Re: Solving 1.1.0 'svn status' perf regression the other way

> On Sat, 14 Aug 2004, Greg Hudson wrote:
>
> > On Sat, 2004-08-14 at 18:54, Erik Huelsmann wrote:
> > > Create a stream for stdout/stderr.
> > > Create a stream which recodes UTF-8 to native, which allocates its own
> xlate
> > > handle.
> > > Forward all output from the recoding stream into the stdio stream.
> > > Recode all UTF-8 to native with the allocated handle (which is stored
> in the
> > > stream baton).
> >
> > I imagine "svn status" also has to do recoding in order to perform
> > filesystem operations, which don't involve stdout/stderr. So we might
> > still be allocating just as many xlate handles even if we take care of
> > stdout/stderr using a stream approach.
> >
> Yes. When I did svn st on my svn trunk WC, there were thousands of
> character translations. The ones needed for console I/O must be a small
> fraction of these. I think you could just use our current APIs for
> translation. We need to improve the caching in some way, so don't
> complicate the stream code.

Well, there is no reason to complicate stream code, since streams can be
stacked on top of each other each performing its own specific action. The
solution IMO is to implement a new stream which does the recoding.

BTW, I think I need to implement this stream anyway to fix issue #1997.

bye,

Erik.

-- 
NEU: WLAN-Router für 0,- EUR* - auch für DSL-Wechsler!
GMX DSL = supergünstig & kabellos http://www.gmx.net/de/go/dsl
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org