diff options
| author | Tom Tromey <tom@tromey.com> | 2025-03-17 12:57:34 -0600 |
|---|---|---|
| committer | Tom Tromey <tom@tromey.com> | 2025-04-03 08:16:23 -0600 |
| commit | 254809b28f1c9c2a43d25ca1a71a5603d4d1bda6 (patch) | |
| tree | dfda90b5ff834c11bf6a9e678f2a0108dff9c7b8 /gdb/testsuite/gdb.python/py-watchpoint.py | |
| parent | 0326df1b9b1f69886dfd0d4d4aba739db5bdbd65 (diff) | |
| download | binutils-254809b28f1c9c2a43d25ca1a71a5603d4d1bda6.zip binutils-254809b28f1c9c2a43d25ca1a71a5603d4d1bda6.tar.gz binutils-254809b28f1c9c2a43d25ca1a71a5603d4d1bda6.tar.bz2 | |
Fix pp.rs test for gccrs
gccrs still can't process all of gdb's Rust tests, but I did manage to
manually test it on a few. In addition to filing some bug reports, I
came up with this patch.
There are two fixes here. First, gccrs emits tuple field names as
integers ("0", "1", etc) whereas rustc uses a leading double
underscore ("__0", "__1", etc). This patch changes gdb to accept the
gccrs output, which IMO makes sense (and for which there's already a
rustc feature request).
Second, it changes rust_struct_anon::evaluate to use check_typedef.
This is a gdb necessity in general, so could be described as an
oversight; but in this case it works around the gccrs oddity that most
named types are emitted as DW_TAG_typedef. I've filed a gccrs bug
report for that.
Diffstat (limited to 'gdb/testsuite/gdb.python/py-watchpoint.py')
0 files changed, 0 insertions, 0 deletions