Tristan Van Berkom authored Oct 03, 2018 and Tristan Van Berkom committed Oct 03, 2018

For some reason, we now receive a SIGINT from the main loop even
when the SIGINT occurred with the handler disconnected in an interactive
prompt. This patch simply ignores any received SIGINT events from the
main loop in the case that we are already in the process of terminating.

This fixes issue #693

_scheduler/scheduler.py: Ignore interrupt events while terminating.

Closes #693

See merge request !852

Tristan Van Berkom authored Oct 03, 2018 and Tristan Van Berkom committed Oct 03, 2018

Allow callers to decide where the temporary file will be created.

Tristan Van Berkom authored Oct 03, 2018 and Tristan Van Berkom committed Oct 03, 2018

Use the designated tempdir when creating refs, we expect that temporary
files are not created in the storage directory ever, they should be
only ever created in the designated temporary directory.

This fixes race conditions where utils._get_dir_size() throws an
unhandled exception when attempting to stat the file which inadvertantly
disappears.

Tristan Van Berkom authored Oct 03, 2018 and Tristan Van Berkom committed Oct 03, 2018

This function assumes that files do not disappear while
walking the given directory.

fix cache size race

See merge request !854

Move the temporary staging directory to artifactdir

See merge request !856

Since the SandboxDummy sandbox is only used in very specific
circumstances and with good reason, this adjusts the SandboxDummy
class to take a reason and to return it if `.run()` has to
raise an exception because this is a dummy sandbox.

Signed-off-by: Daniel Silverstone <daniel.silverstone@codethink.co.uk>

Since Darwin is limited to the dummy sandbox for now due to OSXFUSE
being unsupported as yet, and there being no suitable sandboxing option
for use on OSX, give this as the reason so that it can be reported if
the user attempts a local build.

Signed-off-by: Daniel Silverstone <daniel.silverstone@codethink.co.uk>

To better report issues in the absence of a suitable bwrap, or the
FUSE devices, this refactors the checks for sandboxing in the Linux
platform to cover the various possibilities.  The reasons are then
collated and passed to the dummy sandbox for later reporting to the
user if a local build is attempted.

Signed-off-by: Daniel Silverstone <daniel.silverstone@codethink.co.uk>

On some potentially broken systems, running `bwrap --version` might
fail with an error code.  This patch corrects the oversight ensuring
that we cleanly return False for version checking in such cases.

Signed-off-by: Daniel Silverstone <daniel.silverstone@codethink.co.uk>

Instead of an if/else ladder which is quite complex, this patch
uses the fact that Python supports by-component tuple comparison
to simply compare two (major, minor, patch) tuples

Signed-off-by: Daniel Silverstone <daniel.silverstone@codethink.co.uk>

Since there are use-cases where BuildStream could be installed
onto systems which do not have BubbleWrap (e.g. for remote-build-only
scenarios) it is not correct to assert a dependency on bwrap during
installation.  This patch makes the assertion a warning, and also
clarifies the message somewhat.  This should fix #644

Signed-off-by: Daniel Silverstone <daniel.silverstone@codethink.co.uk>

Make bwrap check runtime only

Closes #644

See merge request !847

Qinusty authored Sep 19, 2018 and Tristan Van Berkom committed Oct 04, 2018

This fixes #634, BuildStream no longer saves workspace configuration on
a failed build.