Adds TemporaryDirectoryWinProof which calls windows_proof_rmtree() on
error.
Use instead of hacky error handling (which might shadow other OSError)
in Compiler.compile().
In most places, we now refer to "builddir/" which is a lot less likely
to make people think it is a subcommand which needs to be used
literally.
This is a regression since commit 276d342eba
due to the existence of new docs which were added later on, using the
wrong form.
Fixes a test when using pkgconf instead of pkg-config.
The .pc file in the pkgconfig-gen test requires it and pkgconf seems be
stricter here and fails to validate.
Some CMake packages fail to find at all if no version is specified.
This commit adds a cmake_version parameter to dependency() to allow you
to specify the requested version.
This value shouldn't be set in the loop, it's probably been making
things slow, but also creates the potential that if choices is an empty
list that value would be unbound. While that's exteremly unlikly to
happen in practice, it is showing up as a warning from pylance.
It's a bit odd we have a test for an external project that's not even in
the meson organization. Regardless, the json file was recently replaced
by a cson file. There is an coffee-script-notation parser in pypi,
but I couldn't get it to work. Just delete the test
As a necessity nix adds a bunch of rpaths to files, this is unavoidable
do to the way nix package management works. Meson doesn't expect this
however, and fails all rpath tests. To correct this we just ignore any
rpath entries that start with `/nix`.
Avoid calling self.collected_failures.append twice, and avoid
inflated indentation by adding a "plain" decorator to mlog.
Fixes: ba71fde18 ("mtest: collect failures regardless of colorized console", 2020-10-12)
It turns out my first attempt to fix this in 00d5ef3191 ("Fix
clang-tidy return value reporting (#7949)") is not sufficient: The
local variable returncode is never updated and stays at 0. This fixes
the returncode calculation.
Fixes: cce172432b ("Use run-clang-tidy when available.")
- mention installing from local sources, not PyPI
- warn against --user installs, which too often screw up users that then
cannot install projects because ~/.local won't be in sudo's PYTHONPATH
- advise installing with sudo -- current versions of pip assume --user
for you rather than failing with permission errors, which is great
unless, like meson, there are compelling reasons to need to install as
root
See https://gitlab.gnome.org/GNOME/glib/-/issues/600
`volatile` was previously mistakenly used in GLib to indicate that a
variable was accessed atomically or otherwise multi-threaded. It’s not
meant for that, and up to date compilers (like gcc-11) will rightly warn
about it.
Drop the `volatile` qualifiers.
Based on a patch by Jeff Law.
See also http://isvolatileusefulwiththreads.in/c/.
Signed-off-by: Philip Withnall <pwithnall@endlessos.org>
env.detect_<lang>_compiler only checks that there is a binary called
whatever that returns a version. There are several cases where the found
binary doesn't work:
1) gcc for ojbc[pp], when support isn't compiled in.
2) the compiler is broken (rust in appveyor somtimes)
Because of that we need to call compiler.sanity_check() as well, and if
we get an EnvironmentException from that skip the language
Also, instead of having a long line of try: ... except: pass, roll all
of the checking up into a loop using getattr(), which is less code and
makes adding a new language easier
These do not go into the command line, and therefore do not matter
for the purpose of avoiding unnecessary rebuilds after meson is
rerun. However, they complicate the task of finding differences
between build lines across meson reruns.
So take the easy way out and sort everything after | and ||.
With this change, there is absolutely no change in QEMU's 40000-line
build.ninja file after meson is rerun.
The order of keys in dictionaries cannot be relied upon, because the hash
values are randomized by Python. Whenever we iterate on dictionaries and
meson.build generates a list during the iteration, the different iteration
orders may cause random changes in the command line and cause ninja to
rebuild a lot of files unnecessarily.
The order of elements in sets cannot be relied upon, because the hash
values are randomized by Python. Whenever sets are converted to lists
we need to keep their order stable, or random changes in the command line
cause ninja to rebuild a lot of files unnecessarily. To stabilize them,
use either sort or OrderedSet. Sorting is not always applicable, but it
can be faster because it's done in C and it can produce slightly nicer
output.
Rewrite the SingleTestRunner to use asyncio to manage subprocesses,
while still using subprocess.Popen to run them. Concurrency is
managed with an asyncio Semaphore; for simplicity (since this is
a temporary state) we create a new thread for each test that is run
instead of having a pool.
This already provides the main advantage of asyncio, which is better
control on cancellation; with the current code, KeyboardInterrupt
was never handled by the thread executor so the code that tried to handle
it in SingleTestRunner only worked for non-parallel tests. And
because executor futures cannot be cancelled, there was no way for
the user to kill a test that got stuck. Instead, without executors
^C exits "meson test" immediately. The next patch will improve things
even further, allowing a single test to be interrupted with ^C.
Distinguish a failure due to user interrupt from a presumable ERROR
result due to the SIGTERM. The test should fail after CTRL+C even if
the test traps SIGTERM and exits with a return code of 0.
ProcessLookupError can also happen from p.kill(). There is also
nothing we can do in that case, so move the "try" for that
exception to the entire kill_process function.
The ValueError case seems like dead code, so get rid of it.
A large part of _run_cmd is devoted to setting up and killing the
test subprocess. Move that to a separate function to make the
test runner logic easier to understand.
Use asyncio futures for the run loop, while still handling I/O in
a thread pool using run_on_executor.
The handling of the test result is not duplicated anymore between
run_tests and drain_futures. Instead, the test result is always processed
and printed by run_test after single_test.run() completes and (in verbose
mode) it cannot interleave with the test output. Therefore the special
case for self.options.num_processes == 1 can be removed.
Michael Brockus
Christoph Reiter
Eli Schwartz
Jussi Pakkanen
Jason Ekstrand
Dylan Baker
Paolo Bonzini
Florian Schmaus
Philip Withnall