| Commit message (Collapse) | Author | Age |
|
|
|
| |
Closes: #851844
|
|
|
|
| |
to make flake8 happy
|
|
|
|
|
|
|
|
| |
This makes sure we use the timestamps of the commit when creating the
tarball. Only applies when not using pristine-tar
Thanks: Ximin Luo for investigating
Closes: #851645
|
|
|
|
|
|
|
|
| |
This allows compressors to use the their default compression level. Only
applies when not using pristine-tar.
Thanks: Antoine Beaupré for investigating
Closes: #820846
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
If we filter a tarball and use filter-pristine-tar we want pristine-tar
to see the correct file name. If a file already exists at that location
move it.
If we wouldn't move the existing tarball but rather e.g. create the
symlink in a temp directory we risk gbp buildpackage picking up the
unfiltered tarball later.
Closes: #558777
|
|
|
|
|
| |
Closes: #850869
Thanks: Simon McVittie
|
|
|
|
|
|
|
|
| |
Instead of unpacking them right away first repack the filtered tarball
so the component tarballs don't end up in the main tarball when using
filters.
Closes: #840602
|
|
|
|
| |
Git-Dch: Ignore
|
|
|
|
| |
Git-Dch: Ignore
|
|
|
|
| |
since we want to avoid the same list being reused
|
|
|
|
| |
Git-Dch: Ignore
|
|
|
|
|
| |
Although none of the callers returns None but raises an exception this
is more defensive.
|
|
|
|
|
|
| |
when importing packages.
Closes: #577810
|
|
|
|
|
| |
We don't iterate over all individual sections but rather let
dpkg-parsechangelog do the hard work.
|
|
|
|
|
|
|
|
|
|
| |
to honor --[no]-git-author
This basically reverts c3258c19c6cfa291c8bebac1dae230f6e3ccca36 but also
adds a test case to not break it again that easily.
Closes: #796913
Thanks: David Kalnischkies for investigating
|
|
|
|
| |
Thanks: Andrea Zagli for the suggestion
|
| |
|
|
|
|
|
|
| |
We're setting the dist to empty in this case since git-pbuilder will the
use either base.cow or base-sid.cow but we should print a proper dist to
the user.
|
|
|
|
|
|
|
|
| |
So far we've been avoiding a pipe for input on empty strings. This would
make commands as 'git mktree' hang forever waiting for standard input.
So use a proper pipe on '' and skip it only for None.
Closes: #850319
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
otherwise 'gbp import-dsc' adds reflog entries without a description.
|
| |
|
| |
|
|
|
|
|
|
| |
If the user running the tests has e.g. a home directory of
/git-buildpackage we'd end up with '/etcHOME/gbp.conf' when replacing on
arbitrary positions. Make the replacement anchored instead.
|
|
|
|
| |
instead of spewing the exception on the console.
|
|
|
|
| |
No need to parse --download of URL is given or it's not a dsc file
|
| |
|
|
|
|
|
| |
This introduces tests that reach out to the network. These are disabled
by default.
|
| |
|
|
|
|
|
|
|
|
| |
So far we would fail empty commits with the confusing error
$ gbp pq export --commit
gbp:info: Generating patches from git (debian/sid..patch-queue/debian/sid)
gbp:error: Error running git commit:
|
| |
|
|
|
|
| |
Don't modify existing ones.
|
|
|
|
|
|
| |
Signed-off-by: Guido Günther <agx@sigxcpu.org>
Close: #845536
|
| |
|
|
|
|
| |
Gbp-Dch: Ignore
|
|
|
|
| |
so we can query things like sections at a later point.
|
| |
|
|
|
|
| |
Gbp-Dch: Ignore
|
|
|
|
| |
Git-Dch: Ignore
|
|
|
|
|
|
| |
Upstream tags are without epoch so we need to filter that out too.
Thanks: Tzafrir Cohen
|
| |
|
|
|
|
|
|
|
|
|
| |
The %patch macro of rpm supports the flag -F<num>, which seems to be
similar to that of patch(1): allow patches with a specific fuzz.
Parse the option but ignore it since we don't want patches with fuzz.
Closes: #846479
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The parser uses optparse and gave confusing error messages on failure
like:
Usage: buildpackage_rpm.py [options]
buildpackage_rpm.py: error: no such option: -F
This now looks like
Usage: gbp.rpm internal patch macro opts parser for -F2
gbp.rpm internal patch macro opts parser: error: no such option: -F
|
| |
|