| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
| |
Also register them to run nightly.
Running tests in a randomized order is a good way of checking whether
our tests depend on each others or not.
Since we order them globally, tests will take longer to run, but that
will give us higher confidence that they are correct.
|
|\
| |
| |
| |
| | |
remove unused functions 1/2
See merge request BuildStream/buildstream!1753
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
| | |
|
|/ |
|
|\
| |
| |
| |
| | |
Fixes for asyncio loops were we don't follow the documentation
See merge request BuildStream/buildstream!1756
|
| |
| |
| |
| |
| |
| |
| |
| | |
Per
https://docs.python.org/3/library/asyncio-policy.html#asyncio.AbstractChildWatcher.add_child_handler,
the callback from a child handler must be thread safe. Not all our
callbacks were. This changes all our callbacks to schedule a call for
the next loop iteration instead of executing it directly.
|
| |
| |
| |
| |
| |
| |
| |
| | |
The documentation
(https://docs.python.org/3/library/asyncio-policy.html#asyncio.AbstractChildWatcher)
is apparently missing this part, but the code mentions that new
processes should only ever be called inside a with block:
https://github.com/python/cpython/blob/99eb70a9eb9493602ff6ad8bb92df4318cf05a3e/Lib/asyncio/unix_events.py#L808
|
|/
|
|
| |
We don't need to keep a reference to the watcher, let's remove it.
|
|\
| |
| |
| |
| | |
scheduler.py: Optimize scheduling by not calling it unnecessarily
See merge request BuildStream/buildstream!1755
|
|/
|
|
|
|
|
|
|
|
|
|
| |
This delays the call to the re-scheduling of jobs until the current
event loop as terminated.
This is in order to reduce the number of time we call this method per
loop, which should reduce the pressure on the loop and allow faster
event handling
Since the call is now delayed, also ensure we only call it once per loop
iteration.
|
|\
| |
| |
| |
| | |
Minor packaging-related fixups
See merge request BuildStream/buildstream!1754
|
|/
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This patch contains the following changes:
* Remove include directive for MAINTAINERS file, that does not exist
anymore. Instead, add the COMMITTERS.rst file, which is the logical
replacement for the MAINTAINERS file.
* Recursively include all requirements files instead of listing them one
by one. This also fixes and issue where the cov-requirements files
were not included in the source distribution.
* Simplify `recursive-include DIR *` to `graft DIR`. These two forms are
functionally equivalent, but the latter is a bit easier to understand.
|
|\
| |
| |
| |
| | |
Add tests for interactive BuildStream operations
See merge request BuildStream/buildstream!1706
|
| | |
|
| |
| |
| |
| |
| | |
Add tests for interactive `bst init` command using
[pexpect](https://pexpect.readthedocs.io).
|
|/ |
|
|\
| |
| |
| |
| | |
_sandboxreapi.py: Reset CasBasedDirectory instead of recreating it
See merge request BuildStream/buildstream!1749
|
| | |
|
| |
| |
| |
| |
| |
| | |
Calling _reset() instead of completely replacing the object fixes
element plugins that use a virtual directory object across Sandbox.run()
calls such as the compose plugin with integration commands.
|
|/
|
|
| |
This reinitializes a CASBasedDirectory object from a directory digest.
|
|\
| |
| |
| |
| | |
testing/runcli.py: Remove unused configure parameter from run() methods
See merge request BuildStream/buildstream!1748
|
|/ |
|
|\
| |
| |
| |
| | |
Multiple CI improvements
See merge request BuildStream/buildstream!1747
|
| |
| |
| |
| |
| | |
We can remove the sast job overwriting and instead 'just' add the
variable it needs to the global variables.
|
| |
| |
| |
| |
| | |
Our image versions tend to drift over time. Let's use a variable to
ensure they are always all at the same version.
|
|/
|
|
|
| |
Contrary to what we thought, those don't get expanded, and we ended up
having the external tests run only for py35, which is incorrect.
|
|\
| |
| |
| |
| | |
Support buildbox-casd running as separate user
See merge request BuildStream/buildstream!1737
|
| |
| |
| |
| |
| |
| |
| |
| | |
To protect the local cache of buildbox-casd from corruption without the
use of FUSE, buildbox-casd has to run as a different user.
Use less restrictive umasks in the source determinism tests to allow
buildbox-casd to function when it is running as a separate user.
|
| |
| |
| |
| | |
This is not necessary and doesn't work with CASD_SEPARATE_USER.
|
| |
| |
| |
| | |
This is not necessary and doesn't work with CASD_SEPARATE_USER.
|
| |
| |
| |
| |
| | |
Linux does not allow hardlinks to read-only files of other users by
default since Linux 3.6 (see /proc/sys/fs/protected_hardlinks).
|
| |
| |
| |
| |
| | |
This is set to True if buildbox-casd is installed with the set-uid bit
and thus, indicates whether buildbox-casd is running as a separate user.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
By default, Linux doesn't allow creating hardlinks to read-only files of
other users since Linux 3.6 (see /proc/sys/fs/protected_hardlinks).
This fixes staging when buildbox-casd is running as a separate user and
the traditional bubblewrap sandboxing backend is used. This combination
is not recommended, however, it's triggered in CI by docker images that
run buildbox-casd as a separate user and a few test cases that override
BST_FORCE_SANDBOX.
|
|/
|
|
|
|
|
|
|
|
|
| |
This is necessary to allow using buildbox-run with userchroot in the
near future, since currently only the owner of the BuildStream process
can access the CASD socket, but the buildbox-casd binary will need to
be setuid' to another user.
This gets around this limitation by allowing the group to access a
symlink, which in turn should point to a directory owned by the CASD
user.
|