[ http://issues.apache.org/jira/browse/GUMP-20?page=all ] Leo Simons closed GUMP-20: --------------------------
Resolution: Fixed Fix Version: Gump3-alpha-3 (was: Gump2) This miscount won't occur in Gump3 due to the different way we "model the model". > Dependecies/Dependees are over counted > -------------------------------------- > > Key: GUMP-20 > URL: http://issues.apache.org/jira/browse/GUMP-20 > Project: Gump > Type: Bug > Components: Python-based Gump > Versions: Gump2 > Reporter: Adam Jack > Fix For: Gump3-alpha-3 > > This list is clearly wrong: > http://lsd.student.utwente.nl/gump/scarab/scarab_details.html#Full+Project+Dependencies > The primary reason is that this count of unique dependency objects (an entity > which links to projects) not the results of the dependency itself -- the > dependent project. Dependency objects are owned (they form a bi-directional > link) and the owner is part of the __eq__ (equality check) which I think is > probably valid (albeit annoying here.) > As such when we attempt to count only unique dependencies and does 'if not > depend in depends:' -- the __eq__ says no. The problem becomes worsened when > on considered jars ids. Are two depends on the same project equal if they > reference different ids? I'd say no. > I suspect the best thing to do is to keep this up at the project level, find > a way to compare two dependencies only by the project they reference. I > wonder if Gumpy needs a way to optimize/merge dependencies, so repeats are > reduced. That said, there are multiple areas (not jsut counters/lists) that > need to have the right answer, and doing this wrong could dork classpaths. > Honestly, I don't know the right way to fix this, hence this JIRA posting to > let the thought sit openly for a while... -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://issues.apache.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]