Zero out padding in custom Default trait implementations #2051
Conversation
|
|
Previously, we were using `std::mem::zeroed()` which unfortunately does not necessarily zero out padding. It'd be better if the padding is zeroed out because some libraries are sensitive to non-zero'd out bytes, especially when forward/backward compatability is involved. This commit ensures all bytes are zeroed out in custom Default trait implementations.
|
Waiting for the CI to run again so I can just copy paste the diffs from other clang versions |
|
I think I fixed up the fixtures. Another CI run should confirm |
|
Bump for CI |
|
Getting close! Bumping again |
|
I really hate that github policy of having to explicitly approve CI runs. I did plan to fix this up and merge it whenever I got a chance, but if you finish it up that's even better of course, thanks! :) |
|
(I should also figure out a way to make runs not stop at the first failure...) |
|
If you want to merge as-is and fixup later, that'd work for me too. |
|
No, I meant pulling your changes, fix up as needed myself, and push with the fixes. I may be able to do it tomorrow if you don't get it green by then, otherwise later this week. |
|
Woohoo! Got all of them |
|
Sweet, thanks! |
Previously, we were using
std::mem::zeroed()which unfortunately doesnot necessarily zero out padding. It'd be better if the padding is
zeroed out because some libraries are sensitive to non-zero'd out bytes,
especially when forward/backward compatability is involved.
This commit ensures all bytes are zeroed out in custom Default trait
implementations.
This closes #2050 .
Note that I kept these commits separate to help with review. They should
probably be squashed before merge so tests are still bisectable.