On my Windows machine, I mount drive D: to the directory C:\src.
The old implementation returns the incorrect 'C:\\??\\Volume{ad2eb15e-f293-4d48-a448-54757d95a97c}' result, which breaks the repo init command.
With the use of os.path.realpath, it can return 'D:\\' correctly.
Change-Id: Ia5f53989055125cb282d4123cf55d060718aa1ff
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/416580
Reviewed-by: Greg Edelston <gredelston@google.com>
Tested-by: Kaiyi Li <kaiyili@google.com>
Commit-Queue: Kaiyi Li <kaiyili@google.com>
Use the requirements logic in the wrapper to load versions out of the
requirements.json file to avoid duplicating them in git_command.py.
Change-Id: Ib479049fc54ebc6f52c2c30d1315cf1734ff1990
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/415617
Reviewed-by: Josip Sokcevic <sokcevic@google.com>
Commit-Queue: Mike Frysinger <vapier@google.com>
Tested-by: Mike Frysinger <vapier@google.com>
os.path.exists returns false on a broken symlink. This is not what repo
needs when checking if a project is setup properly.
For example, if src/foo/.git can't be resolved, repo tries to create
symlink and that results in FileExistsError.
Use lexists which returns True even if symlink is broken. That will
force path where repo checks where symlink is pointing to and will fix
it to the correct location.
Bug: b/281746795
Change-Id: Id3f7dc3a3cb6499d02ce7335eca992ddc7deb645
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/415197
Tested-by: Josip Sokcevic <sokcevic@google.com>
Commit-Queue: Josip Sokcevic <sokcevic@google.com>
Reviewed-by: George Engelbrecht <engeg@google.com>
Reviewed-by: Greg Edelston <gredelston@google.com>
git-1.9.1 was released in 2014 while git-2.7.4 was released in 2016.
Debian Stretch was released in 2017 and Ubuntu Xenial was released in
2016 which are plenty old at this point. Both of those include at
least git-2.7.4.
We will start warning users of Debian Jessie (released in 2015 & EOL
in 2020) and Ubuntu Trusty (released in 2014 & EOL Apr 2024) that
they will need to upgrade.
Change-Id: I6be3809bc45968fdcb02cad3f7daf75ded1bb5b1
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/415137
Tested-by: Mike Frysinger <vapier@google.com>
Reviewed-by: Josip Sokcevic <sokcevic@google.com>
Commit-Queue: Mike Frysinger <vapier@google.com>
Debian 6 Squeeze went EOL in Feb 2016. We don't need to carry support
for that anymore as there have been 6 major releases since. That means
we can bump the required git version from 1.7.2 to 1.7.9. Ubuntu Precise
shipped with the latter.
git-1.7.2 was released in 2010 while git-1.7.9 was released in 2012.
So that shouldn't be a problem either. And we've been warning people
using git versions older than 1.9.1 for 3 years now that they need to
upgrade.
Change-Id: I7712f110ea158297b489b8379b112c6700b21a46
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/415097
Reviewed-by: Josip Sokcevic <sokcevic@google.com>
Tested-by: Mike Frysinger <vapier@google.com>
Commit-Queue: Mike Frysinger <vapier@google.com>
If a project can't be removed from checkout due to uncommitted changes
present, error.project is type of Project and not a string (as it is in
some cases). Project is not JSON serializable, resulting in exception
within exception handler:
TypeError: Object of type Project is not JSON serializable
This change casts project to string as a defensive mechanism. It also
passes project name instead of project object.
Change-Id: Ie7b782d73dc3647975755d5a3774d16ea6cd5348
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/413877
Tested-by: Josip Sokcevic <sokcevic@google.com>
Reviewed-by: Gavin Mak <gavinmak@google.com>
Commit-Queue: Josip Sokcevic <sokcevic@google.com>
The current logic to create checkout layers doesn't work in all cases.
For example, let's assume there are three projects: "foo", "foo/bar" and
"foo-bar". Sorting lexicographical order is incorrect as foo-bar would
be placed between foo and foo/bar, breaking layering logic.
Instead, we split filepaths based using path delimiter (always /) and
then use lexicographical sort.
