Instead of placing the distro name (e.g. 'Fedora Linux') into the
version.in file, use the --with-pkgconfig configure option to set the
disto name.
This does change GDB's version output slightly. Prior to this commit
we might expect to see something like this:
$ gdb --version
GNU gdb (GDB) Fedora Linux 13.1-1-fc38
...
Now we'll see this:
$ gdb --version
GNU gdb (Fedora Linux) 13.1-1-fc38
...
Notice that 'Fedora Linux' moved inside the parenthesis.
There are other benefits to this change, the Python gdb.VERSION
string will now contain '13.1-1-fc38' instead of 'Fedora Linux
13.1-1-fc38', this was reported as an issue in this bug:
https://bugzilla.redhat.com/show_bug.cgi?id=2179554
I've also removed the change to gdb.gdb/selftest.exp as I don't
believe that this was ever being used -- when this test was run we
should have hit an earlier case.
I have added a brand new test that checks our version string. I've
placed this new test into a separate file, so we should (hopefully)
not have any issues maintaining this as an out of tree test.
For implementation, by default we make use of the %dist_name macro,
however, this is not defined on RHEL 9 or earlier, so to support
reusing this commit on RHEL systems, I also detect the %rhel macro,
and in that case use a hard-coded string.
Update gdb-6.3-rh-testversion-20041202.patch.
Update gdb-6.3-bz140532-ppc-unwinding-test.patch.
Update gdb-6.6-buildid-locate.patch.
Update gdb-6.6-buildid-locate-rpm.patch.
Remove 'Recommends: ' line for gcc-gdb-plugin for BZ2149246.
Add 'define _lto_cflags ' to avoid ODR violations.
Add -Wno-stringop-overflow to --enable-gdb-build-warnings to work around
gcc problem.
...so that $_gdb_major and $_gdb_minor will be obtained correctly.
Also, change define of 'use_guile' from empty (which rpmbuild
complained about) to a value of 1. I encountered this problem
while attempting to build on F37.
Bump 'snapgnulib' date.
Drop gdb-rhbz1818011-bfd-gcc10-error.patch.
Drop gdb-rhbz1822715-fix-python-deprecation.patch.
Drop gdb-rhbz1829702-fix-python39.patch.
Drop gdb-rhbz1838777-debuginfod.patch.
Drop gdb-rhbz1844458-use-fputX_unfiltered.patch.
Drop gdb-rhbz1869484-deleted-working-directory.
Adjust build-id related patches.
Ajust VLA patches.
This commit improves the script used to generate the patches from the
upstream git repo. Basically, it removes even more cruft that was
being added by 'git format-patch', making the output really simple and
deterministic (determinism which was a problem before).
Aside from the metadata, nothing else has been changed and therefore
it doesn't justify a new Fedora GDB release.
- chmod +x on the generate-*.sh script.s
- Remove references to gdb-8.0.1 from 'sources' and '.gitignore'.
- Regenerate first line of the patches (remove commit hash).
- Fix empty Source line.
- New rpm option `--with upstream' to drop the Fedora patches for testing.
- Drop some no longer valid .spec file comments.
- Include the Fortran dynamic arrays entry for %changlog of
6.7.50.20080227-1.
