We want to ensure that our automation builds don't pull in libraries
from crates.io, and we need --frozen support in cargo to do that. If we
don't have that support, we shouldn't build.
The base compiler check in python configure does some preprocessing,
which ensures the compiler works to some extent. Autoconf used to have
a more complete test, doing a compile/link. We do have plenty of tests
afterwards that do that anyways, but it's better if we fail early if
the toolchain fails somehow.
This refactors try_compile such that the *_compiler variable themselves
can be used to trigger compiler tests. Eventually, we'll want something
similar for preprocessing and possibly other invocations.
This also removes similar tests from build/autoconf/toolchain.m4.
--HG--
extra : rebase_source : c60d1d6e39b6bd2a377516687affd9b8932ebc12
This patch is really two separate changes.
The first change is that rust crates are large, standalone entities that
may contain multitudes of source files. It therefore doesn't make sense
to keep them in SOURCES, as we have been doing. Moving to use cargo
will require a higher-level approach, which suggests that we need a
different, higher-level representation for Rust sources in the build
system.
The representation here is to have the build system refer to things
defined in Cargo.toml files as the entities dealt with in the build
system, and let Cargo deal with the details of actually building things.
This approach means that adding a new crate to an existing library just
requires editing Rust and Cargo.toml files, rather than dealing with
moz.build, which seems more natural to Rust programmers. By having the
source files for libraries (and binaries in subsequent iterations of
this support) checked in to the tree, we can also take advantage of
Cargo.lock files.
The second is that we switch the core build system over to building via
cargo, rather than invoking rustc directly.
We also clean up a number of leftover things from the Old Way of doing
things. A number of tests are added to confirm that we'll only permit
crates to be built that have dependencies in-tree.
Until now, HAVE_64BIT_BUILD was entirely determined by a compiler check.
But we didn't run the check on e.g. artifact builds, while relying on
its result for some non-compilation related things, leading to subtle
discrepancies.
This changes the configure check to derive HAVE_64BIT_BUILD from bitness
determined by the target CPU, and double checked with a compiler check.
--HG--
extra : rebase_source : 5dc0cf2369ed4457bdd9a15736a70265a771d919
Also, now that we're using modern C++11 compilers, we can just rely on
static_assert, instead of the pile of macros used in the autoconf test.
--HG--
extra : rebase_source : 85d507da653d07e6527a971082277486e3502ea2
Currently, it returns either None or the contents of the compiler's stdout,
which is always expected to be an empty string, and is not very useful. So
instead, return True in the latter case.
--HG--
extra : rebase_source : ee69cdeab38d27178ce759591fb394da65e694ac
This patch is really two separate changes.
The first change is that rust crates are large, standalone entities that
may contain multitudes of source files. It therefore doesn't make sense
to keep them in SOURCES, as we have been doing. Moving to use cargo
will require a higher-level approach, which suggests that we need a
different, higher-level representation for Rust sources in the build
system.
The representation here is to have the build system refer to things
defined in Cargo.toml files as the entities dealt with in the build
system, and let Cargo deal with the details of actually building things.
This approach means that adding a new crate to an existing library just
requires editing Rust and Cargo.toml files, rather than dealing with
moz.build, which seems more natural to Rust programmers. By having the
source files for libraries (and binaries in subsequent iterations of
this support) checked in to the tree, we can also take advantage of
Cargo.lock files.
The second is that we switch the core build system over to building via
cargo, rather than invoking rustc directly.
We also clean up a number of leftover things from the Old Way of doing
things. A number of tests are added to confirm that we'll only permit
crates to be built that have dependencies in-tree.
Now that check_prog, through find_program, returns paths that the build
system can handle, we don't need MT to just be "mt.exe".
However, we still need the PATH to be altered for the other tools we're
not checking in python configure yet (e.g. midl). We also still need
PATH altered for the compiler itself, because for e.g. the amd_x86
version, a necessary DLL is in the amd directory, which means PATH
always needs to be altered for cl.exe.
