- Add cpuid / getXCR0 functions for the cbe to use instead of asm blocks
- Don't cast between 128 bit types during truncation
- Fixup truncation to use functions for shifts / adds
- Fixup float casts for undefined values
- Add test for 128 bit integer truncation
Move the parsing of Root from the parse function to the Parser.parseRoot
method.
Add an empty line between each rule, for consistency with grammar.y.
Ensure that normal doc-comments are always on the top, for consistency.
Add extract-grammar.zig to the tools directory; it is used to extract
the PEG grammar from parse.zig, so that it can be compared with
grammar.y.
* allow file level `union {}` to parse as tuple field
this was found while fuzzing zls.
* before this patch the input `union {}` crashed the parser. after
this, it parses correctly just like `struct {}`.
* adds behavior tests for both inputs `struct {}` and `union {}`,
checking that each becomes a file level tuple field.
these were found while fuzzing zls.
this patch prevents overflow for the following file contents and adds
tests for them.
* `enum(u32)` - causes overflow in std.zig.Ast.fullContainerDecl()
* `*x` - causes overflow in std.zig.Ast.fullPtrType()
* `**x` - causes overflow in std.zig.Ast.firstToken()
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.
Now we can reuse the table between CPU model parsers on Linux and
Windows.
Use similar parsing structure for Windows as we do for Linux. On
Windows, we rely on two entries in the registry per CPU core:
`CP 4000` and `Identifier`. Collating the data from the two allows
us recreating most of the `/proc/cpuinfo` data natively on Windows.
Additionally, we still allow for overwriting any CPU features as flagged
by pulling the feature data embedded in `SharedUserData`.
If a parse node is reserved but never set the node remains
uninitialized and can crash tools doing a linear scan of the nodes
(like ZLS) when switching on the tag.
PR #13101 recently renamed the "i386" architecture to "x86", and it
seems the specific CPU model got swept up in that. "x86" is an umbrella
term that describes a family of CPUs, and the "i386" is the oldest
supported model under that umbrella.
