Avoid parsing key usage flags from hashed key ID subpackets

[DemiMarie]

A key ID subpacket is not a key usage flags subpacket and must not be
interpreted as such. This caused RPM to wrongly reject a signature that
had both key ID and key usage flags subpackets in the hashed section.

@pmatilai I believe this fixes a regression in 4.18 alpha.

[nwalfield]

The missing break is obviously a bug.

It appears that the missing break has existed since support for the issuer subpacket was added in 2001 (3b820b2#diff-b4eac15fda646a3b73f5cd251f33387979eadc71ba52f769bd64b10bd877365cL470). That commit included a comment that control was supposed to fall through, but I think using a comment instead of breaking was misguided. Unfortunately, the comment was later removed :/.
When @DemiMarie added support for the KEY_FLAGS subpacket in 598a771, she understandably missed that the issuer subpacket handling didn't have its own break statement.

I agree that this fixes a bug, the patch should be applied, and, due to unfortunate circumstances, represents a regression in 4.18 alpha.

[nwalfield]

FWIW, I think the commit message should just be 'Add missing break statement, regression introduced in ...'

[DemiMarie]

@nwalfield is the new commit message better? @pmatilai friendly ping, and apologies for getting it wrong the first time. @nwalfield thanks for the testing that caught this before it made it into a release.

[nwalfield]

I think the commit message is better, but in the end @pmatilai needs to decide what is best.

[pmatilai]

The fact that it's a regression fix should be stated in the summary line. That's what people will look at first when looking for relevant changes, and "add missing break" says nothing about the severity of it.

[pmatilai]

Something like "Fix OpenPGP key ID parsing regression" would be more like the headline that such a thing needs. The rest of the message is fine.

[pmatilai]

Thanks for the patch.