On Jun 15, 2006, at 12:44 PM, Martin Vermeer wrote:
It is much better than the previous patch: using the same test, I can
type without any lag with 1 inset nested inside another (i.e., 2
nested insets) and only minimal lag with 3 nested insets. 4 or more
nested insets gives me considerable lag. (How realistic are these
scenarios? I frequently go to 3 nested insets (a note inside a
footnote inside a branch, say), and only rarely to 4.)
This is unlike how things stand without the patch: there, levels of
nesting did not affect typing speed. An everything is faster, even
with just one inset. (Typing in a single inset (unnested) on
unpatched lyx, I can get processor usage up to about 50%; with the
patched version of lyx, I can get it up to 80%.)
This is surprising. The new patch doesn't change anything that could
affect typing speed. It's just cosmetics, nothing else. Are you sure
about thet?
Well, I am sure about the slowdown , but it looks like I made two
mistakes that bear on how to understand the cause of the slowdown.
1. The version of lyx I was comparing against turns out to be 1.4.x,
not 1.5.x, as I thought. Does this imply that 1.5.x is slower than
1.4.x here? ... I'm guessing not: I have Jean-Marc's font caching
patch applied to 1.4.x (which noticeably improves things), but not to
1.5.x; perhaps that accounts for the difference.
2. It looks like the slowdown I was seeing with levels of nesting was
a side effect of having more text on the screen: generally, the more
text, the slower, and the patch (or lyx-1.5.x/no font caching)
amplifies the slowdown.
Bennett
