Forcing serialization on when writing out the build rule makes very
little sense. It was always "forced" on because we mandated a couple of
environment variables due to legacy reasons.
Add an attribute to RunTarget to say that a given target doesn't *need*
those environment variables, and let ninja optimize them away and run
the command directly if set.
That method had nothing specific to the backend, it's purely a Target
method. This allows to cache the OptionOverrideProxy object on the
Target instance instead of creating a new one for each option lookup.
We print a warning if a compilation database isn't successfully
generated, which is good, because that gives some visibility in case the
user really wanted to use the compdb. But warnings default to being
fatal with --fatal-meson-warnings, which is not so good, because this
isn't a very important warning at all, and we'd rather not error out in
such cases when building works fine and a random bonus IDE feature
doesn't work.
Mark this particular warning as non-fatal.
Fixes side issue in https://github.com/mesonbuild/wrapdb/pull/343#issuecomment-1074545609
In the even that all of the inputs are generated, and they're all
generated into the same folder, and there are no subfolders, we would
fail to correctly handle all of the files after the main file. Let's fix
that.t
We currently don't handle subdirectories correctly in
structured_sources, which is problematic. To make this easier to handle
correctly, I've simply changed `structured_sources` to only use Files
and not strings as an implementation detail.
In this case, the test fname might have an implicit extension and cannot
be found by `os.path.isfile()`.
We cannot use `shutil.which()` to handle platform differences, because
not all test fnames are executable -- for example Java jars.
The test representation does have an "is built" attribute which in
theory should work here, because all built targets definitely have their
full filename known to Meson, but it turns out to be misnamed. Rename it
correctly and add an actual "is built" attribute to check.
Tests which aren't built by Meson can be assumed to exist without
consulting their existence on the filesystem.
Fixes#10027
Using future annotations, type annotations become strings at runtime and
don't impact performance. This is not possible to do with T.cast though,
because it is a function argument instead of an annotation.
Quote the type argument everywhere in order to have the same effect as
future annotations. This also allows linters to better detect in some
cases that a given import is typing-only.
Android requires shared modules that use symbols from other shared
modules to be linked before they can be dlopen()ed in the correct
order. Not doing so leads to a missing symbol error:
https://github.com/android/ndk/issues/201
We need to always allow linking for this. Also add a soname, although
it's not confirmed that it's needed, and it doesn't really hurt if it
isn't needed.
For static library crates that depend on other internal static library
crates, all link_with targets get promoted to link_whole targets. Due to
a bug, only link_with targets are considered when generating a Rust
target for the ninja backend. This made it impossible to link a Rust
static library with another internal Rust static library.
This change fixes that issue by chaining link_whole_targets with
link_targets, just like many other languages within the ninja backend.
This was a nice idea in theory, but in practice it had various problems:
- On the only platform where ldconfig is expected to be run, it is
really slow, even when the user uses a non-default prefix and ldconfig
doesn't even have permission to run, nor can do anything useful due to
ld.so.conf state
- On FreeBSD, it bricked the system: #9592
- On cross builds, it should not be used and broke installing, because
ldconfig may not be runnable without binfmt + qemu: #9707
- it prints weird and confusing errors in the common "custom prefix"
layout: #9241
Some of these problems can be or have been fixed. But it's a constant
source of footguns and complaints and for something that was originally
supposed to be just "it's the right thing to do anyway, so just do it
automatically" it is entirely too risky.
Ultimately I do not think there is justification for keeping this
feature in since it doesn't actually make everyone happy. Better for
users to decide whether they need this themselves.
This is anyways the case for cmake and autotools and generally any other
build system, so it should not be too intimidating...
Fixes#9721
Store in TestSerialisation whether a particular test must always be logged
verbosely. This is particularly useful for long-running tests or when a
single Meson test() is wrapping an external test harness. In this case,
TAP can be used by the external harness and Meson will log each subtest as
it runs.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Basically the last thing we did during target processing was to generate
shlib symlinks for e.g. libfoo.so -> libfoo.so.1.
In some cases we would dispatch to another function and return early,
though, which meant we never got far enough to generate the symlinks.
