Re: rpm bug?
On Jul 25, 2007, at 10:05 AM, Elan Ruusamäe wrote: > > perhaps add possibility to query srcheaders to rpm? > Already there rpm -qp foo*.src.rpm What you really want is some means to choose amongst src and multiple binary headers when querying a specfile. Immediately after the request for querying only the srpm header is the request to query only a single subpkg header. The implementation issue is the same even if the usage case is different. The hack to choose 1-of-N headers is at (linenum wrto wrto rpm5.org cvs HEAD) build/spec.c:782 case RPMQV_SPECFILE: for (pkg = spec->packages; pkg != NULL; pkg = pkg->next) { /* If no target was specified, display all packages. * Packages with empty file lists are not produced. */ /* XXX DIEDIEDIE: this logic looks flawed. */ if (target == NULL || pkg->fileList != NULL) xx = qva->qva_showPackage(qva, ts, pkg->header); } break; The srpm header is in spec->sourceHeader. The real design issue is how to pass the 1of-N to the routine from the CLI. hth 73 de Jeff ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm bug?
On Tuesday 24 July 2007 21:02:36 Jeff Johnson wrote: > On Jul 24, 2007, at 1:46 PM, Jakub Bogusz wrote: > > On Fri, Jul 20, 2007 at 11:24:49AM +0200, Tomasz Wittner wrote: > >> On Thu 19. of July 2007, 16:04, Elan Ruusamäe wrote: > >>> On Thursday 19 July 2007 11:44:14 Marcin Król wrote: > > not sure how to put this in proper words, but querying binheader > > results this, one should query srcheaders. > > Anyone brave enough to make requried changes into builder > script? :) > >>> > >>> well. somebody mentioned that builder script should not depend on > >>> any > >>> higher language liker perl so... > >> > >> But why? Better have broken tool than use suitable language? > > > > Better broken in some corner cases than unusable for bootstrap... > > (anyway, redefining Version in middle of .spec is tricky, as you > > can see > > from %{version} behaviour) > > Tricky only because noone has asked. > > I can certainly permit > Version: %%{version} > %define version whatever-you-want > to delay the expansion, and add an additional macro expansion > before adding RPMTAG_VERSION to *.rpm package headers, if that is > desirable. > > A delayed expansion is most definiitely desirable for Release: fields > because of the pesky %{?dist} that has been injected everywhere by > Fedora, and will take years to phase out, sigh. perhaps add possibility to query srcheaders to rpm? > 73 de Jeff -- glen ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm bug?
On Jul 24, 2007, at 1:46 PM, Jakub Bogusz wrote: > On Fri, Jul 20, 2007 at 11:24:49AM +0200, Tomasz Wittner wrote: >> On Thu 19. of July 2007, 16:04, Elan Ruusamäe wrote: >>> On Thursday 19 July 2007 11:44:14 Marcin Król wrote: > not sure how to put this in proper words, but querying binheader > results this, one should query srcheaders. Anyone brave enough to make requried changes into builder script? :) >>> >>> well. somebody mentioned that builder script should not depend on >>> any >>> higher language liker perl so... >> But why? Better have broken tool than use suitable language? > > Better broken in some corner cases than unusable for bootstrap... > (anyway, redefining Version in middle of .spec is tricky, as you > can see > from %{version} behaviour) > Tricky only because noone has asked. I can certainly permit Version: %%{version} %define version whatever-you-want to delay the expansion, and add an additional macro expansion before adding RPMTAG_VERSION to *.rpm package headers, if that is desirable. A delayed expansion is most definiitely desirable for Release: fields because of the pesky %{?dist} that has been injected everywhere by Fedora, and will take years to phase out, sigh. 73 de Jeff ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm bug?
