Rust doesn't have a concept of dependency compile arguments, i.e.
something like headers. Dependencies are linked in and all required
metadata is provided by the linker flags.
During evaluation of codeblocks, we start off with an iteration of
nodes, and then while evaluating them we may update the global
self.current_node context. When catching and formatting errors, we
didn't take into account that the node might be updated from the
original top-level iteration.
Switch to formatting errors using self.current_node instead, to ensure
we can point at the likely most-accurate actual cause of an error.
Also update the current node in a few more places, so that function
calls always see the function call as the current node, even if the most
recently parsed node was an argument to the function call.
Fixes#11643
Extend the "common/include order" test to ensure that the build
directory is preferred over the source directory. For example,
when using `configure_file()`, the resulting file should be
preferred over a file with the same name in the source directory.
Allows getting closer to `./run_project_tests.py -- -Dwerror=true`.
- when argc and argv are not *both* used, there's a standard, compliant
mechanism to mark the variable as unused
- generated code should not build as -Werror
- more thoroughly comment out some commented code
Because we base the pickled data name on the name property of the
command being run... and for built targets, `exe.name` is always just
the name. However, for an ExternalProgram this is just whatever string
we searched for, so, NOT just the basename.
This became a bigger issue once we started using generator() with the
actual program in commit 6aeec80836,
rather than first casting it to a string, because the VS backend
*always* uses the meson_exe approach for various reasons related to VS
being VS.
Outside of that, it's difficult to actually get an ExternalProgram
object passed to meson_exe -- CustomTarget lowers it to a string,
capture is handled via argparse instead of pickling, etc.
Fixes#11593
We will still try to load `meson_options.txt` if `meson.options` doesn't
exist. Because there are some advantages to using `meson.options` even
with older versions of meson (such as better text editor handling)
we will not warn about the existence of a `meson.options` file if a
`meson_options.txt` file or symlink also exists.
The name `meson.options` was picked instead of alternative proposals,
such as `meson_options.build` for a couple of reasons:
1. meson.options is shorter
2. While the syntax is the same, only the `option()` function may be
called in meson.options, while, it may not be called in meson.build
3. While the two files share a syntax and elementary types (strings,
arrays, etc), they have different purposes: `meson.build` declares
build targets, `meson.options` declares options. This is similar to
the difference between C's `.c` and `.h` extensions.
As an implementation detail `Interpreter.option_file` has been removed,
as it is used exactly once, in the `project()` call to read the options,
and we can just calculate it there and not store it.
Fixes: #11176
It was totally subproject-unsafe, and setting a super bad example. This
is bad, because doxygen is annoying to get right and we occasionally
tell people to go use our example test case.
There is a fun nuance here, that makes doxygen unpredictably work on
some versions, and fail on others. Specifically, values must be quoted
in doxygen 1.8, but not in doxygen 1.9, or they break -- but only if the
output directory contains spaces. This was "fixed" in commit
ef91bacb7a
which actually caused it to act like an unquoted OUTPUT_DIRECTORY is not
provided at all, and then fixed for real in commit
eb3d1eb5ad
For portability, it is necessary to quote this just to be on the safe
side.
Fixes#11579
It can only be used for projects that don't have any rules at all, i.e.
they are purely using Meson to:
- configure files
- run (script?) tests
- install files that exist by the end of the setup stage
This can be useful e.g. for Meson itself, a pure python project.
This was added in f774609 to only change the access time of the
coredata file if the coredata struct actually changed. However,
this doesn't work as pickle serializations aren't guaranteed to
be stable. Instead, let's manually check if options have changed
values and skip the save if they haven't changed.
We also extend the associated unit test to cover all the option
types and to ensure that configure does get executed if one of the
options changes value.
This test is intended to test really long output, so it prints 100k
lines of stdout/stderr. It completes in two seconds on my machine, but
the default 30-second timeout is apparently too much for CI, because on
Windows we often get flaky tests due to this. e.g. we'll get within 200
lines of the end.
Bump the CI time by x2. We know this isn't particularly surprising
behavior, and allowing it to request another 30 seconds won't hang the
CI. But it will save us from some spurious failures and restarted jobs.
In commit eaf365cb3e we explicitly sorted
them for neatness, with the rationale that we were restoring intentional
behavior and we only need a set for stylistic purposes.
