| Lists: | pgsql-hackers-win32 |
|---|
| From: | Claudio Natoli <claudio(dot)natoli(at)memetrics(dot)com> |
|---|---|
| To: | 'Andrew Dunstan' <andrew(at)dunslane(dot)net>, Claudio Natoli <claudio(dot)natoli(at)memetrics(dot)com> |
| Cc: | 'Merlin Moncure' <merlin(dot)moncure(at)rcsonline(dot)com>, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, Win 32 hackers PGSQL <pgsql-hackers-win32(at)postgresql(dot)org> |
| Subject: | Re: [PERFORM] scalability issues on win32 |
| Date: | 2004-12-03 03:01:08 |
| Message-ID: | E80765D3BE190745B72F81EFB8332AAB014CFB@harold.memetrics.local |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Lists: | pgsql-hackers-win32 |
> How is it that this has not bitten us elsewhere? Are we only reading
> tiny things from pipes in other cases so that a single read works?
IIRC, this particular use of ReadFile is only to read version strings.
> It seems very strange that a shorter version string should cause the
failure.
Counterintuitive, but appeared entirely reproducible on my WinXP
installation. No doubt there is a perfectly unreasonable explanation.
Interested in seeing another confirmation.
In any case, afaics ReadFile fails to make any guarantee that it will return
an entire line, so pipe_read_line requires attention.
Cheers,
Claudio
| From: | Andrew Dunstan <andrew(at)dunslane(dot)net> |
|---|---|
| To: | Claudio Natoli <claudio(dot)natoli(at)memetrics(dot)com> |
| Cc: | "'Merlin Moncure'" <merlin(dot)moncure(at)rcsonline(dot)com>, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, Win 32 hackers PGSQL <pgsql-hackers-win32(at)postgresql(dot)org> |
| Subject: | Re: [PERFORM] scalability issues on win32 |
| Date: | 2004-12-03 04:12:59 |
| Message-ID: | 41AFE7CB.404@dunslane.net |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Lists: | pgsql-hackers-win32 |
Claudio Natoli wrote:
>>How is it that this has not bitten us elsewhere? Are we only reading
>>tiny things from pipes in other cases so that a single read works?
>>
>>
>
>IIRC, this particular use of ReadFile is only to read version strings.
>
>
>
>
>>It seems very strange that a shorter version string should cause the
>>
>>
>failure.
>
>Counterintuitive, but appeared entirely reproducible on my WinXP
>installation. No doubt there is a perfectly unreasonable explanation.
>Interested in seeing another confirmation.
>
>
I tried but buildfarm stops before we get there :-(
regression failure on join - see
http://www.pgbuildfarm.org/cgi-bin/show_log.pl?nm=loris&dt=2004-12-03%2003:16:39
It's too late for me to do anything about it tonight.
>In any case, afaics ReadFile fails to make any guarantee that it will return
>an entire line, so pipe_read_line requires attention.
>
>
Quite.
cheers
andrew