Open Quuxplusone opened 8 years ago
Bugzilla Link | PR27852 |
Status | NEW |
Importance | P normal |
Reported by | Elena Lepilkina (Elena.Lepilkina@synopsys.com) |
Reported on | 2016-05-24 08:01:34 -0700 |
Last modified on | 2019-08-06 06:16:25 -0700 |
Version | trunk |
Hardware | PC Linux |
CC | Elena.Lepilkina@synopsys.com, llvm-bugs@lists.llvm.org, paul_robinson@playstation.sony.com, thomas.preudhomme@celest.fr |
Fixed by commit(s) | |
Attachments | |
Blocks | |
Blocked by | |
See also |
Given what LABEL is for, combining it with DAG doesn't make sense.
Obviously the test needs to be fixed (go for it!) so I guess the real
question is whether FileCheck should be taught to complain about this.
FileCheck was made to complain about combining -NOT with other things,
because it turned out that was a moderately common mistake. I'm less
convinced that this one is worthwhile.
(In reply to comment #1)
> Given what LABEL is for, combining it with DAG doesn't make sense.
> Obviously the test needs to be fixed (go for it!) so I guess the real
> question is whether FileCheck should be taught to complain about this.
>
> FileCheck was made to complain about combining -NOT with other things,
> because it turned out that was a moderately common mistake. I'm less
> convinced that this one is worthwhile.
Sorry, I wrote this before I saw the llvm-dev thread titled
"RFC: FileCheck Enhancements" and now I see what your intent is
for CHECK-LABEL-DAG. Following up on the email thread.
This bug is just about fixing the test.