This is hopefully more reliable than parsing just the summary line,
and makes available other keys like the commit-hash which we'd like
to pass to the debug symbol automation.
Note that the commit-hash field will have the value 'unknown' for
builds not made out of upstream git. So the key is available with
official and gecko rust builds, but not for example the current
Debian-packaged rustc.
MozReview-Commit-ID: A2G5UPs2ka2
Before bug 1289294, the impossibility to find the version of MT was only
issuing a warning. Warnings in configure are essentially useless, so
since we weren't and still aren't doing anything with the result of that
check, and since there are versions of MT that don't print out a version
number, just remove the check.
--HG--
extra : rebase_source : 4887cebf0f56ca1a297cd02ff1988809c5cb6fdf
And for GCC and clang, try to see if adding -m32, -m64 or --target
works if they don't.
--HG--
extra : rebase_source : 874bc2404a5ccc48e938bc7d9b2fe67ba625cb3e
Since bug 1264482, unknown CPU types end up triggering a ValueError
exception because of the CPU EnumString. Even if somehow the CPU is
valid, the endianness is not and would trigger a ValueError exception
as well.
So, instead of letting the exceptions happen, use a nicer failure mode
with an explicit die().
--HG--
extra : rebase_source : 68432496712075c677de4bf71ea5d420fc70c35c
This removes the unnecessary setting of c-basic-offset from all
python-mode files.
This was automatically generated using
perl -pi -e 's/; *c-basic-offset: *[0-9]+//'
... on the affected files.
The bulk of these files are moz.build files but there a few others as
well.
MozReview-Commit-ID: 2pPf3DEiZqx
--HG--
extra : rebase_source : 0a7dcac80b924174a2c429b093791148ea6ac204
Until now, it's not been possible to do something as straightforward as:
option('--foo', default=delayed_getattr(milestone, 'is_nightly'))
The reason is that option's default needs what it's given, if it's a
@depends function, to depend on --help.
But we can't have every delayed_getattr add dependencies on --help,
because that would make unwanted things to depend on --help and run
when displaying the help.
Until we can totally remove --help dependencies, this change makes the
resulting @depends function created by delayed_getattr depend on --help
if the @depends function it's given already depends on --help.
The mozconfig detection logic has bitten us on many occasions in the
past. The following changes are made to tentatively improve the
situation:
- The API is modified such that autodetection of the mozconfig has
to be a conscious decision made by the caller, and not triggered
any time there is no mozconfig given, which could be a conscious
decision of the opposite.
- mozinfo.json now stores the actual mozconfig (or lack thereof) used
during configure.
--HG--
extra : rebase_source : c7a632afd414f25daf7bbe7e1a66c3736c26e039
Per froydnj in bug 1186064 comment #23, "it makes sense to proceed with removing
MSVC 2013 support." This commit does that.
We also go a step further and require VS2015 Update 2 instead of just
update 1. This temporarily brings us down to just 1 officially supported
Visual Studio version. However, VS2015u3 was just released and is
unofficially supported.
Since MOZ_CRT is no longer set, references to it have been removed.
MozReview-Commit-ID: 8MUR6qLzQA5
--HG--
extra : rebase_source : 8f5061080a3d56dd484f9be03649fb65f0145f67
Per froydnj in bug 1186064 comment #23, "it makes sense to proceed with removing
MSVC 2013 support." This commit does that.
We also go a step further and require VS2015 Update 2 instead of just
update 1. This temporarily brings us down to just 1 officially supported
Visual Studio version. However, VS2015u3 was just released and is
unofficially supported.
Since MOZ_CRT is no longer set, references to it have been removed.
MozReview-Commit-ID: 8MUR6qLzQA5
--HG--
extra : rebase_source : 22ab4f47661ead4995d0c5261104abfb02b82aa2
Per froydnj in bug 1186064 comment #23, "it makes sense to proceed with removing
MSVC 2013 support." This commit does that.
