Re: bad makeinfo slowdown

[Per Bothner]

On 11/06/2010 03:40 PM, Karl Berry wrote:
> Agreed.  We are well aware of the painful slowness (every time I compile
> texinfo.txi ...) and very much want it to be better asap.  However,
> Patrice already did some profiling and there is no obvious inner loop to
> optimize :(.  So it seems any significant improvement will probably come
> from writing parts of it in C, which will imply all kinds of system- and
> Perl-dependent complications.  Not looking forward to that.

I hate to point out the obvious: Going back to the old version
may be the right decision.  I'm sure you've considered it, and
hate the idea.  Presumably there were structural problem with the old code.
If so, one could use the Perl version as a prototype for a C version, or
perhaps a hybrid version that re-uses much of the old code, but whose logic
is like the Perl version.  Unless the slowdown is due to the structure of
the new code, rather than using Perl - in which case the job is easier - 
stay
with the old logic! Though I assume there was dome near-fatal problem
with the old logic?

A 50x slowdown is really bad, considering that the old code worked
pretty well for most of us.  You need a pretty good reason to replace
a working version with one that is 50x slower.  Are there strategic
plans for a next-generation makeinfo that needs to be able to do
something the old one couldn't be make to do?   OTOH a drastic slowdown
precludes some other valuable features such as running makeinfo on
the fly.

Sometimes you have to be willing to say "this was a failed experiment".
--
        --Per Bothner
address@hidden   http://per.bothner.com/