Wasmtime GitHub notifications bot (Apr 27 2025 at 20:25): Wasmtime GitHub notifications bot (Apr 27 2025 at 20:25):SingleAccretion opened PR #10681 from SingleAccretion:DI-Sim to bytecodealliance:main:
We need to record the correct ranges for the CU and include an offset zero entry in the line table, otherwise LLDB won't class our functions as "having debug info".
Fixes #10676.
Wasmtime GitHub notifications bot (Apr 27 2025 at 20:32):SingleAccretion updated PR #10681.
Wasmtime GitHub notifications bot (Apr 27 2025 at 20:47):SingleAccretion submitted PR review.
Wasmtime GitHub notifications bot (Apr 27 2025 at 20:47):SingleAccretion created PR review comment:
debuggers expect this
At least, LLDB expects this. When LLDB stops at a certain address (say due to a step-in), its "has debug info" check is implemented as:
[current address].debug_line_row.IsValid()So even if we have the right DWARF DIE, etc, if we don't have a sequence in the line table that matches the current address, we get behaviors associated with functions that don't have DI.
Wasmtime GitHub notifications bot (Apr 27 2025 at 20:47):SingleAccretion edited PR review comment.
Wasmtime GitHub notifications bot (Apr 27 2025 at 20:49):SingleAccretion has marked PR #10681 as ready for review.
Wasmtime GitHub notifications bot (Apr 27 2025 at 20:49):SingleAccretion requested pchickey for a review on PR #10681.
Wasmtime GitHub notifications bot (Apr 27 2025 at 20:49):SingleAccretion requested wasmtime-core-reviewers for a review on PR #10681.
Wasmtime GitHub notifications bot (Apr 28 2025 at 14:32):alexcrichton submitted PR review.
Wasmtime GitHub notifications bot (Apr 28 2025 at 14:55):alexcrichton merged PR #10681.
Last updated: Dec 06 2025 at 06:05 UTC