- User Since
- Jun 25 2014, 4:17 PM (160 w, 4 d)
Addressed Erik's comments
You might also want to mention the fact that -fsanitizer=vptr requires null in the release notes.
Fri, Jul 21
I'll commit this on behalf of Johannes today as he didn't get his access yet
Wed, Jul 19
This triggered a warning in LLVM itself, in CoverageMapping.h :
Tue, Jul 18
rm early exit bug
Oops, I just realised that now there's a small bug with early exits. Since we don't actually have true lexical order for declarations in the @implementation we might exit early after visiting a method in the @implementation before a function that's actually written before that method. I will probably constraint early exits to avoid this case.
Factor out the lexical ordering code into a new visitor and simplify the implementation of the ast selection visitor
Address review comments.
- Address review comments.
- Remove the Location parameter and ContainsSelectionPoint enum value.
- Stop traversing early when a declaration that ends after the selection range was reached.
Mon, Jul 17
Fri, Jul 14
Yep, LGTM. Thanks.
Nice, I will need something like this for the refactoring stress test tool in the future that verifies that the indexer and renaming engine have a similar view of the code.
Thanks for doing this! I have some comments:
Thu, Jul 13
Wed, Jul 12
I was impatient, so I already started on a patch for diagtool. I'll post it soon.
Right. I was aware of the diagtool before, but didn't really look into what it did. TIL! It would make sense to add this kind of mapping functionality to that tool then.
My script relies on a hack to map the name of the diagnostic to the enum value. We need to come up with a better to map the diagnostic name to the enum value. I propose a new utility tool that would take the name of the diagnostic and map it back to the enum value.
Committed r307771 with correct attribution.
Sure, I'll commit it today.
Tue, Jul 11
LGTM. Do you have commit access or would you like me to commit it on your behalf?
Currently looks like around 200k (4534 @ 33 byte avg length + ptr). If this is too much, we could make it conditional based on NDEBUG or some other macro at compile time.
Mon, Jul 10
Thanks for the patch.
I have just one minor comment:
Thanks, that's pretty cool!
Fri, Jul 7
Can you provide a test that demonstrates what this change fixed/improved?
Thu, Jul 6
Wed, Jul 5
Does this apply to all constexpr global variables? It could potentially fix https://bugs.llvm.org/show_bug.cgi?id=31860 .
A a test-case for implicit declarations.
Tue, Jul 4
LGTM. You can request commit access at http://llvm.org/docs/DeveloperPolicy.html#obtaining-commit-access.
Mon, Jul 3
Are you planning to work on integration with the StmtDataCollector in this patch or would you prefer to follow-up with additional patches?
This is meant to be NFC btw.