When locating a bug in the repository history, I don't always have a failing test script available (or can't easily write one). For example, if the regression is a visual glitch in some particular web browser. In these scenarios, it's often easier for the user to pretend to be the test script himself.
I propose that 'trackdown --interactive' pause for each patch and ask the user "Does this version work correctly? [yn]" If the user answers 'y', it's treated the same as the test passing. If he answers 'n', it's treated as a test failure. Darcs would also have to indicate the temporary directory where the code resides, so the user knows where to test. There would also be a 'q' command to stop the trackdown early. Does this sound useful to others? If so, I'll create an issue for it. Would I be correct to assign this a ProbablyEasy tag? -- Michael _______________________________________________ darcs-users mailing list [email protected] http://lists.osuosl.org/mailman/listinfo/darcs-users