This then led to breakage when people tried to compile against libfoo.so
This surely breaks -uninstalled.pc usage, and also caused problems in
https://github.com/rust-lang/rust/pull/90260
For example:
```
meson builddir \
--native-file vs2019-paths.txt \
--native-file vs2019-win-x64.txt \
--cross-file vs2019-paths.txt \
--cross-file vs2019-win-arm64.txt
```
This was causing the error:
> ERROR: Multiple producers for Ninja target "/path/to/vs2019-paths.txt". Please rename your targets.
Fix it by using a set() when generating the list of regen files, and
add a test for it too.
When a custom_target() uses an env, meson uses a wrapper
script to run the executable. This breaks the console: kwarg
because the wrapper script buffers the output. Fix it by setting
the verbose flag which will not buffer output.
In some cases, init variables that accept None as a sentinel and
immediately overwrite with [], are migrated to dataclass field
factories. \o/
Note: dataclasses by default cannot provide eq methods, as they then
become unhashable. In the future we may wish to opt into declaring them
frozen, instead/additionally.
Names used in init functions are sometimes pointlessly different from
the class instance attributes they are immediately assigned to. They
would make more sense if defined properly.
Since we scan all dependencies for build-only RPATHs now (which are
removed on install), we must take care not to add build-only RPATHs
pointing to directories that dependencies explicitly add -Wl,-rpath
link args for, otherwise the paths will get wiped on install.
Caught by LinuxlikeTests::test_usage_pkgconfig_prefixes
If a pkg-config dependency has multiple libraries in it, which is the
most common case when it has a Requires: directive, or when it has
multiple -l args in Libs: (rare), then we don't add -Wl,-rpath
directives to it when linking.
The existing test wasn't catching it because it was linking to
a pkgconfig file with a single library in it. Update the test to
demonstrate this.
This function was originally added for shared libraries in the source
directory, which explains the name:
https://github.com/mesonbuild/meson/pull/2397
However, since now it is also used for linking to *all* non-system
shared libraries that we link to with absolute paths:
https://github.com/mesonbuild/meson/pull/3092
But that PR is incomplete / wrong, because only adding RPATHs for
dependencies that specify a single library, which is simply
inconsistent. Things will work for some dependencies and not work for
others, with no logical reason for it.
We should add RPATHs for *all* libraries. There are no special length
limits for RPATHs that I can find.
For ELF, DT_RPATH or DT_RUNPATH are used, which are just stored in
a string table (DT_STRTAB). The maximum length is only a problem when
editing pre-existing tags.
For Mach-O, each RPATH is stored in a separate LC_RPATH entry so there
are no length issues there either.
Fixes https://github.com/mesonbuild/meson/issues/9543
Fixes https://github.com/mesonbuild/meson/issues/4372
Meson was passing only the first output and warning about it. To do this
easily, refactor construct_target_rel_path to return a list.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
For libraries installed to libdir, it's not expected to have rpath
hooked up. But for non-default libdirs, the path might not get searched
by default. `ldconfig -m <libdir>` is convenient here, as it will
programmatically add a new directory to search for shared libraries, so
the resulting installed programs work out of the box.
Include the dragonfly BSD platform name, which doesn't match the 'bsd'
catch-all pattern.
Emit a detailed deprecation warning that explains what to do instead.
Also add a unittest.
```
DEPRECATION: target prog links against shared module mymod, which is incorrect.
This will be an error in the future, so please use shared_library() for mymod instead.
If shared_module() was used for mymod because it has references to undefined symbols,
use shared_libary() with `override_options: ['b_lundef=false']` instead.
```
Fixes https://github.com/mesonbuild/meson/issues/9492
This previously worked because we were accidentally doing this via
mutation. However, doing this via mutation is not a good way to do it,
we should be explicit.
Fixes#9542
fixes#6314
in case of backend is vs2017 or vs2019 place LanguageStandard tag with stdcpp version and LanguageStandard_C tag with stdc version in .vcxproj file
We say:
> If version 4.2 or higher of the first is found, targets coverage-text,
> coverage-xml, coverage-sonarqube and coverage-html are generated.
But this is totally untrue. Make it true, by actually checking (and
not generating broken coverage commands when older versions of gcovr are
found).
