Re: Label alignments in shorten_branches

[amylaar at spamcop dot net]

Quoting Bernd Schmidt <bernds_cb1@t-online.de>:
> Let's back up here for a minute.  If this is true, then why don't we
> track the possible padding in INSN_ADDRESSES, rather than using
> align_fuzz?  Shouldn't the two approaches be equivalent, and if not, in
> what cases is the current code beneficial?

It is beneficial when more than one alignment of the same size is spanned,
and also for combinations of alignments of different sizes if the larger
alignment doesn't subsume the smaller one(s) because of an offset between
them.