Some time between 0.56 and 0.57 the TAP parser broke when a test exits
with a nonzero status.
The TAP protocol does not specify this behaviour - giving latitude to
implementers, and meson's previous behaviour was to report the exit
status gracefully.
This patch restores the old behaviour and adds a regression test
Currently we don't handle things correctly if we get a string we should
split, and the linker and needs compiler arguments. It would result in
two unsplit strings in a list, instead of the split arguments in a list
Fixes: #8348
This did work previously, so we need to let it continue working. I'm
proposing removing it in 0.60 because the correct solution has always
worked.
I've also been a bit more defensive here, and made setting
`subproject:opt = foo` in the machine files an error, as we have
`[subproject:built-in options]` or `[subproject:project options]` for
that.
os.path.relpath(f, wd) returns path with \ seperator on Windows, but ninja
targets always uses / separator.
See for example https://gitlab.freedesktop.org/ocrete/libnice/-/jobs/7348274.
Analyzed-by: Xavier Claessens <xavier.claessens@collabora.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Clang has a hand `-Wunused-command-line-argument` switch, which when
turned to an error, gets very grump about `-flto-jobs=0` being set in
the compiler arguments (although `-flto=` belongs there). We'll refactor
a bit to put that only in the link arguments.
GCC doesn't have this probably because, a) it doesn't have an equivalent
warning, and b) it uses `-flto=<$numthreads.
Fixes: #8347
This is a) useless because it's only used to print which options are not
default, and b) harmful because it can result in cases where things
break, like in projects that set a standard that the chosen compiler
doesn't support, but the project (or some subset) can be built with a
different standard.
Fixes: #8360
run_target() does some variable substitutions since 0.57.0. This is a
new behavior, and undocumented, caused by sharing more code with
custom_target(). More consistency is better, so document it now.
custom_target() was doing variable substitution in the past, because it
shared some code with generator(), but that was undocumented. Some
refactoring in 0.57.0 caused it to not replace @CURRENT_SOURCE_DIR@,
@SOURCE_DIR@, and @BUILD_DIR@ anymore. This patch adds back
@CURRENT_SOURCE_DIR@ and document it. It does not add back @SOURCE_DIR@
because it is duplicate with @SOURCE_ROOT@ that has a better name. Also
do not add back @BUILD_DIR@ which is duplicate of @PRIVATE_DIR@, and
not @BUILD_ROOT@ surprisingly, adding to the confusion.
A predicate in two comprehensions was dropped, in what looks like a
rebase error on my part. With the predicate added things work correctly.
Fixes: #8344
Or other language flags that use CPPFLAGS (like CXXFLAGS). The problem
here is actually rather simple, `dict.setdefault()` doesn't work like I
thought it did, I thought it created a weak entry, but it actually is
equivalent to:
```python
if k not in dict:
dict[k] = v
```
Instead we'll use an intermediate dictionary (a default dictionary
actually, since that makes things a little cleaner) and then add the
keys from that dict to self.options as applicable.
Test case written by Jussi, Fix by Dylan
Co-authored-by: Jussi Pakkanen
Fixes: #8361Fixes: #8345
In 0.57.0 installation via polkit won't work anymore. `destdir` is
defaulted to an empty string ('') instead of None. But polkit
installation code incorrectly tests `destdir is None`.
The svr4 linker flag causes issues, especially when compiling c++.
Replace '-z' options with the equivalent non-svr4 flags. When using
-blibpath, we must be careful to include the default system library
path, or the resulting executables will not be able to find libc.
This patch was suggested by @andreaskem in #7581.
It looks like GitLab ignores the suite name and actually uses
the classname. Adjust the output accordingly.
Fixes: #8316
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Avoid that the tasks linger and SingleTestRunner.run() never terminates.
In order to do this, we need read_decode() and read_decode_lines() to be
cancellable, and to handle the CancelledError gracefully while returning
the output they have collected so far.
For read_decode(), this means always operating on a line-by-line basis,
even if console_mode is not ConsoleUser.STDOUT. For read_decode_lines(),
instead, we cannot return an iterator. Rather, read_decode_lines()
returns the output directly (similar to read_decode) and communication
with the parser is mediated by an asyncio.Queue.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
This makes non-parallel tests emit their output on the fly,
similar to ninja console jobs. It also cleans up the code
a bit, avoiding the repetition of "self.options.num_processes"
tests.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
- Ensure the output is terminated with a \n even if the test does not
include one.
- Ensure that stdout is flushed for each reported result
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Fix "meson test --wrapper foo --setup bar", it should work just fine
if the setup does not define a wrapper.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Start the parsing of the output early; this avoids a deadlock
if the test writes to stdout but no one reads from it. It
also reports TAP or Rust subtest results as they happen,
which was the intention all along.
While at it, use a consistent naming conventions for coroutines
vs tasks.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Avoid printing something like "30/-1s" when tests are run without
a timeout or with --timeout-multiplier 0.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Even if the command is not wrapped by meson, it could have been modified
to add java/mono interpreters. This fix potential inconsistency between
wrapped and unwrapped commands.
This has a couple of advantages over rolling it by hand:
1. it correctly handles include_directories objects, which is always
handy
2. it correctly generates a depfile for you, which makes it more
reliable
3. it requires less typing
This allows representing functions like assert(), which take optional
positional arguments, which are not variadic. More importnatly you can
represent a function like (* means optional, but possitional):
```txt
func(str, *int, *str)
```
typed_pos_args will check that all of your types are correct, and if not
provide None, which allow simplifying a number of implementation details
We don't do a very good job of type checking in the interpreter,
sometimes we leave it to the mid layers of backends to do that (layering
violations) and sometimes we just don't check them at all. When we do
check them it's a ton of boilerplate and complicates the code. This
should help quite a bit.
Our approach to unity builds with vala is broken, you cannot unify the
generated C files, as they contain duplicate symbols. We would need to
instead combine the files while they are still in their vala form, then
convert that to C and compile the unified C file.
This does not fix the linked issue, as this removed the ability to do
vala unity builds, but it does allow running vala with `--unity=on`.
Related: #5280
Jussi Pakkanen
Luke Drummond
Dylan Baker
Krzysztof Małysa
Miro Hrončok
Paolo Bonzini
Xavier Claessens
Olexa Bilaniuk
Xℹ Ruoyao
Peter Harris
Luca Weiss
Xiaotian Wu