Re: Re: [COMMITTERS] pgsql: On Windows, when a file is deleted and another process still has

Heikki Linnakangas <heikki(dot)linnakangas(at)enterprisedb(dot)com> writes:
> Here's a patch implementing that, and changing pgrename() to check for
> ERROR_SHARING_VIOLATION and ERROR_LOCK_VIOLATION like pgwin32_open()
> does, instead of ERROR_ACCESS_DENIED.

This looks sane in a quick once-over, though I haven't tested it.
One tiny stylistic suggestion:

if (err != ERROR_SHARING_VIOLATION &&
err != ERROR_LOCK_VIOLATION)
#else
if (errno != EACCES)
#endif
return -1;
if (++loops > 300) /* time out after 30 sec */
return -1;

This is overly cute and will probably confuse both pgindent and ordinary
editors. It's worth one extra line to keep each part of the #if
syntactically independent, ie

if (err != ERROR_SHARING_VIOLATION &&
err != ERROR_LOCK_VIOLATION)
return -1;
#else
if (errno != EACCES)
return -1;
#endif
if (++loops > 300) /* time out after 30 sec */
return -1;

> I wonder if we should reduce the timeout in pgrename(). It's 30 s at the
> moment, but apparently it hasn't been working correctly, failing
> immediately instead if the file is locked.

I have a vague recollection that there was a specific reason for having
such a long timeout --- you might want to check the archives to see the
discussion before that code got committed. However, if nothing turns
up, I wouldn't object to reducing it to 5 or 10 sec.

regards, tom lane