> Last night I reran my appliation regression tests on a win32 database
> with your win32_open fix and that solved the 7 failures I was seeing.
> So with that change *all* of my basic application regression
> tests work when running on the win32 port. (I just changed the one call in
fd.c
> BasicOpenFile to use win32_open).
Great! That's very helpful validation.
> Other issues I have seen in playing around are:
>
> - Errors reported during shutdown related to statistics collector.
Seen this too and noted.
> - Problems with Ctrl-C shutdown not being recognized immediately
> (sometimes if I ctrl-c the postmaster it properly shutsdown,
> othertimes it seems to hang until after another (or sometimes two other)
> connections to the postmaster are attempted (i.e. after I ctrl-c the
> postmaster I need to connect one or twice from psql before it will
> actually respond to the ctrl-c)
Hmm. I'd hazard a guess and say that, given that the ctrl-c responds after a
new connection attempt, this is almost certainly due to the pgstat
APC/socket interaction bug, which I've "patched" in my sources [I'll make a
mental note to remove my patch and try to reproduce.]
> - Sometimes I have seen jdbc errors connecting with the error
> message:
> no pg_hba.conf entry for ??? , but I have only seen this twice and
> don't know yet how to reproduce.
Sorry, no idea on this. We have to reload the hba settings in each backend
under win32, but this happens well before the backend will try to
authenticate. If you can figure out a way to reproduce, please let us know.
Thanks for the reply,
Claudio
---
Certain disclaimers and policies apply to all email sent from Memetrics.
For the full text of these disclaimers and policies see
<a
href="http://www.memetrics.com/emailpolicy.html">http://www.memetrics.com/em
ailpolicy.html</a>