2013/4/15 Dagobert Michelsen <[email protected]>: > I would like to keep the check, there have been occasions where this was > extremely useful > and helped point out issues which were gone by unnoticed. Taking out the > known-to-be-good > pathes is IMHO the best way.
OK, so we're keeping the check and adding a whitelist. The change needs to go as deep as metadata collection, because we're grepping for predetermined strings. We also need to figure out how to represent this, e.g. what's the outcome if we have two occurrences, one good and one bad; and what if the bad one is a substring of the good one (!). I'm not saying this is hard, just that it has to be programmed correctly (unit tests are strongly encouraged). One change will go into package_stats.py around line 29[1] where we define what we're grepping for. Except now not all regexes will be markers of 'bad' any more, so we might want to rename the variable. Then in package_checks.py line 773[2] there will be a modification which will apply some logic, where it'll look at what regexes matched and make a decision whether to throw an error or not. That's where tests are essential, because we need to make sure that we're covering both common and corner cases. I'm guessing it's a simple enough change, so I'll let you guys knock yourselves out with Python coding, but do send your patch out for review before you commit it. We make sure that the code is readable for other humans. Maciej [1] https://sourceforge.net/apps/trac/gar/browser/csw/mgar/gar/v2/lib/python/package_stats.py#L29 [2] https://sourceforge.net/apps/trac/gar/browser/csw/mgar/gar/v2/lib/python/package_checks.py#L773 _______________________________________________ maintainers mailing list [email protected] https://lists.opencsw.org/mailman/listinfo/maintainers .:: This mailing list's archive is public. ::.
