Just checked in to experiments/machine. I''m not happy with the Timeout tests yet, but I wanted to get everything in. They seem to be firing correctly, but the last event from the periodic timer is lost if it overlaps with the timer that stops the reactor. I''m not sure exactly what should happen or what is happening so I''ll have to look into it a bit more. No need to put the added by Francis notes. Life will be easier if we are comfortable with editing each others code... Besides, we can always use ''svn praise'' to see that stuff. How did it go adding the IO stuff? Did the event model and everything fit the job? How do we run the unit test? Testing network code is always a pain, but maybe we can rig up some multi-process testing or something so we can really flex the bottom end. I just started using Mauricio Fernandez rcov utility, and it is really helpful. Give you a text code coverage report as well as a set of html files with highlighted coverage info. Once we move into error reporting, documenting, testing mode for the next API I''ll try to work it in. Ciao, Jeff Copied from the svn changelog: - Made Base a module (was a class) and converted where needed. - Changed to use a regular ruby array rather than the priority queue. - Got rid of the EventManager class that was just used for testing. - Added the Base#verbose method so that each unit test can specify either quiet or verbose. - Added additional @running check to reactor run loop so we don''t block if an event has stopped the reactor. - Renamed timer methods for adding handlers so they don''t overlap with the EventDispatcher interface.
On 5/24/06, Jeff Rose <rosejn at gmail.com> wrote:> > Just checked in to experiments/machine. I''m not happy with the Timeout > tests yet, but I wanted to get everything in. They seem to be firing > correctly, but the last event from the periodic timer is lost if it > overlaps with the timer that stops the reactor. I''m not sure exactly > what should happen or what is happening so I''ll have to look into it a > bit more. > > No need to put the added by Francis notes. Life will be easier if we > are comfortable with editing each others code... Besides, we can always > use ''svn praise'' to see that stuff. > > How did it go adding the IO stuff? Did the event model and everything > fit the job? How do we run the unit test? Testing network code is > always a pain, but maybe we can rig up some multi-process testing or > something so we can really flex the bottom end. > > I just started using Mauricio Fernandez rcov utility, and it is really > helpful. Give you a text code coverage report as well as a set of html > files with highlighted coverage info. Once we move into error > reporting, documenting, testing mode for the next API I''ll try to work > it in. > > Ciao, > Jeff > > Copied from the svn changelog: > > - Made Base a module (was a class) and converted where needed. > - Changed to use a regular ruby array rather than the priority queue. > - Got rid of the EventManager class that was just used for testing. > - Added the Base#verbose method so that each unit test can specify > either quiet > or verbose. > - Added additional @running check to reactor run loop so we don''t block > if an > event has stopped the reactor. > - Renamed timer methods for adding handlers so they don''t overlap with > the EventDispatcher interface. > > > _______________________________________________ > Eventmachine-talk mailing list > Eventmachine-talk at rubyforge.org > http://rubyforge.org/mailman/listinfo/eventmachine-talk >-------------- next part -------------- An HTML attachment was scrubbed... URL: http://rubyforge.org/pipermail/eventmachine-talk/attachments/20060524/7a3af3fa/attachment.htm
Just tried timeout_test.rb. I see three timeout events. Was I supposed to see only two? On 5/24/06, Jeff Rose <rosejn at gmail.com> wrote:> > Just checked in to experiments/machine. I''m not happy with the Timeout > tests yet, but I wanted to get everything in. They seem to be firing > correctly, but the last event from the periodic timer is lost if it > overlaps with the timer that stops the reactor. I''m not sure exactly > what should happen or what is happening so I''ll have to look into it a > bit more. > > No need to put the added by Francis notes. Life will be easier if we > are comfortable with editing each others code... Besides, we can always > use ''svn praise'' to see that stuff. > > How did it go adding the IO stuff? Did the event model and everything > fit the job? How do we run the unit test? Testing network code is > always a pain, but maybe we can rig up some multi-process testing or > something so we can really flex the bottom end. > > I just started using Mauricio Fernandez rcov utility, and it is really > helpful. Give you a text code coverage report as well as a set of html > files with highlighted coverage info. Once we move into error > reporting, documenting, testing mode for the next API I''ll try to work > it in. > > Ciao, > Jeff > > Copied from the svn changelog: > > - Made Base a module (was a class) and converted where needed. > - Changed to use a regular ruby array rather than the priority queue. > - Got rid of the EventManager class that was just used for testing. > - Added the Base#verbose method so that each unit test can specify > either quiet > or verbose. > - Added additional @running check to reactor run loop so we don''t block > if an > event has stopped the reactor. > - Renamed timer methods for adding handlers so they don''t overlap with > the EventDispatcher interface. > > > _______________________________________________ > Eventmachine-talk mailing list > Eventmachine-talk at rubyforge.org > http://rubyforge.org/mailman/listinfo/eventmachine-talk >-------------- next part -------------- An HTML attachment was scrubbed... URL: http://rubyforge.org/pipermail/eventmachine-talk/attachments/20060524/7e95f97d/attachment.htm