Warnings have a location node object (with subdir and lineno
attributes), which is passed as a location: kwarg to mlog.warning() and
formatted in _log_error().
Re-purpose the subdir attribute (path relative to the source root dir,
with an implied filename of 'meson.build'), which is stored into the
node by parser(), to contain a pathname.
(Properly I should rename 'subdir' -> 'file' everywhere, but that's a
lot of churn just to see if this works)
Notes:
The warning location node may also have a colno attribute, which is
currently ignored by _log_error().
We can't currently issue warnings with locations in meson_options.txt
because the filename isn't part of the location (as it's assumed to be
'meson.build).
A MesonException has file, lineno and colno attributes, which get
formatted as a location in mlog.exception().
The file attribute got changed from a path relative to the root source
directory to a pathname (absolute or relative to cwd) in one place in
commit b8fbbf59. Adjust all the other places the file attribute is set
to match.
Also:
Setting MesonException.file seems to be missing in the case where Parser
returned a non-CodeBlockNode object. Fortunately, that looks like it's
unreachable, but add it just in case.
Currently, giscanner and the gicompiler paths are only scanned via pkg-config
if they are first found in the host path.
Add a else statement to fix this oversite.
Currently, meson hard codes the paths of these binaries which results in
cross-compiled environments to run the host versions of these tools.
However, GObject-introspection provides the appropriate paths to these
utilities via pkg-config
find_program is needed in the case g-i is built as a subproject. If
g-ir-scanner or g-ir-compiler are in the build or source directory use those.
If they aren't found in the source directory, use the results from pkg-config.
When a source file for a library is changed without adding new extern
symbols, only that library should be rebuilt. Nothing that uses it
should be relinked.
Along the way, also remove trailing `.` in all Ninja rule
descriptions. It's very confusing to see messages like:
```
Linking target mylib.dll.
```
It's confusing that the period at the end of that is not part of the
filename. Instead of removing that period manually in the tests (which
feels wrong!) just don't print it at all.
We actually use this while linking on Windows, and hence we need to
extract symbols from this file, and not the DLL.
However, we cannot pass it instead of the DLL because it's an optional
output of the compiler. It will not be written out at all if there are
no symbols in the DLL, and we cannot know that at configure time. This
means we cannot describe it as an output of any ninja target, or the
input of any ninja target. We must pass it as an argument without
semantic meaning.
On Windows, the basename is used to determine the name of the PDB
file. So for a project called myproject, we will create myproject.dll
and myproject.exe, both of which will have myproject.pdb. This is
a file collision. Instead, append `_test`, similar to the C# template.
Fixes AllPlatformTest.test_templates on MSVC. This became a hard error
when we started listing PDBs in the implicit outputs list of ninja
targets.
Do the same for a test that was making the same mistake.
This is more correct, and forces the target(s) to be rebuilt if the
PDB files are missing. Increases the minimum required Ninja to 1.7,
which is available in Ubuntu 16.04 under backports.
We can't do the same for import libraries, because it is impossible
for us to know at configure time whether or not an import library will
be generated for a given DLL.
This makes two basic changes, 1 it moves the name of the linker into the
linker class, this should reduce the number of errors and typos, and
ensure that a linker always has one name. This then renames the linkers
to have more consistent names.
Posix/gnu linkers are called ld.<name>: ld.gold, ld.lld, ld.solaris.
Apple linkers are renamed ld64.
This allows users to disable writing out the inbuilt variables to
the pkg-config file as they might actualy not be required.
One reason to have this is for architecture-independent pkg-config
files in projects which also have architecture-dependent outputs.
For example : https://gitlab.freedesktop.org/wayland/weston/issues/269Fixes#4011
This PR significantly improves the handling of the boost library
tags and also ensures that the found libraries are always compatible
(have the same ABI tag). The current setup can also be extended to
filter for additional features (static linking with the runtime, etc.).
Additionally, BOOST_ROOT is better supported (it is now guaranteed
that all found files belong to a single root directory).
Finally, boost.py is now fully annotated (and checked with mypy).
Do this by tracking CMAKE_CURRENT_{SOURCE,BINARY}_DIR variables.
This is achieved by injecting CMake code with CMAKE_PROJECT_INCLUDE
and overriding some builtin functions with a wrapper that adds
additional trace information.
This is a significant speed-up on Windows because terminals are
slow to print things out.
Speed-up in gst-build on Windows:
```
meson install:
before: 5.1 seconds
after: 4.0 seconds
```
In `--only-changed` mode, we were spending half our time inside
`makedirs()`, which is a waste in the case when only a few files have
changed.
Speed-up on Windows with gst-build:
```
meson install --only-changed
before: 1.6 seconds
after: 0.9 seconds
```
This gives a significant speedup in large projects such as gst-build
since now we only search for the tool once. Speed-up on Windows:
```
meson install:
before: 15.3 seconds
after: 5.4 seconds
meson install --only-changed:
before: 11.6 seconds
after: 2.7 seconds
```
If we preserve the existing file, we did not install anything. In that
case, the output now looks like:
```
ninja: Entering directory `.'
ninja: no work to do.
Nothing to install.
Preserved 1732 unchanged files, see meson-logs\install-log.txt for the full list
```
This also avoids an `isfile` lookup, which shaves off a few seconds in
gst-build install.