On Tue, Sep 27, 2022 at 12:44:10AM +0100, Mark Brown wrote: > On Tue, Sep 27, 2022 at 12:46:21AM +0200, наб wrote: > > > If I'm reading the merge right (very much not a given!), > > it seems that the NBD_REPLY_MAGIC (and LO_MAGIC?) constant(s) survived: > > they both need to go for reasons listed in > > bd5926220ffe0: LO_MAGIC doesn't exist > > 82805818898dd: NBD_REPLY_MAGIC is part of the line protocol, > > not a magic number > > > This also reveals that I missed NBD_REQUEST_MAGIC > > (needs to go, same reason as NBD_REPLY_MAGIC) > > in the first pass, but that's unrelated here. > > I preserved NBD_REPLY_MAGIC since the conflict was due to it being > updated by the NBD maintainer, I went with their logic instead. IIRC > they'd also moved it within the file which might make the resolution > harder to read. LO_MAGIC is gone. > > It's not at all clear from what's in the file that your logic about not > including magic numbers defined elsewhere means we shouldn't include > them in the table here, though the commit messages are rather brief so > perhaps there is more to it. It's certainly a very strong definition of > need as far as I can tell. > > > (or, indeed, if it does include... the conflict markers? > > because it does appear to introduce them > > (or, at least, if I leave in the conflict markers and commit a merge, > > it sure looks like what's represented below)?), > > so idk. > > I skipped out on resolving the conflicts in the translated copies of the > file but messed up on resetting the to the base state. It seems I've overestimated the degree to which this matters in this case, my b. Put me down as "don't care". наб
Attachment:
signature.asc
Description: PGP signature