We also go a step further and require VS2015 Update 2 instead of just
update 1. This temporarily brings us down to just 1 officially supported
Visual Studio version. However, VS2015u3 was just released and is
unofficially supported.
Since MOZ_CRT is no longer set, references to it have been removed.
MozReview-Commit-ID: 8MUR6qLzQA5
--HG--
extra : rebase_source : 94527a0be3f107991998185903f1c6cf6e190f9f
Setting MOZ_DEBUG_SYMBOLS as a define was not moved, as this value is not
checked, and exporting MOZ_DEBUG_SYMBOLS was not moved, as this would only
impact nspr, and we're no longer using the nspr build system.
MozReview-Commit-ID: EvBTunhxcsr
This patch adds support for configuring Gonk/B2G with Android-specific
build scripts. This removes duplicated code and simplifies maintenance
of B2G.
The B2G builds will now use libc++ for Gecko; instead of the obsolete
STLport. A side-effect of this patch is the removal of any compile-time
dependency on B2G's bionic library.
MozReview-Commit-ID: 7V6BmC7jlrs
This patch adds support for configuring Gonk/B2G with Android-specific
build scripts. This removes duplicated code and simplifies maintenance
of B2G.
The B2G builds will now use libc++ for Gecko; instead of the obsolete
STLport. A side-effect of this patch is the removal of any compile-time
dependency on B2G's bionic library.
MozReview-Commit-ID: 7V6BmC7jlrs
Before bug 1269513, yasm_version returned a Version object, and it
doesn't anymore, which made the assignment of _YASM_*_VERSION skipped
silently.
Then, configure would go through the yasm version checks as if they were
false and skipping over the AC_MSG_ERRORs, printing out:
/builds/slave/try-m64-0000000000000000000000/build/src/old-configure: line 19615: test: : integer expression expected
which is why this went undetected: the version checks were simply
ignored.
Some shells, however, evaluated the yasm version checks as true, hitting
the AC_MSG_ERRORs.
For most cases, this replaces a value that was set in a way that would ignore
an environment variable, so this restores behavior for values that were
set in confvars.sh.
MozReview-Commit-ID: E31hm8uKq4D
We no longer support MacOS X versions 10.6-10.8. Bumping the
default MACOSX_DEPLOYMENT_TARGET to 10.7 lets us start landing
changes which are incompatible with earlier SDKs.
Currently we build on 10.7, so moving to 10.9 must wait until
our infrastructure is reconfigured to run build jobs on more
recent MacOS versions.
MozReview-Commit-ID: B0CcWVOnnv3
We were unconditionally adding them, now actually check what the
compilers default to and add the flags if they are necessary.
This will, in the future, allow finer grained policy changes, where
we can decide that C++11 and C++14 are fine, downgrading compilers
that do C++17, etc.
At the same time, allow to enable jemalloc 4 with --enable-jemalloc=4.
MOZ_JEMALLOC4 will be deprecated later.
This also changes the semantics for freebsd, where the system jemalloc
is used, relying on MOZ_MEMORY being unset (default on freebsd) and
MOZ_JEMALLOC4 to be set. In this new setup, MOZ_JEMALLOC4 implies
--enable-jemalloc=4, which still works because of the corresponding
changes to old-configure.
While forgetting about it was warned about, having to add every new
environment option to wanted_mozconfig_variables is cumbersome. It turns
out there is a hackish way to make things work without that list, which,
all things considered, is not worse than the hacks around the
wanted_mozconfig_variables function, and are certainly an improvement as
it doesn't require an ever growing list of environment options.
At the same time, we improve things slightly by deriving HOST_CC from CC
in a smarter way, as well as CXX from CC, which we weren't doing
previously.
Many related things are not moved at the same time to keep the patch
somehow "small".
So far, everything was essentially executed at "declaration". This
made the sandbox code simpler, but to improve on the tooling around
python configure (for tests and introspection), we need to have more
flexibility, which executing everything at declaration doesn't give.
