I misspoke above; the example constructed outline has 111110
positions, 50 vnodes, and 5 tnodes. I was distracted in my
terminology because If you were to write the .leo file, you'd
currently write 111110 <v></v> pairs. That should be my next
project. It appears as if the read code already supports *reading* <v/
> nodes, but I haven't begun looking seriously at it yet.
- Stephen
On Apr 2, 10:43 am, thyrsus <[EMAIL PROTECTED]> wrote:
> You're welcome, although I suspect you're getting some placebo
> effect ;-).
>
> My "speedups" are aimed at situations where clones cause the former
> allNodes_iter to explore a potentionally huge tree that produces no
> further result. Here's instructins on how to create an outline that
> will show a *lot* of speedup:
>
> Create a noda "a". Clone it 9 times. 10 vnodes, 1 tnode.
> Create a node "b" as a child of "a". Clone it 9 times beneath "a".
> 110 vnodes, 2 tnodes.
> Create a node "c" as a child of "b". Clone it 9 times beneath "b".
> 1110 vnodes, 3 tnodes.
> Creae a node "d" as a child of "c". Clone it 9 times beneath "c".
> 11110 vnodes, 4 tnodes.
> Create a node "e" as a child of "d". Clone it 9 times beneath "d".
> 111110 vnodes, 5 tnodes.
>
> "all_positions_with_unique_tnodes_iter" will return 5 positions,
> having visited 50 vnodes, where allNodes_iter will return once for
> each 111110 vnodes.
>
> Now, no one would really create an outline like that, but I've got 100
> machines documented by about 6 machine types each which in turn are
> described by about 5 make targets each which are in turn described by
> about 4 makefiles each (different makefiles for different OSs) and
> each make target may be depend on further make targets to a depth of
> about 10 levels or so, and you now see where I am primed to take
> advantage of the above.
>
> My most recent version refactored and consolidated some of the
> iteration code, and it is remotely plausible that the smaller though
> slightly more complex code is giving you a caching advantage - but my
> motivation was "Don't Repeat Yourself", not speed.
>
> I'm delighted to have given a little something back to the project.
>
> - Stephen
>
> On Apr 2, 6:31 am, "Ville M. Vainio" <[EMAIL PROTECTED]> wrote:
>
> > On Sat, Mar 29, 2008 at 12:56 AM, thyrsus <[EMAIL PROTECTED]> wrote:
> > > More replacements of allNodes_iter a.k.a. all_nodes_iter. All unit
> > > tests pass.
>
> > Thanks for all this speedup work. The whole Leo feels much snappier
> > now, even with small documents. Of course I don't know exactly what is
> > causing the speedups ;-)
>
> > --
> > Ville M. Vainio - vivainio.googlepages.com
> > blog=360.yahoo.com/villevainio - g[mail | talk]='vivainio'
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"leo-editor" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at
http://groups.google.com/group/leo-editor?hl=en
-~----------~----~----~----~------~----~------~--~---
