Mike Kupfer <mike.kupfer at sun.com> writes: > Looks okay to me, though I do have 3 questions/comments. > > - how is pylint run? (I can't get to logilab's web site to look at > the documentation.) If pylint needs pointers to sources in other > directories to be effective[1], it would be useful to add a "lint" or > "check" target to the relevant makefile(s).
pylint <list o' source files> (from a directory such that the loadpath is valid) I've been refraining from adding make targets, because, much like the pkg folk, I'm still fighting (on and off) to get its output down to a uniformly sensible level of complaint. I'm mostly losing. > - squishdeltas() could do with a comment saying that the repo write lock > should be held before calling. Good catch (the working copy lock should be held too). The locking in 0.9.4 is recursive, so in theory I could hold them at every level and everything would work... (for reference wlock/lock aren't a read/write pair, but 'lock' is the repo, wlock the working copy (so, lock == "can't commit", wlock == "can't update") > - cdm_arcs(), cdm_bugs(): why do we convert the tuple returned by > filter() into a set? Removes duplicates, that could probably be more explicit though. > Rich> (Do Sun US have next week off?) > > We have Thursday and Friday off as holidays, but some people take > additional days as vacation days. -- Rich
