IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
Before this change, thanks to rpmEVRcmp(), compare_deps() used to treat
EVRs with common EV as equal when one had a release and another hand't.
There were some reasons to use this approach, e.g. to decide that
Requires: %name = %version
is weaker than
Requires: %name = %version-%release
it had firstly to consider them as comparable deps. But, as an
unfortunate consequence of that approach, it was deciding that such deps
are equal, so that addReqProv() was unable to replace former with the
latter.
If one of deps has Epoch and another one hasn't,
we first compare them without Epoch, and if it happens
that they are equal, then the dep that has Epoch wins.
Ignore requirements with RPMSENSE_SENSEMASK containing ~RPMSENSE_EQUAL.
Add strict requirements for requirements on subpackages regardless of
number of alternative providers.
When comparing deps whose senses are identical, Epoch needs special
treatment: if one of deps has Epoch and another one hasn't, we first
compare them without Epoch, and if it happens that they are equal,
the dep that has Epoch wins.
Non-strict interpackage dependencies are now converted to strict when
appropriate.
The %_allowed_nonstrict_interdeps macro support is therefore gone.
This change also downgrades "dependency needs Epoch" error to warning
because it automatically adds strict dependencies.
There is a macro now that can list pairs of allowed non-strict deps.
By default, the macro is not defined so the list is empty and therefore
non-strict deps are not allowed.
In modern rpms both %patch -F <N> and %patch -F<N> are valid option
calls whereas old -F implementation supported -F <N> syntax only.
This patch adds support for %patch -F<N> syntax.
* build/rpmspec.h (OpenFileInfo): Change readBuf to a pointer,
add readBufSize.
(freeOpenFileInfo): New prototype.
* build/spec.c (freeSpec): Initialize readBuf and readBufSize.
(freeOpenFileInfo): New function.
* build/parseSpec.c (readLine): Use getline and freeOpenFileInfo.
(closeSpec): Use freeOpenFileInfo.
SIGPIPE SIG_IGN handler was installed before the fork, which means that,
in autodep scripts, SIGPIPE was ignored as well. This is why in
commands like
cmd1 | cmd2
cmd1 was not killed graceully with SIGPIPE, but instead writing to cmd2
resulted in EPIPE. For which some commands apparently were not ready.
This fixes messages like
/usr/lib/rpm/files.req: line 33: echo: write error: Broken pipe
The check for cycles introduced in 4.0.4-alt100.17-6-ga71e004 appeared
to be incomplete. For example, it fails to recognize a cycle in the
following example:
%package -n test
Requires: libtest = %version-%release
%package -n libtest
Requires: libtest-common = %version-%release
%package -n libtest-common
Requires: libtest = %version-%release
warning: util-linux: non-strict dependency on libblkid
warning: util-linux: non-strict dependency on libuuid
warning: util-linux: non-strict dependency on libmount
warning: mount: non-strict dependency on libblkid
warning: cfdisk: non-strict dependency on libblkid
warning: fdisk: non-strict dependency on libblkid
warning: libblkid: non-strict dependency on libuuid
warning: libblkid-devel: non-strict dependency on libuuid-devel
warning: libmount-devel: non-strict dependency on libblkid-devel
warning: util-linux-debuginfo: non-strict dependency on libblkid-debuginfo
warning: util-linux-debuginfo: non-strict dependency on libuuid-debuginfo
warning: util-linux-debuginfo: non-strict dependency on libmount-debuginfo
warning: mount-debuginfo: non-strict dependency on libblkid-debuginfo
warning: cfdisk-debuginfo: non-strict dependency on libblkid-debuginfo
warning: fdisk-debuginfo: non-strict dependency on libblkid-debuginfo
warning: libblkid-debuginfo: non-strict dependency on libuuid-debuginfo
Note that the right time for this check is after strict dependencies
are propagated, and also after debuginfo dependencies have been lifted.
%package -n test
Requires: libtest = %version-%release
Requires: libtest-common = %version-%release
%package -n libtest
Requires: libtest-common = %version-%release
%package -n libtest-common
Requires: libtest = %version-%release
Here libtest and libtest-common make a cycle, and the "test" package
requires a cycle. Naiive optimization against a cycle can simply
wipe out all dependencies: the dependency on libtest-common will be
optimized out, because libtest-common is already required by libtest;
likewise, the dependency on libtest will be optimized out, because
libtest is already required required by libtest-common.
%_deps_optimization is now treated as optimization level 0, 1, or 2
(the default is 2). Level 1 will make single-package optimizations.
Level 2 will additionally optimize dependencies between subpackages.
Though the intersection check can be useful even for /usr/src/debug
(it indicates binary code duplication and/or poor dependencies between
subpackages), it is too noisy and can eat up quite a bit of output.
The same information is still available directly within the packages
or via contents_index.
Note that checkHardLinks function is now removed. It was unclear
whether it was supposed to verify %lang attributes (returning non-zero
on error) or indicate if all hardlinks are packaged within the package.
It turns out that only a single package in our repo has
PartialHardlinkSets dependency:
$ cd /ALT/Sisyphus/files/x86_64/RPMS/
$ rpm -qp --qf '[%{NAME}\t%{REQUIRENAME}\n]' *.rpm |fgrep 'PartialHardlinkSets'
$ cd /ALT/Sisyphus/files/noarch/RPMS/
$ rpm -qp --qf '[%{NAME}\t%{REQUIRENAME}\n]' *.rpm |fgrep 'PartialHardlinkSets'
freeciv-common rpmlib(PartialHardlinkSets)
$
This probably means that freeciv-common has hardlinks with different
%lang attributes (which probably was supposed to be an error). So
the whole issue should be reconsidered. A leave XXX marks in the
code and suggest new PartialHardlinkSets implementation (however,
the dependency is not being added yet).
When pkg1 has strict dependency on pkg2, all other non-strict/virtual
dependencies on pkg2 will be reaped off pkg1. Note that the are subtle
cases: though I try to make something about merging dependency flags,
it is not always possible, as the dependency on pkg2 can be obtained
via a transitive chain. I see this being a non-critical issue which
is worthwhile to be mentioned.
Basically, if pkg1 has strict dependency on pkg2, the question
is whether pkg1-debuginfo should possess a strict dependency on
pkg2-debuginfo. The answer is "yes" provided that pkg1-debuginfo
is already dependent on pkg2-debuginfo (i.e. has new soname-based
debuginfo dependencies resolved to pkg2-debuginfo). This provides
an opportunity for dependency optimization between dependend
debuginfo subpackages.
The output now is like this:
removing 12 sources from rpm-debuginfo and adding dependency on librpm-debuginfo
removing 11 sources from rpm-debuginfo and adding dependency on librpmbuild-debuginfo
removing 13 sources from rpm-build-debuginfo and adding dependency on rpm-debuginfo
removing 14 sources from rpm-static-debuginfo and adding dependency on rpm-debuginfo
removing 20 sources from librpmbuild-debuginfo and adding dependency on librpm-debuginfo
removing 13 sources from rpm-build-debuginfo and adding dependency on librpm-debuginfo
removing 80 sources from rpm-static-debuginfo and adding dependency on librpm-debuginfo
removing 12 sources from python-module-rpm-debuginfo and adding dependency on librpm-debuginfo
removing 12 sources from rpm-build-debuginfo and adding dependency on librpmbuild-debuginfo
removing 35 sources from rpm-static-debuginfo and adding dependency on librpmbuild-debuginfo
