GNU bug report logs - #78499
tramp-accept-process-output busy-waits

Previous Next

To reply to this bug, email your comments to 78499 AT debbugs.gnu.org.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Daniel Colascione <dancol <at> dancol.org>
To: bug-gnu-emacs <at> gnu.org
Subject: tramp-accept-process-output busy-waits
Date: Mon, 19 May 2025 17:41:00 -0400

tramp-accept-process-output should wait for something to actually happen
to the process from which we're trying to accept output, not just return
immediately after specifying a zero timeout and rely on the caller
to poll.

Message #8 received at 78499 <at> debbugs.gnu.org (full text, mbox):

From: Michael Albinus <michael.albinus <at> gmx.de>
To: Daniel Colascione <dancol <at> dancol.org>
Cc: 78499 <at> debbugs.gnu.org
Subject: Re: bug#78499: tramp-accept-process-output busy-waits
Date: Wed, 21 May 2025 09:47:14 +0200

Daniel Colascione <dancol <at> dancol.org> writes:

Hi Daniel,

> tramp-accept-process-output should wait for something to actually happen
> to the process from which we're trying to accept output, not just return
> immediately after specifying a zero timeout and rely on the caller
> to poll.

tramp-accept-process-output has seen several implementation iterations
ove the years. I'm happy, that it looks stable now.

Could you please show a problem with the current implementation, which
needs to be fixed?

Best regards, Michael.

Message #11 received at 78499 <at> debbugs.gnu.org (full text, mbox):

From: Daniel Colascione <dancol <at> dancol.org>
To: Michael Albinus <michael.albinus <at> gmx.de>
Cc: 78499 <at> debbugs.gnu.org
Subject: Re: bug#78499: tramp-accept-process-output busy-waits
Date: Thu, 22 May 2025 03:43:24 -0400

On May 21, 2025 3:47:14 AM EDT, Michael Albinus <michael.albinus <at> gmx.de> wrote:
>Daniel Colascione <dancol <at> dancol.org> writes:
>
>Hi Daniel,
>
>> tramp-accept-process-output should wait for something to actually happen
>> to the process from which we're trying to accept output, not just return
>> immediately after specifying a zero timeout and rely on the caller
>> to poll.
>
>tramp-accept-process-output has seen several implementation iterations
>ove the years. I'm happy, that it looks stable now.
>
>Could you please show a problem with the current implementation, which
>needs to be fixed?
.
100% CPU use while we're waiting to receive bytes from the peer is a priori a problem that needs to be fixed. Good software doesn't busy wait. We have a perfectly good mechanism for waiting for something to happen with timeouts and we shouldn't avoid them on account of years of accumulated changes or compatibility with ancient and irrelevant bugs. 

Message #14 received at 78499 <at> debbugs.gnu.org (full text, mbox):

From: Michael Albinus <michael.albinus <at> gmx.de>
To: Daniel Colascione <dancol <at> dancol.org>
Cc: 78499 <at> debbugs.gnu.org
Subject: Re: bug#78499: tramp-accept-process-output busy-waits
Date: Thu, 22 May 2025 13:22:24 +0200

Daniel Colascione <dancol <at> dancol.org> writes:

Hi Daniel,

>>tramp-accept-process-output has seen several implementation iterations
>>ove the years. I'm happy, that it looks stable now.
>>
>>Could you please show a problem with the current implementation, which
>>needs to be fixed?
> .
> 100% CPU use while we're waiting to receive bytes from the peer is a
> priori a problem that needs to be fixed. Good software doesn't busy
> wait. We have a perfectly good mechanism for waiting for something to
> happen with timeouts and we shouldn't avoid them on account of years
> of accumulated changes or compatibility with ancient and irrelevant
> bugs.

"A priori" isn't a bug recipe. Pls show me a recipe indicating a
misbehavior, which needs to be fixed.

Best regards, Michael.

Previous Next

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