Age | Commit message (Collapse) | Author |
|
since --set-upstream is deprecated.
|
|
|
|
Move setting of the upstream branch out of the loop. This worked by
accident.
|
|
instead of the deprecated _git_getoutput
|
|
instead of the deprecated _git_inout and clarify the return codes and
exceptions raised.
|
|
No need to leak fds to the child.
|
|
We should set stdin to subprocess.PIPE when piping input according
to the subprocess docs.
|
|
It's shorter and get's rid of the unused dummy variable.
|
|
|
|
There's no need to fill the parser with all the built in defaults since
we can simply copy them to the config from the class dict. This allows
us to first set the values from the old {git,gbp}-* commands and then
only overwrite changed values from the newer "gbp <command>" versions.
Otherwise we'd overwite all old style config with the new style one.
This also fixes the problem where the default section wouldn't be read
at all if the command doesn't even have an empty section in any gbp.conf
file.
|
|
instead of git-import-dsc
|
|
and robustify against invalid modules names.
|
|
|
|
For backward compatibility between {gbp,git}-<subcommand> and "gbp
<subcommand>" make sure we parse the former sections if using the later.
|
|
and not the one from the OptionParser. The instance's default dict this
gets reset to empty when invoking OptionParser.__init__.
|
|
So like git you can now use gbp <command> instead of git-<command> or
gbp-<command>. The manpages and docs aren't adjusted yet.
|
|
Allow diffing a single object, e.g. getting diff from single commit.
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
|
|
Git-Dch: Ignore
|
|
This kills another _git_getoutput. Also deprecate the method since we
don't need to special case the subject.
|
|
Utilize the get_commit_info() method of GitRepository instead of the
deprecated get_subject().
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
|
|
Update working copy and index (to branch HEAD) if we modify the
currently checked-out branch. Otherwise the repository is left in
unclean state when the current branch is upstream or pristine-tar and
the '--no-merge' option is used.
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
|
|
Add a new 'patchname' field to the information returned by
get_commit_info. That is, the subject in a sanitized format, similar to
what git-format-patch uses.
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
|
|
To be consistent with the date property and for easier compatibility
with git dates.
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
|
|
Signed-off-by: Guido Günther <agx@sigxcpu.org>
Closes: #708636
|
|
This fixes the doctest and also makes sure we use an up to date manpage.
Thanks: Daniel Dehennin for the report
|
|
|
|
to suppress the confusiong "fatal: ref HEAD is not a symbolic ref". We
alrady put out a more descriptive error message. This makes using
--ignore-branch look far less dangerous.
|
|
This allows us to fetch upstream and pristine-tar even when in detached
head state. This can be usefull when e.g. building in jenkins which uses
a detached head by default.
|
|
This fixes a problem when running git-dch form a detached HEAD
|
|
and add the error detail to the exception message. This silences the
testsuite and makes sure we have error message and exception output in
sync.
|
|
so we can capture stderr and use it in error messages intead of spewing
it on the console.
|
|
so we have the object creation close to the object itself.
|
|
to avoid accidental usage. Use DebianSource.is_native instead.
|
|
instead of ChangeLog.is_native since the former used
debian/source/format in favor of debian/changelog.
|
|
This reverts commit 0c582cd827f164168a313e0bb4eabf9487bdd088.
|
|
so we can caputure stderr and pass it along with any errors.
|
|
to avoid spurious output on the console and add the error message to the
exception intead.
|
|
and make sure the callers forward the error properly
|
|
but add any error output to the exception instead. All callers have been
audited to properly output the message if needed.
|
|
and use that in the doctest to get rid of the spurious output
|
|
since we don't expect it to succeed.
|
|
None of the gbp script rely on the output being printed. In fact we
improve the exception by appending the stderr output.
|
|
It's not used within gbp itself so no command line tool expects
an error message to be printed.
|
|
This also silences 10_test_get_upstream_tree
|
|
For enabling/disabling all hooks. This option does not affect the
builder command, though.
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
|
|
Determine from changelog if debian/source/format does not provide any
type (1.0 format).
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
|
|
Provides more output options than find_tag().
- longfmt: for getting the tag name in the long format (tag, number of
commits and sha1)
- always: for falling back to sha1 if no (matching) tag name is found
- abbrev: for defining the length of sha1 returned
Change GitRepository.find_tag() to use the the new method.
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
|
|
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
|
|
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
|
|
by using _git_inout instead of GitCommand
|