Eric Blake <ebl...@redhat.com> writes: > On 03/03/2015 12:15 AM, zhanghailiang wrote: > >>>> >>>> Yes, this is the reason ..., agreed, i don't like the abbreviate, >>>> But there is already a 'MigrationState' type defined: >>>> >>> >>>> >>>> So, what about MigrationStatus ? ;) >>> >>> That would be fine with me. >>> >> >> Bad news, this name has also been used :( >> >> In hmp.c: >> >> typedef struct MigrationStatus > > You know, you could always rename the internal-only conflict into > something else so that the publicly exported typename is nice. Yeah, > that makes the series longer,
by *two* patch hunks updating the four occurences of MigrationStatus, > but it should be all mechanical > conversions, right? I'm not going to be too picky about what color we > paint this bikeshed, though. Me neither, but we shouldn't compromise on external interfaces just to avoid a bit of internal churn. Pick a good name, then do what it takes.