Was fine with 6.7 last time tried it, may be missing a backport
but is not worth looking into for 6.6 which is almost over.
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
Just as a quick workaround until maintainer(s) have time.
No need for a revbump given binding operator will cause it
to recheck dependencies when attempting rebuild.
Bug: https://bugs.gentoo.org/927383
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
libgit2-sys crate has a upper bound for <1.8.0 and refuses
to build otherwise, but not seeing(?) any issues with it.
Just a quick emergency fix, will likely be fixed next libgit2-sys
version. Kept full version in the sed's path rather than a glob
to force re-checking this when the crate is bumped.
Alternatively could drop LIBGIT2_NO_VENDOR or set an upper bound
in RDEPEND, but does not seem necessary.
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
libgit2-sys crate has a upper bound for <1.8.0 and refuses
to build otherwise, but not seeing(?) any issues with it.
Just a quick emergency fix, will likely be fixed next libgit2-sys
version. Kept full version in the sed's path rather than a glob
to force re-checking this when the crate is bumped.
Alternatively could drop LIBGIT2_NO_VENDOR or set an upper bound
in RDEPEND, but does not seem necessary.
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
libgit2-sys crate has a upper bound for <1.8.0 and refuses
to build otherwise, but not seeing(?) any issues with it.
Just a quick emergency fix, will likely be fixed next libgit2-sys
version. Kept full version in the sed's path rather than a glob
to force re-checking this when the crate is bumped.
Alternatively could drop LIBGIT2_NO_VENDOR or set an upper bound
in RDEPEND, but does not seem necessary.
Note 9999 is still broken albeit hopefully the crate will just be
fixed upstream in time. 9999 users could keep old libgit2 meanwhile.
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
libgit2-sys crate has a upper bound for <1.8.0 and refuses
to build otherwise, but not seeing(?) any issues with it.
Just a quick emergency fix, will likely be fixed next libgit2-sys
version. Kept full version in the sed's path rather than a glob
to force re-checking this when the crate is bumped.
Alternatively could drop LIBGIT2_NO_VENDOR or set an upper bound
in RDEPEND, but does not seem necessary.
Signed-off-by: Ionen Wolkens <ionen@gentoo.org>
Upstream is libdap. They exist. They have lots of maintained software,
it is simply that oc isn't one of them as far as I can tell.
What happened to it? I don't know. Maybe it's legacy abandoned software.
Their distfiles server still has it, but that's it.
Possibly we should delete the package instead...
Closes: https://bugs.gentoo.org/862906
Signed-off-by: Eli Schwartz <eschwartz93@gmail.com>
Signed-off-by: Sam James <sam@gentoo.org>
LTO makes the compiler better able to spot this. But it isn't something
the entire build should fail due to, not for stable release tags of
existing software.
Closes: https://bugs.gentoo.org/862453
Signed-off-by: Eli Schwartz <eschwartz93@gmail.com>
Signed-off-by: Sam James <sam@gentoo.org>
For several reasons, this effort to control examples/docs via sed'ing
out Makefile.in did not work:
- the seds didn't apply
- the Makefile.in was ignored because it was regenerated via eautoreconf
... and also useless our outright wrong. The examples are 548kb, and the
docs are 156kb. Small files shouldn't be gated on USE flags.
But also, the docs USE flag was attempting to remove the part of the
build system that caused man pages to be shipped, in addition to the
HTML versions of the man pages. This must never be done, manpages are
too important. See QA Policy Guide 0305.
Signed-off-by: Eli Schwartz <eschwartz93@gmail.com>
Signed-off-by: Sam James <sam@gentoo.org>
I would like to report this upstream. However, the only contact method
is their mailing list. And their mailing list may or may not work -- I
find it hard to tell because I've never seen the interface before, but
it is mostly in French and the bit that I recognized doesn't seem to
have working links to e.g. the archive.
Closes: https://bugs.gentoo.org/862903
Signed-off-by: Eli Schwartz <eschwartz93@gmail.com>
Signed-off-by: Sam James <sam@gentoo.org>
This is the 2006 edition of the software. It has a bunch of useful
fixes:
- it fixes Modern C porting issues
- it fixes LTO errors
- it incorporates the glibc-2.10 patch
just by upgrading! It's about time we upgrade, since it is 18 years
later.
Along the way we fix up the ebuild a bit:
- bump to EAPI 8
- acquire an actual SRC_URI, not just the gentoo mirror
- update the homepage along with the SRC_URI. The original homepage is
dead, and the domain redirects to a new domain (whence comes the new
distfile).
It has some shortcomings too:
- the distfile comes with free .svn directories, which have to be
pruned.
- the distfile comes without the 80mb demos
To solve this, we use the old distfile (which I think only exists on
Gentoo's mirror network) and gate it behind a USE=examples flag, in case
anyone needs them.
Closes: https://bugs.gentoo.org/862285
Closes: https://bugs.gentoo.org/884283
Signed-off-by: Eli Schwartz <eschwartz93@gmail.com>
Signed-off-by: Sam James <sam@gentoo.org>
Andreas Sturmlechner
Ionen Wolkens
Matt Turner
Michał Górny
Oliver Freyermuth
Miroslav Šulc
Yixun Lan
Hans de Graaff
Joonas Niilola
Sam James
Eli Schwartz
Alexander Kurakin
orbea