From: | Andrew Dunstan <andrew(at)dunslane(dot)net> |
---|---|
To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Cc: | pgsql-hackers(at)postgresql(dot)org |
Subject: | Re: Speeding up the Postgres lexer |
Date: | 2005-05-23 20:34:54 |
Message-ID: | 42923E6E.3080906@dunslane.net |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Tom Lane wrote:
>[snip - flex is slowed down by backtracking - how to fix ]
>
>What I'm wondering is whether this is really worth doing or not.
>There are currently just two parts of the lexer rules that are affected
>--- the {real} rule illustrated above, and the rules that allow quoted
>strings to be split across lines as the SQL spec requires. But the
>patches are still pretty ugly, and what's really annoying is that there
>doesn't seem to be any way to get flex to complain if someone later
>makes a change that breaks the no-backup-cases property again.
>
>
>
I would be more concerned if there were not reasonable alternatives to
many bulk parsed inserts (COPY, prepared statement).
But I do think it's worth it, even so ... not all client interfaces
support prepared statements (notoriously PHP, although I understand KL
has sent patches to fix that) and not all inserts are suitable for COPY.
cheers
andrew
From | Date | Subject | |
---|---|---|---|
Next Message | Sander Steffann | 2005-05-23 20:37:18 | Re: inet increment w/ int8 |
Previous Message | Simon Riggs | 2005-05-23 20:08:24 | Re: Speeding up the Postgres lexer |