On Friday 25 July 2008 18:14:25 Patrick Bernaud wrote: > Peter TB Brett writes: > > [...] > > Also, a couple of questions: > > > > 06/17 Modify o_attrib_toggle_visibility() to work on a single object. > > 07/17 Modify o_attrib_toggle_show_name_value() to work on a single > > object. > > > > What's the reasoning for this? The patch seems to add a lot of code. > > Is the idea to clean up the API? (I don't object to these changes, just > > wanted to check the rationale). > > Well these two functions are in a file name o_attrib.c, and have names > that have nothing to do with selection. The main purpose is to limit > the spread of selection stuff and get closer to object methods (modulo > their prototype and lack of true objects).
Yes.
> As a result it mostly moves the selection code from o_attrib.c to
> i_callbacks.c (where it is much more appropriate): where is the lot of
> code you see?
Every invocation of those two functions has been replaced by an iteration over
a GList.
> > 14/17 Replace TILE_LOC structure with list of references to TILEs.
> >
> > What, if any, effect does this have on performance?
>
> Obviously it reduces memory allocation: no more TILE_LOC, only
> reference to TILEs. Plus TILE_LOC added nothing that was not already
> in a TILE, plus it eases the access to the TILE (where the interesting
> things are)...
>
Sure. My question was more about whether the TILE_LOC structure was originally
there to increase speed (i.e. reduce the number of pointer dereferences by
one). Ales?
In any case, these aren't really objections, just requests for clarification.
Committed! And once again, please keep up the good work!
Peter :)
--
Peter Brett
Electronic Systems Engineer
Integral Informatics Ltd
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ geda-dev mailing list [email protected] http://www.seul.org/cgi-bin/mailman/listinfo/geda-dev
