https://github.com/ElementsProject/lightning/issues/4203
Seems like gcc is turning the "*r = *a" into memcpy, and various (all?) callers have r == a. Technically disallowed, but gcc may be allowed to make this assumption?
-
rustyrussell commented at 7:27 PM on November 17, 2020: contributor
-
real-or-random commented at 8:53 PM on November 17, 2020: contributor
Ah, that's the same as #776, just not in a
-Osbuild... Sad that it happens here too. -
elichai commented at 8:54 AM on November 18, 2020: contributor
Now gcc too? :(
-
real-or-random commented at 9:58 AM on November 18, 2020: contributor
Oh indeed, so far we've seen this only with clang... Let's track this in #776.
- real-or-random closed this on Nov 18, 2020
