El , Victor Olaya <[email protected]> escribió:
Sounds great to me.
I have given you write access to the SEXTANTE SVN (as I told you, this
might change to another location if it finally gets integrated into
QGIS, but for now it will remain there, at least until the end of
GSoC). Feel free to add your changes there so we can all start testing
them. It affects just a single element of SEXTANTE (though a key
one!), so there is no risk of breaking anything else. Also, SEXTANTE
is still an experimental and unstable plugin, so I see no reason why
you should keep your code in a different place.
Uhm. I am still a bit afraid of breaking stuff, I must admit. I'll post patches attached to blog posts till this is more mature, ie doesn't randomly crash taking QGIS down. I hope we can switch to git soon, and keep a separate branch for mt-related code. Or maybe I should figure out how distributed svn works (never used it that way).
Regarding the progress elements, they are very simple now and in fact
only the main parameters dialog acts as a progress display. We can
maybe leave them like that (passing the progress object to the
algorithm) in a first step, and then implement some sort of signaling
mechanism
Yes, I that is the approach I have taken for now. Eventually there could be a mechanism for the algorithms to fire the signals themselves. It might also be a good thing if they react to cancel signals. Currently, canceled algorithms just keep running and their output is ignored.
BTW, QtConcurrent was mentioned in the other thread as nice, functional, high level api for threading. It has the advantage of keeping a thread pool around and reusing them, and, I believe, the QGIS core uses it. Only problem is, there seem to be no bindings for python, because of some sort of incompatibility with the interpreter I don't understand.
Regards, Camilo
_______________________________________________ Qgis-developer mailing list [email protected] http://lists.osgeo.org/mailman/listinfo/qgis-developer
