Some of this discussion was split out into CB:35463, let's try to move it back here. Basically, it seems there's agreement that we shouldn't have different APIs for the same thing but we also want something that's not endian-specific, which clrsetbits_le32() currently is.

So to move forward, would anyone object if I:

1. remove this file again (doesn't seem like any code is using this yet),
2. add a new clrsetbits32() (and variants) API for this instead, and
3. change all existing instances of clrsetbits_le32() that don't clearly need the endian-specificness to clrsetbits32()?

To view, visit change 36596. To unsubscribe, or for help writing mail filters, visit settings.