On Fri, Aug 20, 2004 at 09:06:32AM +0200, Lars Gullik Bjønnes wrote: > >> >> | So am I right in assuming that the whole make machinery is responsible > >> >> | for the additional 100s? > [...] > | question. It's basically the same we would get from a "flat" > | make within mathed, isn't it? > > Yes, it should be.
So now the question is whether we can take advantage of this. g++ is seemingly able to compile the whole of lyx three times faster than what we see right now Andre' PS: I've been pretty stupid during the meeting. We could have saved quite some time by ssh'ing to soem faster machines just around the corner. Sorry for that. PPS: For those interested in the next 'break everything' patch: So far I have no further drawing regressions (there's still this funny 'boxes have wrong width' thing) and navigation and selection working on outer text. Navigation and selection within nested insets is off wrt to y coordinate, which makes thing look pretty funny. On the pro side: damn fast loading, fast drawing, got rid of all top_y stuff, LyXText::yo_, and simpler RowPainter. And one of the navigation bugs in mathinsets that formerly involved scrolling is magically gone.
