QF5690 updated this revision to Diff 161684.
QF5690 added a comment.
Using `isObjCARCImplicitlyUnretainedType` instead of `isObjCClassType`
Tests in `property-in-class-extension-1.m` relaxed to
`-Wproperty-attribute-mismatch` (was `-Weverything`)
Repository:
rC Clang
https://reviews.llvm.org
QF5690 added a comment.
In https://reviews.llvm.org/D44539#1207982, @rjmccall wrote:
> LGTM.
Thanks! What I should do next? Haven't found any info in docs about it :)
Repository:
rC Clang
https://reviews.llvm.org/D44539
___
cfe-commits mailing
QF5690 added a comment.
> you can ask for commit privileges yourself.
Ok, how do I do it?
Repository:
rC Clang
https://reviews.llvm.org/D44539
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo
QF5690 added a comment.
In https://reviews.llvm.org/D44539#1208838, @rjmccall wrote:
> Please read the developer policy: https://llvm.org/docs/DeveloperPolicy.html
>
> The information is on that page.
> We grant commit access to contributors with a track record of submitting high
> quality p
QF5690 added a comment.
It's been a couple of months now, @rjmccall any ETA's?
Repository:
rC Clang
https://reviews.llvm.org/D44539
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
QF5690 added inline comments.
Comment at: include/clang/Basic/DiagnosticSemaKinds.td:1018
+def warn_objc_property_assign_on_object : Warning<
+ "'assign' attribute must not be of object type, use 'unsafe_unretained'
instead">,
+ InGroup, DefaultIgnore;
rjmccal
QF5690 created this revision.
QF5690 added a reviewer: rsmith.
QF5690 added a project: clang.
Herald added a subscriber: cfe-commits.
There is a problem, that `assign` attribute very often getting out of
attention. For example, consider this code:
@property(nonatomic, strong, readonly, nullabl
QF5690 added reviewers: aaron.ballman, benhamilton.
QF5690 added a comment.
Hey, saw many revisions around obj-c and sema that you are reviewing. Can you
help me with this one?
Repository:
rC Clang
https://reviews.llvm.org/D44539
___
cfe-commits
QF5690 updated this revision to Diff 150487.
QF5690 added a comment.
Remove warning for `Class` type, change warning message.
Repository:
rC Clang
https://reviews.llvm.org/D44539
Files:
include/clang/Basic/DiagnosticGroups.td
include/clang/Basic/DiagnosticSemaKinds.td
lib/Sema/SemaObjC
QF5690 added inline comments.
Comment at: lib/Sema/SemaObjCProperty.cpp:2554
+ PropertyTy->isObjCRetainableType() &&
+ !PropertyTy->isObjCClassType()) {
+Diag(Loc, diag::warn_objc_property_assign_on_object);
rjmccall wrote:
> Please use `isObjCARCIm
QF5690 added a comment.
In https://reviews.llvm.org/D44539#1041993, @benhamilton wrote:
> I wonder if this wouldn't be better as a clang-tidy check:
>
> https://github.com/llvm-mirror/clang-tools-extra/tree/master/clang-tidy/objc
Border between clang-tidy checks and compiler diagnostics is not
QF5690 added a comment.
> I think the problem is there are valid places to use assign on object types
> (especially delegates).
Isn't it better to have unsafe_unretained there? I thought unsafe_unretained
keyword is introduced specifically for that kinds of things.
Ok, here is my last point :)
12 matches
Mail list logo