You've already put back a fix for bug 4483. This needs to be a new bug. > Is this what we want here, or should we start from a fixed list?
It might be useful to have empty categories, just to show that things will eventually be placed there. If you want to support them, then I'd add an option to the script that takes the empty categories as arguments. > 3/ All packages in "opensolaris.org" that should have been of > classification "System/X11 (System)" were previously written > as "System/X11". I thought we got rid of the duplicate "System" on purpose. > http://cr.opensolaris.org/~richb/pkg-4483-v2/ gen_os_files.py: - line 86: "dictory" - line 92: I assume this is chopping off "classification"? Please split on whitespace first, then on "/": line.split(None, 1)[1].split("/") for instance. Maybe you need an rstrip() in there, too, if line isn't stripped already coming in to the method. - line 93: What happens if there is no slash? - line 98: You should explain why you're making these adjustments, as on the face of it, they're pretty silly. - line 114: spurious close-paren? - line 132: "empty category" -> "empty subcategory"? - line 135: I might just return here, rather than having the else clause. - line 142, 145: Don't get in the habit of using non-specific except clauses. You probably want to catch ValueError specifically. - line 223: This might be better off as a single triple-quoted string, like in client.py. - line 314: "subcategoryr"? - line 328: This strikes me as being a bit dangerous -- it's going to pick up a lot of files that aren't import files, and it's going to pick up a lot of import files that aren't relevant to any given build. Danek _______________________________________________ pkg-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
