Re: help wanted reproducing reported performance regression in groff 1.24

"G. Branden Robinson" <[email protected]> Tue, 17 Mar 2026 06:41:50 -0500

Newsgroups	gmane.comp.printing.groff.general
Message-ID	<20260317114150.qvpzlky4tg3t3zwk@illithid>

Hi Morten,

At 2026-03-17T10:16:16+0100, Morten Bo Johansen wrote:
> On 2026-03-17 G. Branden Robinson wrote:
> > At 2026-03-16T19:19:04-0500, G. Branden Robinson wrote:
> > An Arch Linux user has reported a performance regression in man page
> > rendering in groff 1.24.  For "small" man pages (or collections
> > thereof), it's not noticeable, but the reported degradation is
> > quadratic in large inputs, roughly twice as bad as one would expect
> > for 25 copies in a row of the gcc man page.
> > 
> > I'm not able to reproduce the problem.
> 
> I get these numbers:
> 
>   1  0.07  0.08
>   2  0.14  0.17
>   4  0.27  0.41
>   8  0.54  1.06
>  16  1.10  5.03
>  32  2.17  20.80
>  64  4.52  79.39

That's pretty hellacious.

> % uname -a
> Linux mathilde 6.18.18-1-lts #1 SMP PREEMPT_DYNAMIC Fri, 13 Mar 2026 20:22:24 +0000 x86_64 GNU/Linux
> 
> I am also using Arch Linux.

What libc and libstdc++/libc++ does your system use?

> I seemed to nail the performance regression down to this commit:
> 
>   10 months ago 732b07d G. Branden Robinson         ∙ [man]: Fix Savannah #65190 (1/2)

That's the same commit identified in Savannah #68145.  But I think it
may be a case of a change _exposing_ a latent bug (or, in this case,
unperformant selection of an alogorithm or data structure) rather than
the change _causing_ the problem.

(That distinction may ring a bell for readers of this list.  :) )

> Hope this is useful.

Can you re-run your test scenario with the `-Z` option added to groff's
command line?  I'm curious to know how cutting grotty out of the
execution pipeline affects the numbers.

Regards,
Branden

signature.asc (application/pgp-signature, 833 B)