DexExpectStepOrder uses the line to expect a debugger step from the actual line of the command in the Dexter source file. Now Dexter scripts have mainly moved to thier own script files instead of the actual source, there should be a option to override this behaviour to choose your own debugger step location
Details
Diff Detail
- Repository
- rG LLVM Github Monorepo
Event Timeline
We do need to refactor and clean up the delineation between command line of declaration and target line in the code at some point. This is fine in the meantime though and the patch looks good.
LGTM (with an inline question), thank you @BenJMudd!
cross-project-tests/debuginfo-tests/dexter/dex/debugger/DebuggerControllers/ControllerHelpers.py | ||
---|---|---|
22–23 | This change seems a little bit suspicious - as mentioned offline perhaps this was doing the wrong thing before though. It does certainly look like no command classes ever set _on_line, so it looks like this is doing the right thing with your patch. Have you checked the llvm-lit tests in cross-project-tests/debuginfo-tests/dexter/feature_tests pass (preferably on Linux as it has greater coverage)? |
cross-project-tests/debuginfo-tests/dexter/dex/debugger/DebuggerControllers/ControllerHelpers.py | ||
---|---|---|
22–23 | Yep, just tested those tests on Linux both before and after removing this _on_line check, no difference in the output, they all pass, so I think it must have been doing the wrong thing before. |
This change seems a little bit suspicious - as mentioned offline perhaps this was doing the wrong thing before though. It does certainly look like no command classes ever set _on_line, so it looks like this is doing the right thing with your patch.
Have you checked the llvm-lit tests in cross-project-tests/debuginfo-tests/dexter/feature_tests pass (preferably on Linux as it has greater coverage)?