BUG=b:325119758
TEST=./run_tests, manual sync on chromiumos repository
Change-Id: I76924c3cc6ba2bb860d7a3e48406a6bba8f58c10
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/412338
Tested-by: Josip Sokcevic <sokcevic@google.com>
Commit-Queue: Josip Sokcevic <sokcevic@google.com>
Reviewed-by: George Engelbrecht <engeg@google.com>
In some cases (e.g. in a CI system), it's desirable to be able to
instruct repo to force checkout. This flag passes --force flag to `git
checkout` operations.
Bug: b/327624021
Change-Id: I579edda546fb8147c4e1a267e2605fcf6e597421
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/411518
Commit-Queue: Josip Sokcevic <sokcevic@google.com>
Reviewed-by: Gavin Mak <gavinmak@google.com>
Reviewed-by: George Engelbrecht <engeg@google.com>
Tested-by: Josip Sokcevic <sokcevic@google.com>
If a repo manifest is updated so that project B is placed within a
project A, and if project A had content in new B's location in the old
checkout, then repo sync could break depending on checkout order, since
B can't be checked out before A.
This change introduces checkout levels which enforces right sequence of
checkouts while still allowing for parallel checkout. In an example
above, A will always be checked out first before B.
BUG=b:325119758
TEST=./run_tests, manual sync on ChromeOS repository
Change-Id: Ib3b5e4d2639ca56620a1e4c6bf76d7b1ab805250
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/410421
Tested-by: Josip Sokcevic <sokcevic@google.com>
Reviewed-by: Greg Edelston <gredelston@google.com>
Commit-Queue: Josip Sokcevic <sokcevic@google.com>
Reviewed-by: Gavin Mak <gavinmak@google.com>
If trace2.eventTarget was set to the empty string,
match git behavior and don't write a trace.
Bug: 319673783
Change-Id: I02b3884ad97551f8a9d7363c2cbe6b0adee6f73e
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/410518
Reviewed-by: Josip Sokcevic <sokcevic@google.com>
Commit-Queue: Josip Sokcevic <sokcevic@google.com>
Tested-by: Peter Collingbourne <pcc@google.com>
git-trace2 events contain additional information what git is doing under
the hood, and repo doesn't have visibility into.
Instead of relying on timestamp information to match REPO_TRACE with
git-trace2 events, add SID information into REPO_TRACE.
Change-Id: I37672a3face81858072c7a3ce34ca3379199dab5
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/410280
Tested-by: Josip Sokcevic <sokcevic@google.com>
Reviewed-by: Gavin Mak <gavinmak@google.com>
Commit-Queue: Josip Sokcevic <sokcevic@google.com>
Some (most?) Linux repos don't have /usr/bin/python, unless
python-is-python3 is installed. While package owners can adjust shebang,
we have seen an increase in number of bugs filed as extra steps are
required.
Per PEP-0394, python3 is acceptable and should be available if python3
is supported. We no longer support python2, and repo no longer works
with python2, so this change makes that explicit.
Bug: 40014585
Change-Id: I9aed90fd470ef601bd33bd596af3df69da69ef5d
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/407497
Commit-Queue: Josip Sokcevic <sokcevic@google.com>
Reviewed-by: Jason Chang <jasonnc@google.com>
Reviewed-by: Gavin Mak <gavinmak@google.com>
Reviewed-by: Максим Паймушкин <maxim.paymushkin@gmail.com>
Tested-by: Josip Sokcevic <sokcevic@google.com>
Prior to this change RepoChangedException would be caught and re-rasied
as a different exception. This would prevent RepoChangedException
handler from running in main.py
Bug: b/323232806
Change-Id: I9055ff95d439d6ff225206c5bf1755cc718bcfcc
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/407144
Tested-by: Josip Sokcevic <sokcevic@google.com>
Reviewed-by: Josip Sokcevic <sokcevic@google.com>
Commit-Queue: Josip Sokcevic <sokcevic@google.com>
UpdateManifestError inherits from RepoExitError which inherits
from BaseRepoError. None of them takes project as kwargs
causing the error like "UpdateManifestError() takes no keyword
arguments" in b/317183321
[1]: https://gerrit.googlesource.com/git-repo/+/449b23b698d7d4b13909667a49a0698eb495eeaa/error.py#144
Bug: b/317183321
Change-Id: I64c3dc502027f9dda56a0824f2712364b4502934
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/398997
Commit-Queue: Yiwei Zhang <yiwzhang@google.com>
Tested-by: Yiwei Zhang <yiwzhang@google.com>
Reviewed-by: Josip Sokcevic <sokcevic@google.com>
Reviewed-by: Jason Chang <jasonnc@google.com>
If a project contains files not owned by the current user, remove will
fail. In order to ensure repo sync continues to work, rename the
affected project instead, and let user know about it.
