summaryrefslogtreecommitdiff
path: root/mesonbuild/mdist.py
Commit message (Collapse)AuthorAgeFilesLines
* use real pathlib moduleDylan Baker2020-11-201-1/+1
| | | | | We added the _pathlib module to work around defeciencies in python 3.5's implementation, since we now rely on 3.6 lets drop this
* pathlib: Fix resolve() by overriding it in Python 3.5Daniel Mensinger2020-10-041-1/+1
|
* mdist: actually constrain the choices of the --formats optionDylan Baker2020-09-221-1/+1
| | | | Since we already have that list available.
* environment: use ExternalProgram to find ninjaPaolo Bonzini2020-09-041-6/+6
| | | | | | | | | | | This allows the NINJA environment variable to support all the Windows special cases, especially allowing an absolute path without extension. Based on a patch by Yonggang Luo. Fixes: #7659 Suggested-by: Nirbheek Chauhan <nirbheek@centricular.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
* wrap: Refactor to split wraps dictionary into 3 separate dictsXavier Claessens2020-07-011-1/+1
| | | | | It makes the code cleaner to have 3 separate dictionaries for packagename, dependency and programs.
* mdist: Filter out buildtype to avoid warningNirbheek Chauhan2020-06-131-1/+1
| | | | | | | | | | Since we parse buildoptions.json to pass options, we end up passing -Dbuildtype and also -Doptimization and -Ddebug which triggers the warning: WARNING: Recommend using either -Dbuildtype or -Doptimization + -Ddebug [...] Filter out buildtype. It is redundant.
* dist: Print location of generated tarballsXavier Claessens2020-03-291-0/+1
| | | | | Otherwise the location is not discoverable without reading online documentation.
* dist: Add --no-tests optionXavier Claessens2020-03-291-2/+6
| | | | | It is not always needed to test generated tarballs, for example when they are generated by CI that already does its own testing.
* Cherry-picking - (rm python2 % add python3 .format) (#6725)Michael Brockus2020-03-031-3/+3
|
* mdist: Use windows_proof_rmtree when deletingNirbheek Chauhan2020-02-171-4/+4
|
* mdist: Don't delete builddir on dist failureNirbheek Chauhan2020-02-171-26/+31
| | | | Closes https://github.com/mesonbuild/meson/issues/1976
* Exclude .hg* from dist tarballsRichard Laager2020-02-101-1/+10
| | | | | | | | | | | | | | | | | | Debian's lintian checker complaints when upstream tarballs contain an .hgtags file. This excludes ".hg[a-z]*". This is mostly consistent with the git handling in del_gitfiles() which deletes .git*. hg archive --help shows an example of -X ".hg*". However, instead of ".hg*", I have used ".hg[a-z]*" to keep the automatically added hg_archival.txt. This file may be useful to link the tarball to the Mercurial revision for either manual inspection or in case any code interprets it for a --version or similar. This also excludes .hgignore and other things like .hgflow, which seems desirable. Fixes #6575
* Set source and build dir envvars when running dist.Jussi Pakkanen2020-01-231-4/+6
|
* dist: Fix --include-subprojects when .wrap file has directory valueXavier Claessens2019-12-101-1/+5
|
* dist: Add --include-subprojects optionXavier Claessens2019-12-081-8/+36
|
* mdist.py: import lzma only when neededThomas Petazzoni2019-10-191-1/+1
| | | | | | | | Source of patch: http://lists.busybox.net/pipermail/buildroot/2019-October/262990.html http://code.bulix.org/gdxcu4-914360 Fixes https://github.com/mesonbuild/meson/issues/6012
* dist: Add '-C' option and print better error messageXavier Claessens2019-08-071-2/+9
| | | | It was not obvious that 'meson dist' must be run from build directory.
* dist: Pass right build options when configuring dist directoryThibault Saunier2019-07-301-2/+6
| | | | Fixes https://github.com/mesonbuild/meson/issues/5732
* Create multiple different archive types with dist.Jussi Pakkanen2019-07-231-47/+59
|
* Made dist a top level command.Jussi Pakkanen2019-07-231-0/+209
View Lorry Controller Status