-module-build > +CHECKPKG_OVERRIDES_CSWperldoc += > file-collision|/opt/csw/share/man/man3/Module::Build::Platform::os2.3perl|CSWperldoc|CSWpm-module-build > +CHECKPKG_OVERRIDES_CSWperldoc += > file-collision|/opt/csw/share/man/man3/Module::Build::Platform::RiscOS.3perl|CSWperldoc|CSWpm-module-build > +CHECKPKG_OVERRIDES_CSWperldoc += > file-collision|/opt/csw/share/man/man1/json_pp.1|CSWperldoc|CSWpm-json-pp > +CHECKPKG_OVERRIDES_CSWperldoc += > file-collision|/opt/csw/share/man/man3/Module::Build::Platform::EBCDIC.3perl|CSWperldoc|CSWpm-module-build > +CHECKPKG_OVERRIDES_CSWperldoc += > file-collision|/opt/csw/share/man/man3/Module::Build::Authoring.3perl|CSWperldoc|CSWpm-module-build > +CHECKPKG_OVERRIDES_CSWperldoc += > file-collision|/opt/csw/share/man/man3/Parse::CPAN::Meta.3perl|CSWperldoc|CSWpm-parse-cpan-meta > +CHECKPKG_OVERRIDES_CSWperldoc += > file-collision|/opt/csw/share/man/man3/Module::Build::Version.3perl|CSWperldoc|CSWpm-module-build > +CHECKPKG_OVERRIDES_CSWperldoc += > file-collision|/opt/csw/share/man/man3/CPAN::Meta::YAML.3perl|CSWperldoc|CSWpm-cpan-meta-yaml > +CHECKPKG_OVERRIDES_CSWperldoc += > file-collision|/opt/csw/share/man/man3/Module::Build::Platform::VMS.3perl|CSWperldoc|CSWpm-module-build > +CHECKPKG_OVERRIDES_CSWperldoc += > file-collision|/opt/csw/share/man/man3/Module::Build::Platform::aix.3perl|CSWperldoc|CSWpm-module-build > +CHECKPKG_OVERRIDES_CSWperldoc += > file-collision|/opt/csw/share/man/man3/Module::Build::PPMMaker.3perl|CSWperldoc|CSWpm-module-build > +CHECKPKG_OVERRIDES_CSWperldoc += > file-collision|/opt/csw/share/man/man3/Module::Build::API.3perl|CSWperldoc|CSWpm-module-build > +CHECKPKG_OVERRIDES_CSWperldoc += > file-collision|/opt/csw/share/man/man3/Module::Metadata.3perl|CSWperldoc|CSWpm-module-metadata > +CHECKPKG_OV
Re: [csw-devel] SF.net SVN: gar:[20019] csw/mgar/pkg/texlive/trunk/Makefile
2013/1/4 : > Revision: 20019 > http://gar.svn.sourceforge.net/gar/?rev=20019&view=rev > Author: pfelecan > Date: 2013-01-04 08:09:19 + (Fri, 04 Jan 2013) > Log Message: > --- > texlive/trunk: hopefully last adjustments, for this cycle, suggested > by checkpkg and coping with the mess in reinplace. > > Modified Paths: > -- > csw/mgar/pkg/texlive/trunk/Makefile > > Modified: csw/mgar/pkg/texlive/trunk/Makefile > === > --- csw/mgar/pkg/texlive/trunk/Makefile 2013-01-03 20:16:05 UTC (rev 20018) > +++ csw/mgar/pkg/texlive/trunk/Makefile 2013-01-04 08:09:19 UTC (rev 20019) > @@ -444,6 +444,18 @@ > CHECKPKG_OVERRIDES_CSWtexlive-common += > file-with-bad-content|/usr/share|root/opt/csw/share/texmf-dist/doc/latex/turnthepage/Makefile > CHECKPKG_OVERRIDES_CSWtexlive-common += > file-with-bad-content|/usr/share|root/opt/csw/share/texmf-dist/scripts/bundledoc/arlatex > CHECKPKG_OVERRIDES_CSWtexlive-common += > file-with-bad-content|/usr/share|root/opt/csw/share/texmf-dist/scripts/fragmaster/fragmaster.pl > +CHECKPKG_OVERRIDES_CSWtexlive-common += > file-with-bad-content|/usr/share|root/opt/csw/share/man/man1/chkweb.1 > +CHECKPKG_OVERRIDES_CSWtexlive-common += > file-with-bad-content|/usr/share|root/opt/csw/share/man/man1/chktex.1 > +CHECKPKG_OVERRIDES_CSWtexlive-common += > file-with-bad-content|/usr/share|root/opt/csw/share/texmf/doc/man/man1/chkweb.1 > +CHECKPKG_OVERRIDES_CSWtexlive-common += > file-with-bad-content|/usr/share|root/opt/csw/share/texmf/doc/man/man1/chktex.1 > +CHECKPKG_OVERRIDES_CSWtexlive-common += > file-with-bad-content|/usr/share|root/opt/csw/share/texmf/doc/man/man1/purifyeps.1 > +CHECKPKG_OVERRIDES_CSWtexlive-common += > file-with-bad-content|/usr/share|root/opt/csw/share/texmf/doc/man/man1/pdfjam.1 > +CHECKPKG_OVERRIDES_CSWtexlive-common += > file-with-bad-content|/usr/share|root/opt/csw/share/texmf/doc/man/man1/arlatex.1 > +CHECKPKG_OVERRIDES_CSWtexlive-common += > file-with-bad-content|/usr/share|root/opt/csw/share/texmf/doc/info/eplain.info > +CHECKPKG_OVERRIDES_CSWtexlive-common += > file-with-bad-content|/usr/share|root/opt/csw/share/texmf/doc/info/dvips.info Is there a need to list each file separately in this case? You could just write: CHECKPKG_OVERRIDES_CSWtexlive-common += file-with-bad-content ...and that would match all the file-with-bad-content error tags in texlive-common. ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[20019] csw/mgar/pkg/texlive/trunk/Makefile
2013/1/4 Peter FELECAN : >> Is there a need to list each file separately in this case? You could just >> write: >> >> CHECKPKG_OVERRIDES_CSWtexlive-common += file-with-bad-content >> >> ...and that would match all the file-with-bad-content error tags in >> texlive-common. > > Thank you. This is exactly what I looked after but didn't succeed to > find. Where is this documented? The wiki, section Overrides → in GAR. http://wiki.opencsw.org/checkpkg#toc7 ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[20027] csw/mgar/gar/v2
Yann, this change is not a merge, it reverts all the recent changes. Contact me ASAP to get this sorted out. ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[20039] csw/mgar/gar/v2/lib/python/testdata
So we missed them after all? LOL ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[20089] csw/mgar/gar/v2/lib/python/inspective_package.py
2013/1/11 : > -ldd_output[binary] = ldd_info > > + ldd_output[binary] = ldd_info > + > return ldd_output Could you also include a unit test? ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: [csw-devel] SF.net SVN: gar:[20219] csw/mgar/pkg
2013/1/25 : > WHOLE TREE: Delete again svn:externals Can it be that mgar newpkg is still creating them? ___ devel mailing list devel@lists.opencsw.org https://lists.opencsw.org/mailman/listinfo/devel
Re: Package GAR status
[+devel for some posterity] 2015-01-04 11:08 GMT+00:00 Maciej (Matchek) Bliziński : > Hi Dago, > > Is this report up to date? > > http://www.opencsw.org/package-gar-status.html > > If not, can we make it be up to date? Where's the code? > > Maciej >