Bug: 321273512
Change-Id: I0779d61fc67042308a0226adea7d98167252a5d3
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/404372
Reviewed-by: Mike Frysinger <vapier@google.com>
Tested-by: Josip Sokcevic <sokcevic@google.com>
Commit-Queue: Josip Sokcevic <sokcevic@google.com>
There are still some verbose messages (e.g. "remote: ...") when doing
repo sync after a couple days. Let's hide them behind verbose flag.
Bug: N/A
Test: repo sync
Change-Id: I1408472c95ed80d9555adfe8f92211245c03cf41
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/400855
Reviewed-by: Josip Sokcevic <sokcevic@google.com>
Tested-by: Tomasz Wasilczyk <twasilczyk@google.com>
Commit-Queue: Tomasz Wasilczyk <twasilczyk@google.com>
Symbolic references need to be checked each time sync is called, not
only for newly created repositories. For example, it is possible to
change a project name to the already existing name, and that will result
in a broken git setup without this patch: refs/ will still point to the
old repository, whereas all objects will point to the new repository.
Bug: 40013418
Change-Id: I596d29d182986804989f0562fb45090224549b0f
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/395798
Reviewed-by: Mike Frysinger <vapier@google.com>
Tested-by: Josip Sokcevic <sokcevic@google.com>
Commit-Queue: Josip Sokcevic <sokcevic@google.com>
After a6413f5d a GitCommandError is raised.
Since there were no user identity were set up,
it fails:
- ReviewableBranchTests from test_project.py
- ResolveRepoRev and CheckRepoRev from test_wrapper.py
Test: ./run_tests
Change-Id: Id7f5772afe22c77fc4c8f8f0b8be1b627ed42187
Signed-off-by: Vitalii Dmitriev <vitalii.dmitriev@unikie.com>
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/398658
Reviewed-by: Mike Frysinger <vapier@google.com>
Tested-by: Vitalii Dmitriev <dmit.vitalii@gmail.com>
Commit-Queue: Vitalii Dmitriev <dmit.vitalii@gmail.com>
Before the change, repo normalizes the urls
with a following format only:
git@github.com:foo/bar
It doesn't cover the following case:
<remote name="org" fetch="git@github.com:org/" />
<project name="somerepo" remote="org" />
Results to:
error: Cannot fetch somerepo
from ssh://git@github.com/org/git@github.com:org/somerepo
Current change fixes it by normalizing this format:
git@github.com:foo
Test: ./run_tests tests/test_manifest_xml.py
Change-Id: I1ad0f5df0d52c0b7229ba4c9a4db4eecb5c1a003
Signed-off-by: Vitalii Dmitriev <vitalii.dmitriev@unikie.com>
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/398337
Commit-Queue: Vitalii Dmitriev <dmit.vitalii@gmail.com>
Tested-by: Vitalii Dmitriev <dmit.vitalii@gmail.com>
Reviewed-by: Mike Frysinger <vapier@google.com>
repo blocks indefinitely until trace collector receives trace events,
which is not desired. This change adds a fixed timeout to connect and
send operations. It is possible that some events will be lost. repo logs
any failed trace operation.
Bug: b/316227772
Change-Id: I017636421b8e22ae3fcbab9e4eb2bee1d4fbbff4
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/398717
Tested-by: Josip Sokcevic <sokcevic@google.com>
Commit-Queue: Josip Sokcevic <sokcevic@google.com>
Reviewed-by: Jason Chang <jasonnc@google.com>
This simplifies check_python_version() since there is no point in trying
to fall back to python3, as we are already running using some Python 3
interpreter.