Fixes#9505
Since they will never be used outside of the build directory, they do
not need to literally contain the .o files, and references will be
sufficient.
This covers a major use of object libraries, which is that the static
library would potentially take up a lot of space by including another
copy of every .o file.
Fixes#9292Fixes#8057Fixes#2129
This replaces the absolute hack of using
```
install_subdir('nonexisting', install_dir: 'share')
```
which requires you to make sure you don't accidentally or deliberately
have a completely different directory with the same name in your source
tree that is full of files you don't want installed. It also avoids
splitting the name in two and listing them in the wrong order.
You can also set the install mode of each directory component by listing
them one at a time in order, and in fact create nested structures at
all.
Fixes#1604
Properly fixes#2904
UseMultiToolTask allows parallelism inside a project, without requiring cl.exe
internal multi-threading (which meson generated projects currently can't use,
mainly due to specifying output filenames for each object).
TODO:
- think about making behaviour conditional on msbuild version / add comment
why not
The main reason for this move is to make it easier to merge the copies of
project generation. But as far as I can tell, the Microsoft.Cpp.props import
also belongs before the ItemDefinitionGroup.
Originally the order seems to have been that way, but 431a9ea664 changed it in
the course of other changes.
This patch adds a new meson built-in option for cython, allowing it to
target C++ instead of C as the intermediate language. This can, of
course, be done on a per-target basis using the `override_options`
keyword argument, or for the entire project in the project function.
There are some things in this patch that are less than ideal. One of
them is that we have to add compilers in the build layer, but there
isn't a better place to do it because of per target override_options.
There's also some design differences between Meson and setuptools, in
that Meson only allows options on a per-target rather than a per-file
granularity.
Fixes#9015
This really is more of a struct than a dict, as the types are disjoint
and they are internally handled, (ie, not from user input). This cleans
some things up, in addition I spotted a bug in the ModuleState where the
dict with the version and license is passed to a field that expects just
the version string.
When setup creates a Visual Studio environment, a message is logged
which contains a path to the build directory. Typically, this path is
converted to a relative path prior to printing. If the path cannot be
converted to a relative path (e.g., because buildpath is on a different
drive from the cwd), print out the full path instead of failing with an
unhandled exception.
When a static library B to a static library A with generated sources, B
directly references the object file corresponding to the generated source in
A. For that reference in B object_filename_from_source() is used. But A did
not specify the object file name, ending up with cl.exe's default.
Fixes: #9235
Since we changed to using a json file to avoid over long command lines
we created a situation where the generated files may not be ready when
the depscan happens. To avoid that, we need to add all of the generated
sources as order deps.
Fixes: #9258
Alias commands did not work with the vs backend, due to trying to access
target.command[0] with an empty command. Fix this by just not emitting a
CustomBuild node for alias targets - the project references are enough to
trigger the necessary actions.
Fixes: #9247
This caught a couple of cases of us doing:
```python
for i in range(len(x)):
v = x[i]
```
which are places to use enumerate instead.
It also caught a couple of cases of:
```python
assert len(x) == len(y)
for i in range(len(x)):
xv = x[i]
yv = y[i]
```
Which should instead be using zip()
```python
for xv, yv in zip(x, y):
...
```
We have a lot of these. Some of them are harmless, if unidiomatic, such
as `if (condition)`, others are potentially dangerous `assert(...)`, as
`assert(condtion)` works as expected, but `assert(condition, message)`
will result in an assertion that never triggers, as what you're actually
asserting is `bool(tuple[2])`, which will always be true.
This didn't actually catch what it's supposed to, which is cases of:
```python
for x in dict.keys():
y = dict[x]
```
But it did catch one unnecessary use of keys(), and one case where we
were doing something in an inefficient way. I've rewritten:
```python
if name.value in [x.value for x in self.kwargs.keys() if isinstance(x, IdNode)]:
```
as
``python
if any((isinstance(x, IdNode) and name.value == x.value) for x in self.kwargs):
```
Which avoids doing two iterations, one to build the list, and a
second to do a search for name.value in said list, which does a single
short circuiting walk, as any returns as soon as one check returns True.
Ben Brown