Previously we expected to only find decl refs in a `foo.bar.baz`
type of expression. This would crash when trying to render something
like `@typeInfo(T).Int.bits`. We now properly account for builtins
and other components.
As part of the Opaque Pointers upgrade documentation, LLVM says that the
function LLVMGetGEPSourceElementType() can be used to obtain element
type information in lieu of LLVMGetElementType(), however, this function
actually returns the struct type, not the field type. The GEP
instruction does store the information we need, however, this is not
exposed in the C API. It seems like they accidentally exposed the wrong
field, because one would never need the struct type since one must
already pass it directly to the GEP instruction, so one will always have
it handy, whereas one will usually not have the field type handy.
Removed the copy of param_names inside of Fn and changed to
implementation of getParamName to fetch to parameter name from the ZIR.
The signature of getParamName was also changed to take an additional
*Module argument.
From https://man7.org/linux/man-pages/man7/inotify.7.html
> **IN_MASK_CREATE** (since Linux 4.18)
>
> Watch pathname only if it does not already have a watch associated with it; the error EEXIST results if pathname is already being watched.
This change provides a basic implementation of #2349 for stage2. There's
still quite a lot of work before this logic is as complete as what's in
Clang (b364535304/clang/lib/CodeGen/CGStmt.cpp (L2304-L2795)),
particularly considering the diversity of constraints across targets.
It's probably not worth doing the complete work until there's a clearer
picture for constraints in Zig's future dedicated ASM syntax, but at
least this gives us a small improvement for now.
As a bonus, this also fixes a bug with how we were handling `_`
identifiers.
If you have multiple llvm-config executables in your path, and all of
them cause failures, then only the last failure will be printed. This
can cause confusion when the multiple llvm-config executables are from
different major LLVM versions, i.e. LLVM 13 and 14, which might mask an
error that happened on the LLVM 14 llvm-config with an unrelated error.
This commit makes it so that all errors are collected into a list and
printed all at once; this way, you can see how each llvm-config
executable failed to configure properly. Note that the failures still
won't be printed if a successful configuration is found.
We now warn the user if config.h could not be located.
This also updates the search to stop early upon encountering a
`.git` directory, so that we avoid recursing outside of the zig
source if possible.