This actually wasn't true, because we never sorted them to begin with
(we did sort the version numbers), but sorting them is fine. The bigger
issue is that we actually used a set to avoid printing the same feature
type multiple times. Now we do print them multiple times -- because each
registered feature includes the unique node.
Fix this by using both sorted and a set.
Fix tests that should in retrospect have flagged this as an issue, but
were added later on in the same series to check something else entirely,
happen to cover this too, and were presumably copied directly from
stdout as-is...
We need to know the project minimum version before evaluating the rest
of the function. There's three basic approaches:
- try to set it inside KwargInfo
- just run a minimal version of func_project for this, then load
everything after
- drop down to the AST and set it before anything else
In order to handle FeatureNew emitted by a FunctionNode evaluated
before project() due to being inlined, such as `version: run_command()`,
only option 3 suffices, the rest all happen way too late. Since we have
just added AST handling support for erroring out, we can do that to set
the version as well.
If we add new kwargs to a function invoked on the first line, we also
need to validate the meson_version before erroring out due to unknown
kwargs. Even if the AST was successfully built.
Amusingly, we also get to improve the error message a bit. By passing
the AST node instead of an interpreter node, we get not just line
numbers, but also column offsets of the issueful meson_version. That
broke the stdout of another failing test; adapt it.
If the meson.build file is sufficiently "broken", even attempting to lex
and parse it will totally fail, and we error out without getting the
opportunity to evalaute the project() function. This can fairly easily
happen if we add new grammar to the syntax, which old versions of meson
cannot understand. Setting a minimum meson_version doesn't help, because
people with a too-old version of meson get parser errors instead of
advice about upgrading meson.
Examples of this include adding dict support to meson.
There are two general approaches to solving this issue, one of which
projects are empowered to do:
- refactor the project to place too-new syntax in a subdir() loaded
build file, so the root file can be interpreted
- teach meson to catch errors in building the initial AST, and just load
enough of the AST to check for meson_version advice
This implements the latter, allowing to future-proof the build
grammar.
These are necessary for projects outside Meson itself that want to
extend the 'meson install' functionality as meson-python does to
assemble Python package wheels from Meson projects.
Fixes#11426.
We do not need the python module's find_installation() for this, as this
does various things to set up building and installing python modules
(pure python and C-API). This functionality is already tested in the
python tests.
Elsewhere, when we just need an interpreter capable of running python
scripts in order to guarantee a useful scripting language for custom
commands, it suffices to use find_program(), which does not run an
introspection script or do module imports, and is thus faster and
a bit cleaner.
Either way, both methods are guaranteed to find the python3 interpreter,
deferring to mesonlib.python_command for that guarantee.
test "71 summary" can sometimes return the python command with the
".exe" part all uppercased for mysterious Windows reasons. Smooth this
over with ExternalProgram.
Rustc as of version 1.61.0 has support for controlling when
whole-archive linking takes place, previous to this it tried to make a
good guess about what you wanted, which worked most of the time. This is
now implemented.
Additionally, rustc makes some assumptions about library names
(specifically static names), that meson does not keep. This can be fixed
with rustc 1.67, where a new +verbatim modifier has been added. We can
then force rustc to use the name we give it. Before that, we can sneak
through `/WHOELARCHIVE:` in cases of dynamic linking (into a dll or
exe), but we can't force the archiver to do what we want (rustc
considers the archiver to be an implementation detail). The only
solution I can come up with is to copy the library to the format that
rustc expects. I've run into some issues with that as well, so we warn
in that case.
The decisions to leave static into static broken on MSVC for 1.61–1.66
was made because:
1) The work around is non-trivial, and we would have to support that
workaround for a long time
2) The number of users of Rust in Meson is small
3) The number of users of Rust in Meson on Windows, with MSVC is tiny
4) Using rustup to update rustc on windows is trivial, and solves the
problem completely
Fixes: #10723Fixes: #11247
Co-authored-by: Nirbheek Chauhan <nirbheek@centricular.com>
This method allows meson.build to introspect on the changed options.
It works by merely exposing the same set of data that is logged by
MesonApp._generate.