Change-Id: I9dfdd002b4ef5567e064d3d6ca98ee1f3410fd48
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/397759
Reviewed-by: Mike Frysinger <vapier@google.com>
Tested-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Commit-Queue: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
The EnvironmentError exception was changed to OSError in commit
ae824fb2fc.
Change-Id: I1b4ff742af409ec848131e82900e885c9f089f0c
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/397757
Reviewed-by: Mike Frysinger <vapier@google.com>
Tested-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Commit-Queue: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Most times a repo sync after some time (week+) results in a bunch of
messages, which are not very useful for average user:
- discarding 1 commits
- Deleting obsolete checkout.
Bug: N/A
Test: repo sync
Change-Id: I881eab61f9f261e98f3656c09e73ddd159ce288c
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/397038
Reviewed-by: Josip Sokcevic <sokcevic@google.com>
Commit-Queue: Josip Sokcevic <sokcevic@google.com>
Tested-by: Tomasz Wasilczyk <twasilczyk@google.com>
It's not a valid pylint config
Change-Id: Ida480429a3a86637f26e9fc3a0d6fa2d225d952a
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/396921
Reviewed-by: Scott Lee <ddoman@google.com>
Commit-Queue: Josip Sokcevic <sokcevic@google.com>
Tested-by: Josip Sokcevic <sokcevic@google.com>
In the case of a project being removed from the manifest, and in the
path in which the project used to exist, and symlink is place to another
project repo will start to warn about partial syncs when a partial sync
did not occur.
Repro steps:
1) Create a manifest with two projects. Project a -> a/ and project b -> b/
2) Run `repo sync`
3) Remove project b from the manifest.
4) Use `link` in the manifest to link all of Project a to b/
Bug: 314161804
Change-Id: I4a4ac4f70a7038bc7e0c4e0e51ae9fc942411a34
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/395640
Reviewed-by: Gavin Mak <gavinmak@google.com>
Tested-by: Matt Schulte <matsch@google.com>
Commit-Queue: Gavin Mak <gavinmak@google.com>
Python 3.11 introduces PYTHONSAFEPATH and the -P flag which, if enabled,
does not prepend the script's directory to sys.path by default.
This breaks repo because main.py expects its own directory to be part of
Python's import path.
This causes problems with tools that add PYTHONSAFEPATH to python
programs, most notably Bazel.
We will simply prepend main.py's path to PYTHONPATH instead.
Bug: 307767740
Change-Id: I94f3fda50213e450df0d1e2df6a0b8b597416973
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/391236
Tested-by: Duy Truong <duytruong@google.com>
Commit-Queue: Mike Frysinger <vapier@google.com>
Reviewed-by: Mike Frysinger <vapier@google.com>
When a repo client is initialized with --standalone-manifest, it doesn't
have merge branch defined. This results in mergeBranch being None.
Bug: b/308025460
Change-Id: Iebceac0976e5d3adab5300bd8dfc76744a791234
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/393716
Tested-by: Josip Sokcevic <sokcevic@google.com>
Reviewed-by: Mike Frysinger <vapier@google.com>
Commit-Queue: Josip Sokcevic <sokcevic@google.com>
This updates RepoLogger.log_aggregated_errors to print out the error
message the RepoExitError when there is not a list of aggregated
errors.
Previously it would log out:
=======================================================================
Repo command failed: ManifestParseError
This told us what class of error occurred but missed the helpful error
message that developers put in the error. After this change it will now
print out the error message:
=======================================================================
Repo command failed: ManifestParseError
error parsing manifest /path/to/manifest.xml: no element found:
line 197, column 0
Change-Id: I4805540fddb5fa9171dbc8912becfa7fdfb1ba67
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/392614
Commit-Queue: Aravind Vasudevan <aravindvasudev@google.com>
Tested-by: Joshua Bartel <josh.bartel@garmin.com>
Reviewed-by: Aravind Vasudevan <aravindvasudev@google.com>
Avoids treating the operation as if it were acting on a bare repository, thereby triggering failures when the Git client is configured with `safe.bareRepository=explicit`. Repo doesn't actually use a bare repository, but pointing at the gitdir acts as if it had.
