Fix ICE (issue 8262). #4025
Fix ICE (issue 8262). #4025
Conversation
| @@ -0,0 +1,16 @@ | |||
| // TBD: define expected output here. | |||
There was a problem hiding this comment.
The trouble with putting each issue in its own file is the test suite then takes forever to run. Please add it into one of the existing files, then the incremental testing time is pretty much zero.
There was a problem hiding this comment.
Where's the best place to put this particular test?
There was a problem hiding this comment.
Usually they go in xtest46 if they don't have and special compiler flag requirements. And the test should be reduced to not use phobos.
There was a problem hiding this comment.
What does the 46 stand for?
There was a problem hiding this comment.
Haha who knows. It's the megatest.
There was a problem hiding this comment.
Shouldn't it be called test42 then?
There was a problem hiding this comment.
Shouldn't it be called test42 then?
Actually come to think of it there is a test42 that is also ginormous.
There was a problem hiding this comment.
Hmm. There are a bunch of testxxx's that appear to be named after their respective bug numbers. So I looked up issue 42 and 46 on bugzilla... seems to be some random issues long since fixed.
There was a problem hiding this comment.
Those test files most likely predate bugzilla.
|
I suspect that if the alias resolves to a tuple, then the tuple must be one element, and it is resolved to that one element, otherwise an error. |
|
Righto, I'll take a look in that direction. Thanks for the feedback! |
quickfur
force-pushed
the
issue8262
branch
2 times, most recently
from
0d50ae8
to
9a4e884
Compare
|
Don't have time to work on this right now... plus, this will mostly be superceded by the multiple |
FIxes: https://issues.dlang.org/show_bug.cgi?id=8262
This PR is still incomplete, as it still doesn't quite implement the
alias thisin the expected manner (the output of the test program in the bug isS(0)as opposed to0, which would be the case if the extra level of indirection viaSeq!swas removed), and the error appears to be swallowed. At least the compiler won't segfault anymore, but probably this isn't the correct fix for the problem.