-
Notifications
You must be signed in to change notification settings - Fork 13.6k
Open
Labels
A-debuginfoArea: Debugging information in compiled programs (DWARF, PDB, etc.)Area: Debugging information in compiled programs (DWARF, PDB, etc.)C-discussionCategory: Discussion or questions that doesn't represent real issues.Category: Discussion or questions that doesn't represent real issues.T-compilerRelevant to the compiler team, which will review and decide on the PR/issue.Relevant to the compiler team, which will review and decide on the PR/issue.WG-debuggingWorking group: Bad Rust debugging experiencesWorking group: Bad Rust debugging experiences
Description
In #129403 I'm seeing that using array-based SIMD types changes the debug info you get from them, and thus #128866 has in fact already changed the debuginfo you get from platform types like __m256
.
// gdb-command:print vi8x16
-// gdb-check:$1 = simd::i8x16 (0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15)
+// gdb-check:$1 = simd::i8x16 ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15])
Should debug-info notice the Vector
ABI and emit something that doesn't get double-braced? Or does it not matter and the ([1, 2])
is completely fine so it's not worth changing?
workingjubilee
Metadata
Metadata
Assignees
Labels
A-debuginfoArea: Debugging information in compiled programs (DWARF, PDB, etc.)Area: Debugging information in compiled programs (DWARF, PDB, etc.)C-discussionCategory: Discussion or questions that doesn't represent real issues.Category: Discussion or questions that doesn't represent real issues.T-compilerRelevant to the compiler team, which will review and decide on the PR/issue.Relevant to the compiler team, which will review and decide on the PR/issue.WG-debuggingWorking group: Bad Rust debugging experiencesWorking group: Bad Rust debugging experiences