Make requirement on rpm-build strong on OpenSUSE #1
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Git-buildpackage-rpm now always updates the 'VCS:' tag in the exported spec file. A new config option 'spec-vcs-tag' controls the format: - if empty, no 'VCS' tag is inserted and possible old 'VCS' tag is removed - otherwise, a 'VCS' tag is inserted or the old 'VCS' tag is updated - '%(tag)s' expands to the long tag name (from git-describe) Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Add a new commandline option to ignore untracked files. Untracked files are ignored and do not cause an error. However, an error is given if there are changes to tracked files. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Add support for reading the local config file(s) from a given git tree-ish. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Read the local gbp config (.gbp.conf, debian/gbp.conf) from the exported tree instead of the current working copy. This makes sure that we build correctly even if building something else than the current working copy. 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>
Use the --all option of git-add so that all tracked files are updated in all conditions. Previously deletion of tracked files was not staged, for example. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
With this option you can either only update already tracked files to index the (untracked=False) or add new files, too. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Whether to ignore untracked files (untracked=False) or not (untracked=True). Now, clones the temporary index file from the actual index, in order to correctly update files. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com> Signed-off-by: Ed Bartosh <eduard.bartosh@intel.com>
Add support for building different kind of "working copies", when using the --git-export option: - 'WC.TRACKED': only take files that are already tracked - 'WC.UNTRACKED': take untracked files, too - 'WC.IGNORED': also add files that'd normally be ignored Using '--git-export=WC' behaves like before. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Add new possible keywords to be used in packaging-tag format string: 'nowtime', 'authortime', 'committime', 'nowtimenum', 'authortimenum' and 'committimenum'. The '*timenum' keyword denote the corresponding '*time' keyword appended with an additional incremental serial number. E.g. if 'nowtime' would produce '20120531', 'nowtimenum' would (the first time) produce '20120531.1'. If a new tag would be created on the same date, the field would expand to '20120531.2' etc. What is completely missing is a way to support these new tag keywords in git-import-srpm tool. So, if you use these tags, git-import-srpm is not able to reproduce the tags, (and, it is not necessarily able to correctly tell if you've already imported a certain version of the package). Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Redirect stdout and stderr of the (child) command to sys.stdout and sys.stderr of the caller, respectively. This change is mainly for the unit tests. It makes Python nose to correctly capture the output of the child command, too, which in turn suppresses a lot of spurious output when running nosetests. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Reduces spurious output from rpmbuild. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Separate keywords for packaging and upstream tag names. Also, add more tests for tagging options. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
The name of the package containing zipmerge varies, e.g. 'libzip' in Fedora 22 vs. 'libzip-tools' in Fedora 23. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Corresponding the --skip-debian-tag options of import-dsc. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Prevent a crash when the author (for a raw diff) was None because no name/email could be determined from git config or environment variables. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Add some missing (build) dependencies if unit tests are enabled as part of the build process. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Similar to what the option does in git-import-orig-[rpm]. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
This updates all remote-tracking branches (for the remote that is fetched from) whose local branch name is identical to the remote branch name. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Makes it possible to only checkout certain files from a branch, instead of switching to the branch. Add a new method instead of extending checkout() in order to keep it consistent. That is, otherwise checkout() would have totally different outcome depending on whether paths were defined of not. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Treat non-checked-out branches similarly to the current branch when forcing update. That is, do git merge, instead of just setting the ref. Also, renames fast_forward_branch() to update_branch() to better match the functionality. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
The new option can be used to "copy" files from the packaging branch to the patch-queue branch when doing pq import. The files defined with this option will appear as new files in one monolithic commit in the development/patch-queue branch. By default, the local gbp conf files are imported in order to try to ensure that gbp sees the same settings on the patch-queue branch as on the packaging branch. NOTE: This option does not affect the importing (i.e. apply and commit) of patches. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
