Jean-Marc Lasgouttes wrote:
"Helge" == Helge Hafting <[EMAIL PROTECTED]> writes:
Helge> Did that. There seems to be no display errors. Editing
Helge> documents, and resizing the userguide is fine.
Helge> The big file consisting of small paragraphs is still ok to
Helge> edit.
Is this with or without Martin's patch?
Without. I tend to test only _one_ patch at a time, unless told otherwise.
That's my way of trying to not mess up my tree too much.
I misunderstood, I thought Martins patch got merged.
Helge> The big file with a single gargantuan paragraph seems worse, I
Helge> out-typed it (with normal sentences, not garbage) and waited 10
Helge> seconds before it appeared. Even typing in the few initial
Helge> small paragraphs were awfully slow. It wasn't this bad last
Helge> time I tested.
This is surprising.
Testing with the patch reverted shows that pure yesterdays cvs
is just as bad. So, André's patch definitely didn't make anything
worse for me.
Testing with both Martins and Andrés patches at the same time,
with the "case6/case 8" removed from Andrés patch:
Speed is nice in the big document with normal paragraphs. I couldn't
even create cpu load here.
The document with a 125k paragraph is fine as long as I
don't work in the monster paragraph. Working in the
initial small paragraph is fine. I can still out-type lyx
in the unrealistic monster paragraph.
Moving around in the userguide (and resizing it) is nice. No
display artifacts of any kind. Horizontal resizing is quick
enough that opaque resizing works fine.
I would be surprised that this patch alone makes a big difference. I
guess we should stick with martin's for now (1.4.0). However the
approach in Andre's patch can only improve speed, but should be taken
further.
I'd say, apply them both. Martins patch make a big difference, as
editing one paragraph no longer is impeded by the size of other paragraphs.
Andrés patch makes less of an impact for me, but a linear speedup
with no ill effects is a very good thing. The effect may be much more
noticeable for slower machines, where the doubled speed might make
the difference between sluggish and ok. Taking it furter could
be a new patch, I see no problem applying the existing.
Helge Hafting
