GNU bug report logs - #47283
Performance regression in narinfo fetching

Previous Next

Full log

Message #16 received at 47283 <at> debbugs.gnu.org (full text, mbox):

From: Ludovic Courtès <ludo <at> gnu.org>
To: Christopher Baines <mail <at> cbaines.net>
Cc: 47283 <at> debbugs.gnu.org
Subject: Re: bug#47283: Performance regression in narinfo fetching
Date: Sun, 21 Mar 2021 22:10:40 +0100

Hi!

Christopher Baines <mail <at> cbaines.net> skribis:

> Christopher Baines <mail <at> cbaines.net> writes:
>
>> I haven't looked in to this yet, but maybe it would be possible to
>> adjust the code so that it doesn't perform so badly, but still tries to
>> handle possible exceptions.
>>
>> The two ideas I have is to rewrite the (let ...) bit in terms of a fold,
>> maybe that would perform better, or stop using let for iteration and
>> setup the exception handling, then process each request, using set! to
>> update the state. I haven't tested either of these.
>
> I tried something, neither of these things, but just not calling (loop
> ...) within the catch block. I don't know why this might work, but it
> seems to make guix weather much faster.

Oh yes, that’s also because calling ‘loop’ from within ‘catch’ made it a
non-tail call, so we kept accumulating exception handlers, and the ‘lp’
loop in ‘raise-exception’ would have an ever increasing list of handlers
to traverse.

> Here's the patch [1], I've just realised it's broken, as it'll loose the
> result value (and use an old one) when the connection is closed. I'll
> send a updated patch without this issue in a moment.
>
> 1: https://issues.guix.gnu.org/47288

OK, thanks.  I’ll reply to your other message first.  :-)

Ludo’.

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.