With this change, only @depends functions depending on --help, as
well as templates, are executed at the moment the moz.configure
files are included in the sandbox. The remainder is executed at the
end.
- HW_THREADS doesn't appear to be doing anything.
- USE_ARM_KUSER is not used since bug 675078. This also removes the configure
flag that sets it.
- HAVE_SETBUF and HAVE_SNPRINTF are leftover from bug 944935.
- MOZ_MEMORY_GONK is leftover from bug 804303.
- DEVELOPER_OPTIONS, INTEL_CC, INTEL_CXX, MOZ_ENABLE_QTMOBILITY,
GTK_CONFIG are or even were never used outside configure.
- MOZ_PROFILELOCKING which gradually became a no-op over the years. This
also removes the configure flag that sets it.
- XULRUNNER_STUB_NAME is xulrunner-only, and xulrunner is gone. This
also removes the configure flag that sets it.
- The only use of MOZ_CAN_RUN_PROGRAMS was removed in bug 780561.
- AR_LIST and AR_DELETE have not been used since bug 584474.
- MOZ_COMPONENT_NSPR_LIBS is leftover from bug 1036894.
- MOZ_PNG_ARM_NEON_CHECK is not used since bug 980488.
- MOZ_WEBRTC_LEAKING_TESTS has been no-oped by bug 825510.
- VPX_NEED_OBJ_INT_EXTRACT and NO_INTEGRATED_AS_CFLAGS are not used since
bug 1151175.
- WCHAR_CFLAGS is not used since bug 904985.
Because closing the handler is not enough (it eventually reopens
itself), the dumping of config.log overwrites the file, and confuses
the process of reading it to dump it in the first place, showing
incomplete logs. The intent from the start was that nothing would be
logged in the FileHandler, so on top of closing it, actively remove it.
When reading config.log, with old-configure output, we may get non-ascii
strings, but that currently fails because we're using plain open() to
read it. So use encoded_open() instead (which does the same job for
other files in the same script).
Because the build system can be encapsulated in mach, python configure
can have a pipe as stdout/stderr, and in that case, sys.stdout/stderr
have an ascii encoding, failing to print out anything that doesn't
fit in ascii, consequently failing to print the things we've read from
config.log. So reopen stdout and stderr with the right encoding in
the configure output handler.
Gonk, Android, and the generic cross-compilation setup all were using a
different yet similar way to prefix the toolchain. The latter was even
wrong, since the target and target alias usually don't match actual
toolchain prefixes (which don't include the machine part of the target).
Note that this removes force-setting cross_compiling to yes in
old-configure, which wasn't working because every AC_TRY_COMPILE
resets it with $ac_cv_prog_cc_cross or $ac_cv_prog_cxx_cross.
The initial goal of templates was to provide a way to write shorter
constructs for some generic tasks during configure. But the limitations
of the sandbox and the properties of templates made them used for more
general functions.
Consequently, this led to templates having to be available from
anywhere, which, in turn, led to difficult to introspect constructs.
With bug 1257823, we've made almost everything use set_config and
similar functions from the global scope, but we don't enforce that
those primitives are only used at the global scope.
This change does that: it enforces that primitives are only used at
the global scope. Or in templates.
Now, since templates were used for other purposes than generic uses
of other primitives, we now allow non-template functions to be declared.
Those can be used everywhere, but don't have access to the sandbox
primitives.
So far, we've been using the lowercase of the variable name, but it's
not enough for some special cases. Those special cases could do their
own business, but then, they'd have to duplicate 90% of check_prog,
which is less desirable.
While DummyFunction is descriptive of what the instances are (and they
can't even be called), the various uses of isintance(obj, DummyFunction)
are kind of confusing, especially when they are in moz.configure land
(and this bug is about to add another one).
And since the file is also used for old-configure, close our handle on
the file before spawning old-configure, and make old-configure append
there instead of truncating the file.