Bug: 307559774
Change-Id: I2c142275b2726a59526729c0b2c54faf728f125d
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/391554
Commit-Queue: Jason R. Coombs <jaraco@google.com>
Tested-by: Jason R. Coombs <jaraco@google.com>
Tested-by: Emily Shaffer <emilyshaffer@google.com>
Reviewed-by: Emily Shaffer <emilyshaffer@google.com>
Reviewed-by: Mike Frysinger <vapier@google.com>
The superproject is initialized as a bare repo in Superproject:_Init().
That means that later operations must treat it as a bare repository,
specifying the gitdir and setting 'bare' appropriately when launching
GitCommand()s. It's also OK not to specify cwd here because GitCommand()
will drop cwd if bare == True anyways.
With this change, it's possible to run `repo init` and `repo sync` with the
Git config 'safe.bareRepository' set to 'explicit'. This config strengthens
Git's security posture against embedded bare repository attacks like
https://github.com/justinsteven/advisories/blob/main/2022_git_buried_bare_repos_and_fsmonitor_various_abuses.md.
Bug: b/227257481
Change-Id: I954a64c6883d2ca2af9c603e7076fd83b52584e9
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/389794
Reviewed-by: Mike Frysinger <vapier@google.com>
Tested-by: Jason R. Coombs <jaraco@google.com>
Tested-by: Emily Shaffer <emilyshaffer@google.com>
Reviewed-by: Emily Shaffer <emilyshaffer@google.com>
Commit-Queue: Jason R. Coombs <jaraco@google.com>
Currently we don't deal with schema-less URIs like
`git@github.com:foo` at all resulting in a scenario where we append
them to the manifest repo URL.
In order to deal with this, we munge both the manifest URL and the
fetch URL into a format we like and proceed with that.
Bug: https://g-issues.gerritcodereview.com/issues/40010331
Change-Id: I7b79fc4ed276630fdbeb235b94e327b172f0879b
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/386954
Reviewed-by: Mike Frysinger <vapier@google.com>
Tested-by: Michael Kelly <mkelly@arista.com>
Commit-Queue: Mike Frysinger <vapier@google.com>
Change-Id: I41371feb88c85e9da0656b9fab04057c22d1dcf4
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/391514
Tested-by: Jason R. Coombs <jaraco@google.com>
Reviewed-by: Mike Frysinger <vapier@google.com>
Commit-Queue: Jason R. Coombs <jaraco@google.com>
- Bump minimum version to Python 3.6.
- Use f-strings in a lot of places.
Change-Id: I2aa70197230fcec2eff8e7c8eb754f20c08075bb
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/389034
Tested-by: Jason R. Coombs <jaraco@google.com>
Reviewed-by: Mike Frysinger <vapier@google.com>
Commit-Queue: Jason R. Coombs <jaraco@google.com>
Found via pylint:
W0231: __init__ method from base class 'Transport'
is not called (super-init-not-called)
Just fixed for code correctness and to avoid potential future bugs.
Change-Id: Ie1e723c2afe65d026d70ac01a16ee7a40c149834
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/390676
Reviewed-by: Mike Frysinger <vapier@google.com>
Tested-by: Daniel Kutik <daniel.kutik@lavawerk.com>
Commit-Queue: Daniel Kutik <daniel.kutik@lavawerk.com>
When a new shared project is added to manifest, there's a short window
where objects can be deleted that are used by other projects.
To close that window, set preciousObjects during git init. For
non-shared projects, repo should correct the state in the same execution
instance.
Bug: 288102993
Change-Id: I366f524535ac58c820d51a88599ae2108df9ab48
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/390234
Commit-Queue: Josip Sokcevic <sokcevic@google.com>
Tested-by: Josip Sokcevic <sokcevic@google.com>
Reviewed-by: Mike Frysinger <vapier@google.com>
In Python 3, these exceptions were merged into OSError, so switch
everything over to that.
Change-Id: If876a28b692de5aa5c62a3bdc8c000793ce52c63
Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/390376
Reviewed-by: Aravind Vasudevan <aravindvasudev@google.com>
Commit-Queue: Mike Frysinger <vapier@google.com>
Tested-by: Mike Frysinger <vapier@google.com>
