Re: (Font-lock) slowness and syntax properties

>>>>> "Jan" == Jan Vroonhof <jan(a)xemacs.org&gt; writes: Jan> update_syntax_cache basically Jan> - Computes the Qsyntax_table property using Fget_char_property Jan> - Estimates the boundaries of the region where this table Jan> is valid using Fprevious_extent_change and Jan> Fnext_extent_change (note that this is a vast underestimate, Jan> most of the time this will be (point-min) and (point-max)). In fact, most of the time in everything except cperl mode Qsyntax_table should be nil (or maybe the default). So one dirty trick is to simply arrange to dirty a flag in any buffer where the syntax-table property is ever set. That would be a lot faster than using Fget_char_property. Jan> 3. IIRC correctly syntax property extents cannot overlap, Jan> therefore the boundaries of the extent where we found the Jan> current property value are good values for the cache Jan> boundaries and much more likely to be bigger (possibly even Jan> the whole buffer). This is probably true. However, even if not there's a way to just walk the extents that overlap this one. Note that if this one is big, then we do a big computation once, and avoid further computation for a long time. When the profit is small, the computation is also small. Also, you may want the dirty flag anyway (to invalidate these cache boundaries). -- Institute of Policy and Planning Sciences http://turnbull.sk.tsukuba.ac.jp University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Don't ask how you can "do" free software business; ask what your business can "do for" free software.