Fetching latest commit…
Cannot retrieve the latest commit at this time.
|Failed to load latest commit information.|
As I learned Qt, I finally understood how they want you to do things, so this is prettty much a complete rewrite from previous versions in the thread of development. Everything is pretty much done with signals now. In addition, if you're writing a conventional Qt application and just want twisted as an addon, you can get that by calling reactor.runReturn() instead of run(). Run blocks, but blocks artifically. Stoppping and rerunning the reactor seems to work fine... Qt is the event engine that drives it all regardless. There is a way to run trial using a gui... in bin, there is a routine gtrial. Put that in the same directory as trial and it pops up a trivial gui... hit the buton and it all runs the same... don't use the --reactor option when calling gtrial... but all the other options appear to work. This was just to make sure there wasn't anything strange with guis which there doesn't appear to be using twisted's trial also works... make sure the qtreactor directory is in your PYHONPATH and just use the --reactor=qt4 option Note that if a QApplication or QCoreApplication instance isn't constructed prior to calling reactor run, an internally owned QCoreApplication is created and destroyed. This won't work if you call runReturn instead of run unless you take responsibility for wacking QCoreApplication yourself... However, most users want this reactor to do gui stuff so this shouldn't be an issue. Included in this directory is a diff of trial's stdout between the default reactor and the Qt4 reactor. The diff is for the full test # trial --reactor=qt4 twisted 2>&1 | tee /tmp/qt4trial.trial # trial twisted 2>&1 | tee /tmp/selecttrial.trial # diff /tmp/select.trial /tmp/qt4.trial > qtvSelect.diff There are significant differences in time for the full suite of tests. Not entirely sure why yet... the profiler seems unhappy with the mail suite and getting avatar info... but the numbers don't really add up. If anyone has any insights... -glenn -- Glenn H. Tarbox, PhD firstname.lastname@example.org