Re: libpq SSL with non-blocking sockets

From: Steve Singer <ssinger_pg(at)sympatico(dot)ca>
To: Martin Pihlak <martin(dot)pihlak(at)gmail(dot)com>
Cc: Robert Haas <robertmhaas(at)gmail(dot)com>, PG Hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: libpq SSL with non-blocking sockets
Date: 2011-06-30 13:25:14
Message-ID: BLU0-SMTP93E0029461ABF3EFF6DDF98E580@phx.gbl
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On 11-06-28 02:14 PM, Martin Pihlak wrote:
>
> Hmm, I thought I thought about that. There was a check in the original
> patch: "if (conn->sslRetryBytes || (conn->outCount - remaining)> 0)"
> So if the SSL retry buffer was emptied it would return 1 if there was
> something left in the regular output buffer. Or did I miss something
> there?
>

The issue I saw in the original patch was that at that point in the
function, sslRetryBytes could be zero (if the data was sent) but
conn->outCount - remaining would be an incorrect value since "remaining"
is the number of bytes left to send from sslRetryBuf NOT
conn->outBuffer. This is no longer an issue in the updated patch.
I will try to take a closer look at your updated patch in the next few days.

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Alvaro Herrera 2011-06-30 14:40:33 Re: Bug in SQL/MED?
Previous Message Robert Haas 2011-06-30 12:47:11 Re: how to call the function--pqCatenateResultError()