Initializing an array with [0, ..-1] compiles, but causes a segfault at runtime #6977

Activity

1. Aatch commented on Jun 6, 2013

   

   Aatch

   on Jun 6, 2013

   Contributor

   More actions

   Seems like a simple oversight, we should make sure that the length is >= 0. I am surprised it doesn't fail trying to emit a -1 alloca though.
2. luqmana commented on Jun 6, 2013

   

   luqmana

   on Jun 6, 2013

   Contributor

   More actions

   Hmm yea, the count part is just casted to a uint so this is doing the equivalent of [0, ..(MAX_UINT)]
3. thestinger commented on Jun 6, 2013

   

   thestinger

   on Jun 6, 2013

   Contributor

   More actions

   Ah, so this is really just crashing because it hits the bug we have on very large arrays. Perhaps this isn't really a separate issue than the integer literal overflow and large stack array ones?
4. Aatch commented on Jun 6, 2013

   

   Aatch

   on Jun 6, 2013

   Contributor

   More actions

   @thestinger either way, negative-length arrays aren't a sane thing to write and should still fail as it's own thing.
5. bors

   added a commit that references this issue on Jun 7, 2013

   auto merge of #6985 : Aatch/rust/fixed-vec-6977, r=thestinger

   ...

   This issue will close once commit 4abd83b is merged into the 'main' branch.

   4abd83b
6. Aatch

   closed this as completedon Jun 7, 2013
7. flip1995

   added a commit that references this issue on Apr 8, 2021

   Auto merge of rust-lang#6977 - flip1995:or_patterns_msrv, r=llogiq

   ...

   6f2a6fe
8. U007D

   added a commit that references this issue on Aug 21, 2026

   Merge rust-lang#6977

   ...

   ab343e4