Fixes#10898
This adds two new methods, that are conceptually related in the same way
that `enable_auto_if` and `disable_auto_if` are. They are different
however, in that they will always replace an `auto` value with an
`enabled` or `disabled` value, or error if the feature is in the
opposite state (calling `feature(disabled).enable_if(true)`, for
example). This matters when the feature will be passed to
dependency(required : …)`, which has different behavior when passed an
enabled feature than an auto one.
The `disable_if` method will be controversial, I'm sure, since it
can be expressed via `feature.require()` (`feature.require(not
condition) == feature.disable_if(condition)`). I have two defences of
this:
1) `feature.require` is difficult to reason about, I would expect
require to be equivalent to `feature.enable_if(condition)`, not to
`feature.disable_if(not condition)`.
2) mixing `enable_if` and `disable_if` in the same call chain is much
clearer than mixing `require` and `enable_if`:
```meson
get_option('feat') \
.enable_if(foo) \
.disable_if(bar) \
.enable_if(opt)
```
vs
```meson
get_option('feat') \
.enable_if(foo) \
.require(not bar) \
.enable_if(opt)
```
In the first chain it's immediately obvious what is happening, in the
second, not so much, especially if you're not familiar with what
`require` means.
It's always been strange to me we don't have an opposite method of the
`disable_auto_if` method, but I've been pressed to find a case where we
_need_ one, because `disable_auto_if` can't be logically contorted to
work. I finally found the case where they're not equivalent: when you
don't want to convert to a boolean:
```meson
f = get_option('feat').disable_auto_if(not foo)
g = get_option('feat').enable_auto_if(foo)
dep1 = dependency('foo', required : f)
dep2 = dependency('foo', required : g)
```
gpgme has decided that config-tool is bad, which makes sense. They've
also decided that they will only install theirs, if gpg-error also
installs one, which is a bit... confusing. Anyway, it's impossible to
know whether it should or should not exist, so just accept that this
test is ready to be skipped on distros that currently no longer have
this ancient config-tool script.
Victory is within reach!
llvm-config is unsuitable for standard cross-compile,
because we need to build llvm especially for it, which
is not done is almost any distros, so, for example,
standard bootstrap chroot will be unsuitable.
This patch is trying to acheive feature parity between
config-tool searching of LLVM and CMake-based one,
which is arch-agnostic.
Signed-off-by: Konstantin <ria.freelander@gmail.com>
This test case checks stdout and demands a `dependency()` lookup fail.
The resulting error message can be different depending on whether cmake
is installed, or not. For cmake-specific tests we would simply skip the
test if cmake is not installed, but here we can just fine-tune the
pattern matching we use to determine if the test failed "correctly".
Fixes#11320
Currently Meson allow the following (Muon does not):
```meson
option('foo', type : 'boolean', value : 'true')
option('bar', type : 'integer', value : '42')
```
This is possibly a holdover from very old code, but it's a bad idea and
we should stop doing it. This deprecation is the first stop on that
journey.
We make use of allow_unknown=True here, which allows us to only look at
the common arguments in the main option parser, and then look at the
specific options in the dispatched parsers. This allows us to do more
specific checking on a per overload basis.
It is often more useful to generate shell script than dumping to stdout.
It is also important to be able to select the shell format.
Formats currently implemented:
- sh: Basic VAR=prepend_value:$VAR
- export: Same as 'sh', but also export VAR
- vscode: Same as 'sh', but without substitutions because they don't
seems to work. To be used in launch.json's envFile.
The code below this already handles being passed an Executable or
ExternalProgram, and it does it correctly, since it handles host
binaries that need an exe_wrapper correctly, while the code in the
generator paths doesn't.
The xcode backend is, like always, problematic, it doesn't handle things
the same way as the ninja and vscode backends, and generates a shell
script instead of using meson as a wrapper when needed (it seems likely
that just forcing the meson path for xcode would be better). I don't
have a working mac to develop a fix for, so I've left a todo comment
there.
Fixes: #11264
Sebastian Dröge
Josh Soref
HiPhish
Eli Schwartz
Christoph Reiter
Barnabás Pőcze
Jussi Pakkanen
Dylan Baker
Xavier Claessens
Daan De Meyer
Daniele Nicolodi
L. E. Segovia
Charles Brunet
Nirbheek Chauhan
Dudemanguy
Konstantin
Gijs Peskens
Thomas Li
Paolo Bonzini