On Jun 30, 12:22 am, William Stein <wst...@gmail.com> wrote:
> On Tue, Jun 30, 2009 at 3:11 AM, Dan Drake<dr...@kaist.edu> wrote:
> > On Mon, 29 Jun 2009 at 05:21PM -0700, John H Palmieri wrote:
> >> How about #6429: sagedoc: make search_src and friends less OS
> >> dependent? This should fix the doctest failure for sagedoc.py on ppc
> >> OS X 10.4 (I hope).
>
> > I'd also like to see some testing of that patch on slower computers. The
> > patch uses os.walk and splitlines to search every line in the Sage
> > library...on my work computer it's plenty fast, but it's possible that
> > on some computers it's quite slow. But I have no idea how slow that
> > might be.
>
> It would be very nice if it built a reverse-lookup database, so that after the
> first search it is really fast. When Alex Clemesha wrote a full text
> search system for Sage a long time ago he did that (I think via
> sqlite), and it was nice.
> The drawback is that when the Sage library changes, one has to know to
> not use the cache (or to properly update it).
>
>
>
> > I don't think being slow should hold up the patch, since "works, is
> > portable, but slow" should always beat "fast but broken/unreliable",
> > right? :)
>
> It depends on how slow it is. If it is a hundred times slower, I wouldn't be
> so sure.
On my Intel iMac:
Before the patch:
sage: time s = search_src('matrix', interact=False)
CPU times: user 1.53 s, sys: 2.46 s, total: 4.00 s
Wall time: 4.51 s
After the patch:
sage: time s = search_src('matrix', interact=False)
CPU times: user 2.15 s, sys: 0.16 s, total: 2.31 s
Wall time: 2.31 s
It would be nice to see timings on other systems.
John
--~--~---------~--~----~------------~-------~--~----~
To post to this group, send email to sage-devel@googlegroups.com
To unsubscribe from this group, send email to
sage-devel-unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URLs: http://www.sagemath.org
-~----------~----~----~----~------~----~------~--~---
