Simon Riggs <simon@2ndquadrant.com> writes:
>> So thinking we should test XLogArchiveCheckDone() for both
>> stopxlogfilename and history file and then stat for the stop WAL file:
> This seems better.
Somehow I missed the 5-Apr patch that introduced this bogosity.
Next time you make a fundamental change in the behavior of a function,
how about updating its comment to match?
After studying it for a minute, I think that XLogArchiveCheckDone no
longer even has a clearly explainable purpose; it needs to be split
into two functions with two different behaviors. I plan to revert
XLogArchiveCheckDone to its previous behavior and introduce
XLogArchiveIsBusy (flipping the sense of the result) to use in
pg_stop_backup.
regards, tom lane