I hit upon a nice way to speed up idle-summary mode in semantic.
The parsing of hoards of files has been a bit of a drag for me in my
big code base at work. The idea is that summary-mode will disable the
'unloaded option in semanticdb-find-default-throttle while it does
As such, idle-summary will be very fast. If you then use one of the
many other semantic commands that access the full database, then those
unloaded tables will then get loaded.
It was a bit tricky as the simple concept conflicted with all the
caching mechanisms I added late last year.. I got it worked out
though and all the inter-file dependencies now track correctly. Yay!
Those who use this feature alot and use CVS/CEDET. Please give it a
try and let me know how it goes. Thanks!
In other news, I checked in the second half of my SRecode plan,
which is a template file map that keeps itself up-to-date. This means
that user templates (in ~/.srecode) will be automatically tracked and
loaded as needed. That makes it easy to override particular templates
from the core templates, thus allowing customizations of higher-level
templates. Yay again!
If that sounds confusing, well... Sorry. At the moment SRecode is
yet another template writing / insertion system that happens to be a
bit complex. For me it's more than that, as the core of a generalized
code generation system. Hints of that system appear in a couple of
the simple srecode applications such as "srecode-getset.el".
The srecode.texi file should now be up to date with most of the
latest new features.
Eric Ludlam: eric@...
Siege: http://www.siege-engine.com Emacs: http://cedet.sourceforge.net