By requiring the source file to be null-terminated, we avoid extra
branching while simplifying the logic at the same time.
Running ast-check on a large zig source file (udivmodti4_test.zig),
master branch compared to this commit:
* 4% faster wall clock
* 7% fewer cache misses
* 1% fewer branches
The motivation for this commit is that there exists source files which
produce ast-check errors, but crash stage1 or otherwise trigger stage1
bugs. Previously to this commit, Zig would run AstGen, collect the
compile errors, run stage1, report stage1 compile errors and exit if
any, and then report AstGen compile errors.
The main change in this commit is to report AstGen errors prior to
invoking stage1, and in fact if any AstGen errors occur, do not invoke
stage1 at all.
This caused most of the compile error tests to fail due to things such
as unused local variables and mismatched stage1/stage2 error messages.
It was taking a long time to update the test cases one-by-one, so I
took this opportunity to unify the stage1 and stage2 testing harness,
specifically with regards to compile errors. In this way we can start
keeping track of which tests pass for 1, 2, or both.
`zig build test-compile-errors` no longer works; it is now integrated
into `zig build test-stage2`.
This is one step closer to executing compile error tests in parallel; in
fact the ThreadPool object is already in scope.
There are some cases where the stage1 compile errors were actually
better; those are left failing in this commit, to be addressed in a
follow-up commit.
Other changes in this commit:
* build.zig: improve support for -Dstage1 used with the test step.
* AstGen: minor cosmetic changes to error messages.
* stage2: add -fstage1 and -fno-stage1 flags. This now allows one to
download a binary of the zig compiler and use the llvm backend of
self-hosted. This was also needed for hooking up the test harness.
However, I realized that stage1 calls exit() and also has memory
leaks, so had to complicate the test harness by not using this flag
after all and instead invoking as a child process.
- These CLI flags will disappear once we start shipping the
self-hosted compiler as the main compiler. Until then, they can be
used to try out the work-in-progress stage2.
* stage2: select the LLVM backend by default for release modes, as long
as the target architecture is supported by LLVM.
* test harness: support setting the optimize mode
* Add command line help for "-mexec-model"
* Define WasmExecModel enum in std.builtin.
* Drop the support for the old crt1.o in favor of crt1-command.o
Signed-off-by: Takeshi Yoneda <takeshi@tetrate.io>
This logic was a workaround to prevent cache deadlocks which happened
from always using exclusive file locks. Now that the Cache system
supports sharing cached artifacts, this workaround is no longer needed.
Closes#7596
Normally we rely on importing std to in turn import the root
in the start code, but when using the stage1 won't happen,
so in order to run AstGen on the root we put it into the
import_table here.
There is now a distinction between `@import` with a .zig extension and
without. Without a .zig extension it assumes it is a package name, and
returns error.PackageNotFound if not mapped into the package table.
This change reduces the amount of divergence in the compiler's main
pipeline logic enough to run AstGen for all files in the compilation,
regardless of whether the stage1 or stage2 backend is being used.
Practically, this means that all Zig code is subject to new compile
errors, such as unused local variables.
Additionally:
* remove leftover unsound asserts from recent hash map changes
* fix sub-Compilation errors not indenting correctly
Previously, Zig did not properly communicate the target CPU features for
RISC-V to clang assembler, because Clang has a different way to pass CPU
features for C code and for assembly code. This commit makes Zig pass a
RISC-V -march flag in order to communicate CPU features to Clang when
compiling assembly files.
* stage1 backend allows configuring the uwtables function attr
via a flag rather than its own logic.
* stage2 defaults to enabling uwtable attr when
linking libunwind, or always on windows
* stage2 makes link_eh_frame_hdr true automatically if uwtable
attr is set to be on for zig functions
* CLI: add -funwind-tables and -fno-unwind-tables to allow the user to
override the defaults.
* hook it up to `zig cc`
closes#9046
closes#9034
These options were listed under the
"Debug Options (Zig Compiler Development)" heading. Anything in this
section should be considered unstable and can be modified at any time
at any developer's discretion.
* then, in `link/Wasm.zig` map `CRTFile` to full emulated libs name
* move logic for removing any mention of WASI snapshot
`wasi_snapshot_preview1` from `Compilation.zig` into `link/Wasm.zig`
Move parsing of system libs into `main.zig` next to where we decide
if we should link libC, and, if targeting WASI, if the specified
libname equals one of the emulated components, save it on the side
and remove it from the system libs. Then, build *only* those parts
of WASI libc that were preserved in the previous step.
This also fixes building of different crt1 bits needed to support
reactors and commands.
- hash/eql functions moved into a Context object
- *Context functions pass an explicit context
- *Adapted functions pass specialized keys and contexts
- new getPtr() function returns a pointer to value
- remove functions renamed to fetchRemove
- new remove functions return bool
- removeAssertDiscard deleted, use assert(remove(...)) instead
- Keys and values are stored in separate arrays
- Entry is now {*K, *V}, the new KV is {K, V}
- BufSet/BufMap functions renamed to match other set/map types
- fixed iterating-while-modifying bug in src/link/C.zig
Clang has a completely inconsistent CLI for its integrated assembler for
each target architecture. For x86_64, for example, it does not accept
an -mcpu parameter, and emits "warning: unused parameter". However, for
ARM, -mcpu is needed in order to properly lower assembly to machine code
instructions (see new standalone test case provided thanks to @g-w1).
This is a compromise between
b8f85a805bf61ae11d6ee2bd6d8356fbc98ee3ba and
afb9f695b1bdbf81185e7d55d5783bcbab880989.
This reverts commit afb9f695b1bdbf81185e7d55d5783bcbab880989.
I don't think this was ever actually verified to fix the thing it
purported to, and it started causing warnings for unused command line
parameters.
- more support for linux, android, freebsd, netbsd, openbsd, dragonfly
- centralize musl utils; musl logic is no longer intertwined with csu
- fix musl compilation to build crti/crtn for full archs list
- fix openbsd to support `zig build-lib -dynamic`
- initial dragonfly linking success (with a warning)
ancillary:
- fix emutls (openbsd) tests to use `try`
This matches the behaviour of other languages and leaves us
the ability to create actual static Wasm archives with
```
zig build-lib -static some.zig
```
which can then be combined with other Wasm object files and linked
into either a Wasm lib or executable using `wasm-ld`.
Update langref to reflect the fact we now ship WASI libc.
Rename include dir to match the convention:
from `wasm32-wasi` to `wasm-wasi-musl`
Add building stubs which will be used to build and cache WASI
libc sysroot.
We do need to link the system libc if linking system libraries as
they may potentially be compiled against e.g. a newer glibc version
than zig can provide. However if not linking system libraries, using
the zig provided libc is more reliable as it does not depend on any
quirks of the host system or being able to invoke the system cc to
find include dirs.
When scanDecls happens, we create stub Decl objects that
have not been semantically analyzed. When they get referenced,
they get semantically analyzed.
Before this commit, when they got unreferenced, they were completely
deleted, including deleted from the containing Namespace.
However, if the update did not cause the containing Namespace to get
deleted, for example, if `std.builtin.ExportOptions` is no longer
referenced, but `std.builtin` is still referenced, and then `ExportOptions`
gets referenced again, the Namespace would be incorrectly missing the
Decl, so we get an incorrect "no such member" error.
The solution is to, when dealing with a no longer referenced Decl
objects during an update, clear them to the state they would be in
on a fresh scanDecl, rather than completely deleting them.
Previously, stage2 used a global decl_table for all Decl objects, keyed
by a 16-byte name hash that was hopefully unique. Now, there is a tree
of Namespace objects that own their named Decl objects.