This allows initialization of a GitRepository object, even if the current working directory (or path given to GitRepository) is not the top level directory of the git repository. Don't guess the git meta data dir, but, take it as reported by git itself. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Add a new _git_inout2() helper method that returns the git output (stdout) as a generator - instead of all stdout data in one string. Useful for handling git commands that are expected to have a lot of stdout data, like git-archive. Also, changes the private __git_inout() method to return a generator. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Return tar data as a generator object, if the 'output' option is not defined. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Make dump_tree() utilize the GitRepository.archive() method - similarly to git_archive_submodules() - instead of ad-hoc Python pipes. This makes dump_tree() work independent of the callers current working directory. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
This should make the usage of gbp-pq more flexible and prevent mistakes caused by e.g. running gbp-pq import under some subdirectory. 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>
Add tests for features that didn't have one, yet. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Three split packages: git-buildpackage-{common,rpm,doc}
Signed-off-by: Junchun Guan <junchunx.guan@intel.com>
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Better compatibility with 3rd party modules that have their own logging initializations. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
This is the first tool in an effort of enabling gbp in the BitBake build environment. Gbp-import-bb is a tool for importing packages from a BitBake-based "combined" distro repository into individual per-package Git repositories. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
This is a tool for managing patch-queues for packages maintained in the BitBake packaging format (.bb recipes). Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Initial version of the tool for building BitBake packages from Git. NOTE: The buildpackage-bb tool itself is able to operate even without an initialized BitBake build environment although the build likely fails in this case. However, this makes it possible to export the packaging meta data, for example. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
This is a new tool for helping to clone remote per-package Git repositories when working in BitBake-based "full distro" build environment. This is useful in the case that individual packages are actually maintained in per-package Git repositories (like Tizen). That is, the full distro repository that the developer operates in is composed of the packaging meta data from the individual per-package repositories. When willing to contribute to a package the developer would use clone-bb to clone the correct per-package repository and make his changes there. NOTE: clone-bb uses GBP_PACKAGING_REPO variable to determine the remote repository URI. This variable should be defined in the package recipes in order to make clone-bb usable. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
For forcing the creation of annotated tags. Causes the an editor to be spawned if no message is given. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
This is a Tizen-specific tool for creating and pushing special submit tags. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Change-Id: I3123ca6f024dcdd380e745f3bbadaba3f5bfc098 Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Old versions of python-six shipped on e.g. Debian 7, Ubuntu 12.04 and openSUSE 12.3 do not contain urllib. Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Even though OpenSUSE is the only rpm-based distro that supports so called weak requirements requirement on rpm-build should be strong. Otherwise gbp just cannot build the package.
|
I'm not sure I agree, because one doesn't necessarily use 'rpmbuild' to build the package. E.g. in the case of openSUSE you possibly use 'osc' and don't need rpm-build at all. |
|
Hm, maybe I misunderstand something. Here's how I usually use gbp to build a package:
This flow works like a charm for building the packages on Jenkins for all the distros we are shipping our products to. The trick is that The question is: why is OpenSUSE any different from the rest of RPM based distros? I'm not using OBS, but even if I would - from what I know it uses rpmbuild internally anyway. |
|
git-buildpackage-rpm does not necessarily use rpmbuild at all. You can define the builder command with '--git-builder'. rpmbuild is the default builder but git-buildpackage-rpm works perfectly fine without it. In openSUSE people might use the 'osc' tool for building packages. Similarly, on Fedora systems people probably want to use the 'mock' tool for building packages. These tools do not require the presence rpmbuild on the host system. This reminds me that rpm in newer Fedora systems support weak dependencies, as well, so rpmbuild could be a weak dependency there, too. |
It looks like with these comments I made it even worse - instead of getting a strong dependency on I'm currently trying to create a generic solution to build the RPM's for |
Can't you just install git-buildpackage-rpm AND rpm-build, if you're solution is depending on rpmbuild?
Yeah, why not. OTOH, you could try to push your patches directly to upstream, i.e. |
marquiz
force-pushed
the
experimental-rpm
branch
2 times, most recently
from
9253830
to
ea6a3f7
Compare
marquiz
force-pushed
the
experimental-rpm
branch
from
ea6a3f7
to
3b2b4db
Compare
marquiz
force-pushed
the
experimental-rpm
branch
from
3b2b4db
to
edb20b4
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Even though OpenSUSE is the only rpm-based distro
that supports so called weak requirements
requirement on rpm-build should be strong.
Otherwise gbp just cannot build the package.