Additionally, fix build.
Once you finish making changes to this review, switch it back to ready for review.
- Queries
- All Stories
- Search
- Advanced Search
- Transactions
- Transaction Logs
Advanced Search
Yesterday
Verify with Static Analyser developers.
Switching status of review, once you will be ready with changes (or your decision), just mark it ready for review again.
Wed, Mar 22
Tue, Mar 21
Looks fine, and looks like it's working.
Mon, Mar 20
Overall LGTM, give it few days, maybe something will pop up.
I will try to run this change on my project and check if it finds any false-positives or real-issues.
I have feeling that hidden checks that none other check depend on should be enabled by default.
Also I have feeling that those hidden checks shouldn't be visible to end user.
But need to confirm this first with Static Analyser developers.
Sun, Mar 19
Review comment fixes
Update documentation, changed default check configuration.
Update documentation, and config examples, add "Traditional Tokens Representation" section.
Sat, Mar 18
There are 2 issues found in llvm repository:
- llvm/lib/Target/AMDGPU/AMDGPULibFunc.cpp:721:22: warning: reference variable 'S' extends the lifetime of a just-constructed temporary object 'const StringRef', consider changing reference to value [readability-reference-to-constructed-temporary]
- clang/lib/Sema/SemaChecking.cpp:10096:36: warning: reference variable 'AT2' extends the lifetime of a just-constructed temporary object 'const analyze_printf::ArgType', consider changing reference to value [readability-reference-to-constructed-temporary]
Change name of isStatic* to isImmutable*
Removed some whitespace & added tests
Requesting changes due to lack of support for base class initializes & got some concerns related lambdas used in initializers.
Fri, Mar 17
I will check this more deeply during weekend.
Thu, Mar 16
Regarding release-notes:
I would probably wrote something like "fixed handling for designated initializers", nonene will understand things like "sequence point".
If you decide to change this, the just commit this as an [NFC] without review.
And checks in this section should be in alphabetical order (so this one shouldn't be at the end), but that's a different story that can be corrected later for all those entrys.
Tue, Mar 14
Looks good to me, consider adding info to Release notes, that this check has been improved/fixed support for designated initializers.
Mon, Mar 13
Thank you for information, I will look into this....
Indeed test for this scenario is missing.
LGTM
Sun, Mar 12
Release notes:
./clang-tools-extra/docs/ReleaseNotes.rst
My only complain is documentation, changes in code and tests are correct.
Build failed mainly due to conflicts in previous change.
It should get green once prev patch will land, and this will be re-based.
You will need to rebase this, there are conflicts...
"The dummy implementations of std::format and std::print require C++20,
so it's easier to test them in a separate
redundant-string-cstr-format.cpp check file. This also requires a small
improvement to the basic_string_view implementation in the <string>
dummy header.
Correct commit message and this could land.
Rebase + Fix order of checks in ReleaseNotes
Update documentation
Review fixes
Rebase on top of main + Change llvm::Optional into std::optional
@ccotter Do we commit this, or you plan to still change something here ?
In D145312#4187195, @carlosgalvezp wrote:@PiotrZSL I believe you have landed some patches of this chain. Would you mind sharing how you do it? I'm not an expert in Phabricator and simply doing arc patch D145312 leads to cherry-pick conflicts, and when I solve them I don't end up having this patch as HEAD. With this knowledge I will be able to help out closing these patches!
This need to be rebase due to conflict to land.
- Failed tests
- Missing tests for this change
- Missing configuration option for this change (changing default behavior)
- Missing release notes
- Missing documentation update
- auto& , auto&&, auto* still wont be ignored (that's also auto type).
Sat, Mar 11
"Do you think that the change can land like in its current state first? Or would you prefer that the configuration option is added at the same time?"
Review comments fix