The reason the "checking" string always appears is that @depends
functions are always called, regardless of the value of the dependency.
This introduces a new decorator @depends_true, which works like
@depends, but the decorated function is not called unless one of the
dependency value resolves to True.
The new decorator can also be used to replace many cases where we do
@depends(foo)
def bar(foo):
if foo:
...
For the same reasons as set_config is being moved to the global scope,
we're moving set_define to the global scope here. An additional change
is that set_define is now part of the sandbox itself instead of being
defined within the sandbox, which makes it share the implementation
details with set_config.
The way set_config is set currently makes it difficult to introspect
moz.configure files to know what configuration items are being set,
because they're hidden in the control flow of functions.
This makes some of the moz.configure more convoluted, but this is why
there are templates, and we can improve the recurring cases afterwards.
The way functions are being sandboxed in moz.configure land is that
their global namespace is being replaced with a limited and identifiable
dict. And we avoid re-wrapping a function that already received this
treatment.
The problem is that template functions have their global namespace
replaced, and any function that is defined within the template inherits
that global namespace. So when it comes time to wrap those functions
defined in templates with e.g. depends, we detect that they're already
wrapped although they are not, because we look if their global namespace
is of the recognizable type we use when replacing it.
So instead of looking at the global namespace type, keep track of all
functions that are wrapped.
This also adds a GRADLE_FLAGS environment variable for use in
automation.
Manually tested.
MozReview-Commit-ID: 8nDkqz2VnJn
--HG--
extra : rebase_source : 32626a7dc0c0a6a440e300d92c31670f14319325
extra : amend_source : fe134e25f079851b4c648b53a7a485ee20c15c18
Currently, if a @depends function doesn't have a return statement or
return None, a result is automatically set from all the set_config()
called from the function.
As we're going to move set_config to the global namespace, and as this
feature is only used once, and it's only used for something that was
written before ReadOnlyNamespace was exposed to the sandbox, we can
"safely" get rid of it.
This change adds a `Version` type to moz.configure which is a small
wrapper around `distutils.version.Version`. It's suitable for wrapping
version numbers in configure checks and doing equality or greater-than
less-than comparisons in a sensible way.
MozReview-Commit-ID: BOL6yvemulG
--HG--
extra : rebase_source : 3b463eac0499086f8acffda0d01418b6ab17f3d6
extra : amend_source : aebd6e40c408d9f868623b2f53fcdf7455e2fff5
In Python parens around an expression without a trailing comma is not a tuple,
so ('foo') == 'foo'. This is really easy to screw up with check_progs, which
coerces progs to a list and would give you ['f','o','o'] in this case. This
patch enforces that the progs argument is a tuple or list and errors if it
is not.
MozReview-Commit-ID: 7BJZuF9B8D5
--HG--
extra : rebase_source : 5db9a6b4bb9ef7195c2513407810093bff5e9174
extra : amend_source : f67ab46c2ac00a2a95cfc67e9763ac12b690ac14
old-configure and js/src/old-configure interestingly didn't handle both
the same way. But vtune support is only actually implemented in js/src,
so only the rules from js/src/old-configure matter (nothing was
enforcing the decistion from old-configure to js/src/old-configure), and
this is what is implemented here.
Until we stop relying on the raw_cpu and raw_os values from target and
host, we need to keep normalizing them like old-configure.in did, which
involves filtering them through config.sub.
Bug 1038639 removed the option, but since autoconf doesn't barf for
unknown options, the option MOZ_ARG_WITH_STRING was kept to emit a
AC_MSG_ERROR. This is not necessary anymore.
This aligns with the triplets used by clang/llvm. Technically, this
won't break iOS builds still using -darwin triplets until we move
MOZ_IOS_SDK to moz.configure and actively reject non iOS targets with
the iOS sdk.
Also allow to distinguish iOS and OSX with target.os.
