On Mon, Aug 03, 2009 at 04:53:44PM -0500, Shawn Walker wrote: > [email protected] wrote: >> On Mon, Aug 03, 2009 at 04:20:24PM -0500, Shawn Walker wrote: >>> [email protected] wrote: >>>> On Mon, Aug 03, 2009 at 02:13:56PM -0700, Danek Duvall wrote: >>>>> On Mon, Aug 03, 2009 at 01:50:26PM -0700, [email protected] wrote: >>>>> >>>>>> Folks, >>>>>> I have another small code review. This fixes a couple of transport >>>>>> issues, a traceback in the CLI, and provides more information when we >>>>>> fail assertions in the progress tracker. The total change here is less >>>>>> than 100 lines. >>>>>> >>>>>> http://cr.opensolaris.org/~johansen/webrev-10411/ >>>>> progress.py: >>>>> >>>>> - Is the base ProgressTracker class the right place to put messaging? I >>>>> would expect this to go in FancyUNIXPT and CommandLinePT, but not the >>>>> others. This raises the question of whether FUPT should inherit from >>>>> CLPT. >>>> Given that the traceback from failing the assertion is going to go to >>>> stderr, I assumed that it wouldn't be a problem to have the actual >>>> numbers get printed there too. Am I missing something? >>> The GUI. >> >> That's not a helpful answer. Every time we get bugs reported against >> the GUI that might actually be in the api, we tell people to re run the >> command and either capture the stacktrace that the GUI dumps to a >> terminal window, or reproduce the problem in the CLI. Has the GUI >> stopped printing tracebacks to stderr? > > Yes, it often presents them in a dialog to the user after the operation > fails.
Are these just the error messages, or does it include the traceback too? Are they using the sys traceback stuff, or are they redirecting stderr into a buffer for later use? Danek proposed we raise an exception here instead, and I've followed up with some additional comments. What do you think of that approach? -j _______________________________________________ pkg-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