On Fri, Jul 20, 2007 at 11:24:49AM +0200, Tomasz Wittner wrote: > On Thu 19. of July 2007, 16:04, Elan Ruusamäe wrote: > > On Thursday 19 July 2007 11:44:14 Marcin Król wrote: > > > > not sure how to put this in proper words, but querying binheader > > > > results this, one should query srcheaders. > > > > > > Anyone brave enough to make requried changes into builder script? :) > > > > well. somebody mentioned that builder script should not depend on any > > higher language liker perl so... > But why? Better have broken tool than use suitable language? Better broken in some corner cases than unusable for bootstrap... (anyway, redefining Version in middle of .spec is tricky, as you can see from %{version} behaviour) -- Jakub Boguszhttp://qboosh.pl/ ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm bug?
On Thu 19. of July 2007, 16:04, Elan Ruusamäe wrote: > On Thursday 19 July 2007 11:44:14 Marcin Król wrote: > > > not sure how to put this in proper words, but querying binheader > > > results this, one should query srcheaders. > > > > Anyone brave enough to make requried changes into builder script? :) > > well. somebody mentioned that builder script should not depend on any > higher language liker perl so... But why? Better have broken tool than use suitable language? -- Tomasz Wittner ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm bug?
On Thursday 19 July 2007 11:44:14 Marcin Król wrote: > > not sure how to put this in proper words, but querying binheader results > > this, one should query srcheaders. > > Anyone brave enough to make requried changes into builder script? :) well. somebody mentioned that builder script should not depend on any higher language liker perl so... -- glen ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm bug?
> not sure how to put this in proper words, but querying binheader results > this, > one should query srcheaders. Anyone brave enough to make requried changes into builder script? :) M. ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm bug?
On Thursday 19 July 2007 10:50, Marcin Król wrote: > Maybe this is feature required for something else to work, I don't know. > For me its a bug which should be nailed :) not sure how to put this in proper words, but querying binheader results this, one should query srcheaders. see how this script works: http://cvs.pld-linux.org/cgi-bin/cvsweb.cgi/test/specinfo.pl [EMAIL PROTECTED] pld/test $ ./specinfo.pl ~/rpm/pld/SPECS/mozilla-firefox.spec PACKAGE_NAME mozilla-firefox PACKAGE_VERSION 2.0.0.4 PACKAGE_RELEASE 1.1 mozilla-firefox-2.0.0.4-1.1 /home/glen/rpm/pld/SRPMS/mozilla-firefox-2.0.0.4-1.1.src.rpm /home/glen/rpm/pld/RPMS/mozilla-firefox-2.0.0.4-1.1.i686.rpm /home/glen/rpm/pld/RPMS/mozilla-firefox-libs-2.0.0.4-1.1.i686.rpm /home/glen/rpm/pld/RPMS/mozilla-firefox-lang-en-2.0.0.4-1.1.i686.rpm /home/glen/rpm/pld/RPMS/mozilla-firefox-addon-tidy-0.8.3.9-1.1.i686.rpm just using %dump you'll get last binpkg. -- glen ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
rpm bug?
Hello. I'm not good in uderstanding how rpm internally works so don't blam me if I'm writing obvious things :) I think I've found bug in .spec processing. I'll show that on example. Our builder script is using following command to get package name and version for auto- CVS tag: rpmbuild --macros /usr/lib/rpm/macros:/usr/lib/rpm/i686-linux/macros:/etc/rpm/macros.*:/etc/rpm/macros:/etc/rpm/i686-linux/macros:~/etc/.rpmmacros:~/.rpmmacros:/home/users/krol/.builder-rpmmacros --nodigest --nosignature --nobuild --define 'prep %{echo:dummy: PACKAGE_NAME %{name} }%dump' --nodeps mozilla-firefox.spec Main package in mozilla-firefox.spec has version 2.0.0.5 however above command returns PACKAGE_VERSION = 0.8.4.0 which is version of mozilla-firefox-tidy subpackage. Since it is last subpackage defined in spec it seems like rpm just take last "Version: something" as package version instead of using version of main package. Maybe this is feature required for something else to work, I don't know. For me its a bug which should be nailed :) M. ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en