fix_jar() tries to remove an existing Class-Path entry from the jar
manifest by postprocessing the manifest and passing it to `jar -um`.
However, `jar -um` can only add/replace manifest entries, not remove
them, and it also complains loudly when replacing an entry:
Dec 13, 2022 7:11:19 PM java.util.jar.Attributes read
WARNING: Duplicate name in Manifest: Manifest-Version.
Ensure that the manifest does not have duplicate entries, and
that blank lines separate individual sections in both your
manifest and in the META-INF/MANIFEST.MF entry in the jar file.
Thus fix_jar() produces one such warning for each entry in the manifest
and accomplishes nothing else.
Use jar -uM instead. This completely removes the manifest from the jar
and allows adding it back as a normal zip member, fixing fix_jar() and
avoiding the warnings.
Fixes: https://github.com/mesonbuild/meson/issues/10491
Fixes: c70a051e93 ("java: remove manifest classpath from installed jar")
Due to a deficiency in upstream clang-format, our automatic target for
`ninja clang-format-check` runs clang-format, then compares the bytes of
the file before and after to see if anything changed. If it did change,
we rewrite the file back to its original form and error out.
Since clang-format 10, there is an option to report warnings instead of
writing the reformatted file, and also, to make those warnings fatal.
This is a much better user experience, to see *what* is wrong, not just
that something is wrong, and also gets rid of a pretty gross "modify
your files when you didn't ask for it" behavior that is vulnerable to
getting interrupted.
Let's switch over to the new approach, if we can.
Which adds the `use-set-for-membership` check. It's generally faster in
python to use a set with the `in` keyword, because it's a hash check
instead of a linear walk, this is especially true with strings, where
it's actually O(n^2), one loop over the container, and an inner loop of
the strings (as string comparison works by checking that `a[n] == b[n]`,
in a loop).
Also, I'm tired of complaining about this in reviews, let the tools do
it for me :)
Those classes are used by wrapper scripts and we should not have to
import the rest of mesonlib, build.py, and all their dependencies for
that.
This renames mesonlib/ directory to utils/ and add a mesonlib.py module
that imports everything from utils/ to not have to change `import
mesonlib` everywhere. It allows to import utils.core without importing
the rest of mesonlib.
This does force a number of uses of `# pylint: disable` comments, but it
also finds a couple of useless global uses and one place (in the
previous commit) that an easy refactor removes the use of global. Global
is a code smell, so forcing adding a comment to disable helps force
developers to really consider if what they're doing is a good idea.
The `global` statement is only needed to assign to global variables, not
read or mutate them. So calling `global.mutate()` is fine, but not
`var = foo`, which would otherwise shadow `var`.
Meson internally knows about many languages and tools, and *FLAGS
variables, and which languages to use them for. Instead of duplicating
this logic, import it from mesonbuild.*
This logic was originally standalone, but now that it is merged into the
Meson tree we can have a single source of truth.
In commit 4ca9a16288 we added unreliable
support (it warns you if you try it) for gcc-compatible treatment of
uppercase-C files being C++ instead of C. In order to handle it
correctly, we needed to evaluate can-compile by special-casing "C" to
avoid lowercasing it for comparisons.
This didn't cover all cases where we check if "C" is a C++ language
file. We also straight-up check the language of a file (rather than
working backwards to see if a C++ compiler can compile it) when doing
module scanning, and this needs to special-case "C" as well.
We also had one case where we only checked lowercase fortran extensions,
but not lowercase C++ extensions. While we are at it, use lowercase for
C++ as well, except the "C" special case.
Fixes#10629
The old implementation assumed a path is of Windows iff the second
character is a colon. However, that is not always true.
First, a colon can be included in a non-Windows path. For example, it is
totally fine to have a directory named ':' on Linux 5.17.0 tmpfs.
Second, a Windows path may start with \\ for UNC or extended length.
Use pathlib to handle all of these cases.
The check for if the project supports the -j flag was needlessly
complex. We support two types of project:
- waf, always supports -j
- make, if GNU, supports -j
We never checked waf, and the make check assumed that the entire
command, rather than just the last component, was "make". It also
neglects "gmake".
Since any possible build command *may* support -j, always run the
--version check. Detect either build command in the output.
- Remove duplicated code in mdevenv.py
- Change the limit to 1024 instead of 2048 which is what has been
tested.
- Skip shortening if it is already short enough.
- Skip shortening with wine >= 6.4 which does not seems to have that
limitation any more.
- Downgrade exception to warning in the case WINEPATH cannot be
shortened under 1024 chars, it is possible that it will still work.
gcovr will read this file anyway, but if it exists we don't need to
assume that the project wishes to exclude subprojects/ -- they can
determine that themselves.
Fixes#3287Closes#9761
lcov doesn't read the config file by default, but we can do the smart
thing here.
Fixes#4628
This way, we can ensure that gtk-doc parses the --path argument
correctly when passed in from the cmd.exe console, since ':' is normally
used to denote that a drive is being used on local paths.
We can immediately short-circuit if there is no destdir, as we simply
return the prefix unchanged.
If there is some kind of destdir and the prefix contains a drive letter,
then no matter what we need to remove the drive letter before joining.
Technically, if the destdir is a relative path e.g. `destdir\` and
`C:\prefix`, we should still install to `destdir\prefix` without the
drive letter.
But... we also guarantee that destdir is an absolute path (or empty)
anyway. And even if we didn't, non-absolute destdir is a broken concept
for a variety of complicated reasons. So none of this matters in
practice.
One way or another, we don't need to actually check whether destdir is
an absolute path before cutting off the prefix drive letter.
These are only used for type checking, so don't bother importing them at
runtime.
Generally add future annotations at the same time, to make sure that
existing uses of these imports don't need to be quoted.
This reverts commit e257a870fe.
The PR adding this command had infinitely hanging CI, and now that it is
merged to master we cannot get any CI on any PR to succeed.
Don't assume itstool, msgfmt et al. are just magically on the path.
Normally for commands being processed in build.ninja we'd look up the
program in order to run it. Offer the same guarantee for programs being
passed through an awkward script wrapper.
This is basically a rewrite of the gnome.yelp target to remove the
ad-hoc script, which generates multiple issues, including meson
not knowing which files were installed.
Closes#7653Closes#9539Closes#6916Closes#2775Closes#7034Closes#1052
Related #9105
Related #1601
When installing with 'meson install --quiet' I'd get the following output:
This file does not have an rpath.
This file does not have a runpath.
(It turns out that of the couple hundred of binaries that are installed,
this message was generated for /usr/lib/systemd/boot/efi/linuxx64.elf.stub.)
There doesn't seem to be any good reason for this output by default. But those
functions can still be used for debugging. Under a debugger, returning the
string is just as useful as printing it, but more flexible. So let's suppress
printing of anything by default, but keep the extractor functions.
The code was somewhat inconsistent wrt. to when .decode() was done. But it
seems that we'll get can expect a decodable text string in all cases, so
just call .decode() everywhere, because it's nicer to print decoded strings.
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
Benjamin Gilbert
Eli Schwartz
Dylan Baker
Jussi Pakkanen
Xavier Claessens
Alf Henrik Sauge
Akihiko Odaki
Staz M
Seungha Yang
Chun-wei Fan
Thomas Klausner
Matthias Klumpp
Pablo Correa Gómez
Zbigniew Jędrzejewski-Szmek
Christian Clauss