Bug 32073 – We have two mark-fixed commands

• Status: RESOLVED
• Resolution: FIXED
• Component: Tools / Tests

[Adam Barth]

There should be only one!

It seems like we're been favoring putting more commands into bugzilla-tool, where they can share a bunch of infrastructure.  What do you think?  Should we port the extra functionality from mark-bug-fixed to bugzilla-tool mark-fixed, or should we kill the bugzilla-tool command?

[Eric Seidel (no email)]

I never intended for there to be 2.  We could easily just delete the bzt one.  I think some of the apple folks who use the mark-as-fixed command prefer it to be a stand-alone command, so we should kill the bzt one.  However, we could/should consider replacing the guts of mark-as-fixed with a python implementation which uses all our nifty infrastructure. :)

[Adam Barth]

mark-bug-fixed does use Python!  Go look at it.  :)

The problem is that it's been left behind in the march of bugzilla-tool technology.  It's using stone knives and bear skins.

[David Kilzer (:ddkilzer)]

The mark-BUG-fixed script was moved to a separate script so that it would be less...confusing...to developers who didn't like deciphering the cacophony of subcommands in bugzilla-tool --help.

However, I don't think that this experiment was a success, and since the script was not updated with the other bugzilla-tool refactoring, and because bugzilla-tool now has help for individual sub-commands, I'm slightly in favor of it being rolled back into bugzilla-tool.

[Eric Seidel (no email)]

We can make it easy to make any Command object into an individual script.

Something like this:

mark-as-fixed
#!/bin/python

from modules.commands.mark_as_fixed import MarkAsFixed

if __name__ == "__main__":
    MarkAsFixed().check_arguments_and_execute(sys.argv, None)

That won't work out of the box yet, because Commands don't deal well with tool being None, but that can easily be fixed.  We'd probably want to call some special run_as_stand_alone_command(sys.argv) call instead of check_arguments_and_execute directly anyway.

[Eric Seidel (no email)]

Heck, we could even write a script to auto-gen individual command wrapper files from bugzilla-tool sub-commands if necessary. :)

[Eric Seidel (no email)]

(In reply to comment #2)
> mark-bug-fixed does use Python!  Go look at it.  :)
> 
> The problem is that it's been left behind in the march of bugzilla-tool
> technology.  It's using stone knives and bear skins.

Wow, yeah, I *totally* forgot about bug 28910!

[Adam Barth]

Created attachment 44858 [details]
Patch

[Eric Seidel (no email)]

Comment on attachment 44858 [details]
Patch

Hmm.. OK.  Assuming it still works.   This seems like a only semi-useful stop-gap measure.

[Adam Barth]

Committed r52142: <http://trac.webkit.org/changeset/52142>

[Adam Barth]

(In reply to comment #9)
> Committed r52142: <http://trac.webkit.org/changeset/52142>

Actually committed in r52148