Because some of the existing mozconfigs may be setting some variables,
we need to inject those that are handled by moz.configure now. It likely
doesn't matter for the variables currently in moz.configure, but it will
soon become important when more things are moved to moz.configure.
In fact, it is necessary for GENISOIMAGE and DSYMUTIL that we're going
to move in this bug, set in automation mozconfigs.
The implementation is cumbersome and quite horrible. We could do better
by changing the execution model in mozbuild.configure, which is probably
necessary for other reasons as well, but that requires more work and
testing.
It's an opt-in flag that allows to display where the build is in
terminal window titles. The fact that it's opt-in and likely unknown
makes it very low-value, and the fact that it was added in an era where
builds were not very well parallelized made it have a meaning, but now
that builds are parallelized, its meaningfulness is diminished.
Let's just remove it.
With all the things that still depend on all the variables derived from
--host and --target in both old-configure and moz.build, we still need
to keep variables such as OS_ARCH, OS_TARGET, CPU_ARCH, OS_TEST, etc.
Eventually, we'd settle on the output of split_triplet.
This /tries/ to preserve the current values for all these variables,
while also trying to make things a little more consistent. It also
effectively rejects OSes such as HPUX or AIX, because it is unclear
the decades old accumulated scripts related to them still do anything
useful, and we might as well have them start again from scratch, which,
in the coming weeks, will be even easier.
The data we get out of mozconfig can be unicode, and needs to be in a
form the shell used to run old-configure will be able to recognize,
which is likely utf-8 on UNIX (that's what we settled on), and mbcs on
Windows.
So far, we've been passing down all configure_args from mozconfig as
well as every flag appearing on sys.argv. This is overly broad and
causes problems for some options, like --enable-application.
However, we don't need all these options to be passed.
For the top-level old-configure, we need to pass the flags it can
handle, as well as the flags that we want passed down to
js/src/configure.
For js/src/old-configure, we only need to pass the flags it can handle.
The flags an old-configure can handle is defined by the list of flags
in @old_configure_options. The list of flags to pass down to
js/src/configure is defined by extra_old_configure_args.
And since the mozconfig configure_args are being injected into python
configure processing, the list of values we get in old_configure includes
the mozconfig configure_args.
While the long term goal is that js and top-level use the same configure
and the same overall setup, including the possibility to use mozconfigs,
figuring out what we want to do wrt mozconfig vs. command line and
environment variable is not a clear-cut case, and it's more important to
fix the immediate problem mozconfig causes to js developers by
"temporarily" returning to the previous behavior of not loading the
mozconfig for the js configure.
This was already done in the case of running it as a subconfigure, this
extends the exception. Unfortunately, there is no direct way to tell
whether the running configure is the js configure. The indirect way is
to look at the OLD_CONFIGURE path, which points to
js/src/old-configure.
I expect we'll have figured things out for mozconfigs well before
old-configure dies.
The implementation is a bit circumvoluted, but we do need to share
options between the top-level and js/src configures, possibly with
different defaults, and to properly pass things down from one to
the other. Until we are further down the road and can actually merge
both configures, this is a necessary evil.
Because --enable-application is the current way to do things, transpose
it to configure.py, but since --enable-application=js doesn't make
sense, make it an alias of a new --enable-project option.
This only partially moves --enable-application out of old-configure.in
because there are a lot of other things intertwined with it.
This moves all the reading mozconfig, finding autoconf, refreshing the
old configure, and running the old configure into sandboxed
moz.configure. This effectively bootstraps the sandboxed python configure.
Mike Hommey
Nathan Froyd
Wes Kocher
Carsten "Tomcat" Book
Jan Beich
Ralph Giles
Fabrice Desré
Tom Tromey
Benjamin Smedberg
Gregory Szorc
Ted Mielczarek
Chris Manchester
Thomas Zimmermann
Sebastian Hengst
Christoph Diehl
Ms2ger
Mike Shal
Hector Zhao
Nick Alexander
Blake Winton
Nicolas B. Pierron
Patrick Cloke