| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
Introduces std.zig.ErrorBundle which is a trivially serializeable set
of compilation errors. This is in the standard library so that both
the compiler and the build runner can use it. The idea is they will
use it to communicate compilation errors over a binary protocol.
The binary encoding of ErrorBundle is a bit problematic - I got a little
too aggressive with compaction. I need to change it in a follow-up
commit to use some indirection in the error message list, otherwise
iteration is too unergonomic. In fact it's so problematic right now that
the logic getAllErrorsAlloc() actually fails to produce a viable
ErrorBundle because it puts SourceLocation data in between the root
level ErrorMessage data.
This commit has a simplification - redundant logic for rendering AST
errors to stderr has been removed in favor of moving the logic for
lowering AST errors into AstGen. So even if we get parse errors, the
errors will get lowered into ZIR before being reported. I believe this
will be useful when working on --autofix. Either way, some redundant
brittle logic was happily deleted.
In Compilation, updateSubCompilation() is improved to properly perform
error reporting when a sub-compilation object fails. It no longer dumps
directly to stderr; instead it populates an ErrorBundle object, which
gets added to the parent one during getAllErrorsAlloc().
In package fetching code, instead of dumping directly to stderr, it now
populates an ErrorBundle object, and gets properly reported at the CLI
layer of abstraction.
|
|
There are no dir components, so you would think that this was
unreachable, however we have observed on macOS two processes racing to
do openat() with O_CREAT manifest in ENOENT.
closes #12138
|
|
Multiple processes can sit waiting for the exclusive lock at the same
time, so we want to recheck whether it needs to be updated whenever
we get an exclusive lock.
This also fixes a race condition between one process truncating the
cache file and another process reading it without atomic locking.
|
|
|
|
- Fix assertion failure if AstGen failed on a multi-module file
- Cap number of per-error reference notes and total multi-module errors each at 5
- Always put "root of package" reference notes first
Resolves: #14499
|
|
|
|
|
|
|
|
|
|
|
|
closes #14553
|
|
This finishes the work started in #14502 where atoms are owned by the
linker themselves. This now makes debug atoms fully owned by dwarf,
and no information is left stored on the decl.
|
|
* std.zig.parse is moved to std.zig.Ast.parse
* the new function has an additional parameter that requires passing
Mode.zig or Mode.zon
* moved parser.zig code to Parse.zig
* added parseZon function next to parseRoot function
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Previously, if a source file was referenced from multiple packages, it
just became owned by the first one AstGen happened to reach; this was a
problem, because it could lead to inconsistent behaviour in the compiler
based on a race condition. This could be fixed by just analyzing such
files multiple times - however, it was pointed out by Andrew that it
might make more sense to enforce files being part of at most a single
package. Having a file in multiple packages would not only impact
compile times (due to Sema having to run multiple times on potentially a
lot of code) but is also a confusing anti-pattern which more often than
not is a mistake on the part of the user.
Resolves: #13662
|
|
By @Vexu's suggestion, since fetching the name from the parent package
is error-prone and complex, and optimising Package for size isn't really
a priority.
|
|
This is a simple starting version of the optimization described in #168
where the fields are just sorted by order of descending alignment.
|
|
Closes #12403
|
|
|
|
Closes #12272
|
|
|
|
|
|
Closes #12535
Closes #12765
Closes #12927
|
|
Closes #14123
|
|
|
|
* revert changes to Module because the error set is consistent across
operating systems.
* remove duplicated Stat.fromSystem code and use a less redundant name.
* make fs.Dir.statFile follow symlinks, and avoid pointless control
flow through the posix layer.
|
|
This avoids extra syscalls.
|
|
|
|
|
|
There are still a few occurrences of "stage1" in the standard library
and self-hosted compiler source, however, these instances need a bit
more careful inspection to ensure no breakage.
|
|
|
|
closes #13535
|
|
In general, we prefer compiler code to use relative paths based on open
directory handles because this is the most portable. However, sometimes
absolute paths are used, and sometimes relative paths are used that go
up a directory.
The recent improvements in 81d2135ca6ebd71b8c121a19957c8fbf7f87125b
regressed the use case when an absolute path is used for the zig lib
directory mixed with a relative path used for the root source file. This
could happen when, for example, running the standard library tests, like
this:
stage3/bin/zig test ../lib/std/std.zig
This happened because the zig lib dir was inferred to be an absolute
directory based on the zig executable directory, while the root source
file was detected as a relative path. There was no common prefix and so
it was not determined that the std.zig file was inside the lib
directory.
This commit adds a function for resolving paths that preserves relative
path names while allowing absolute paths, and converting relative
upwards paths (e.g. "../foo") to absolute paths. This restores the
previous functionality while remaining compatible with systems such as
WASI that cannot deal with absolute paths.
|
|
The `sema.inst_map` datastructure is very often accessed. All
instructions that reference the result of other instructions does a
lookup into this field. Because of this, a significant amount of time,
is spent in `std.HashMap.get`.
This commit replaces the `HashMap` with a simpler data structure that
uses the zir indexes to index into a slice for the result. See the data
structure doc comment for more info.
|
|
|
|
|
