Exactly what it says on the tin! The comments in the code detail this a little more too.
Details
- Reviewers
NoQ xazax.hun rnkovacs dcoughlin Charusso baloghadamsoftware - Commits
- rG967583bc087c: [analyzer] Note last writes to a condition only in a nested stackframe
rC368817: [analyzer] Note last writes to a condition only in a nested stackframe
rL368817: [analyzer] Note last writes to a condition only in a nested stackframe
Diff Detail
- Repository
- rL LLVM
Event Timeline
Yup, looks good!
clang/include/clang/StaticAnalyzer/Core/BugReporter/BugReporterVisitors.h | ||
---|---|---|
146–149 ↗ | (On Diff #208275) | I definitely don't don't don't don't dislike quadruple negations. |
I have one small question otherwise looks good.
clang/lib/StaticAnalyzer/Core/BugReporterVisitors.cpp | ||
---|---|---|
1457 ↗ | (On Diff #210340) | I wonder if "nested" is a good term in the title of this revision. We could also show notes in the caller right? So it might also be an "enclosing" frame. And if so, is it desired? Since the "enclosing" frames should already be visible for the user without the additional notes. But correct me if I'm wrong :) |
clang/lib/StaticAnalyzer/Core/BugReporterVisitors.cpp | ||
---|---|---|
1447–1454 ↗ | (On Diff #210340) | Yeah, interesting, i guess we should also not track the value further. Otherwise it'd be weird that one piece in the middle is missing but everything else is still there. |
clang/lib/StaticAnalyzer/Core/BugReporterVisitors.cpp | ||
---|---|---|
1447–1454 ↗ | (On Diff #210340) | That is what D64271 was for, and I think it would be a bad idea. What we should rather do is have a better note message: |
clang/test/Analysis/track-control-dependency-conditions.cpp | ||
263 ↗ | (On Diff #210340) | Like here. We'd lose this note if we didn't track further. I'm already gathering data tho! |