Bobbie, Looking better, a few more comments below ...
Paul Bobbie wrote: > New webreve posted at > http://cr.opensolaris.org/~blong/logilab-astng/ > > Paul Cunningham wrote on 05/07/09 02:46: >> Bobbie wrote: >>> >>> I am working on porting the logilab-astng python libraries. These >>> libraries are used by pylint (python lint). If I could get some >>> review and feeback, I'd appreciate it. >>> >>> http://cr.opensolaris.org/~blong/logilab-astng/ >> >> === Start of comments === >> >> 1. usr/src/Targetdirs >> & usr/src/lib/Makefile >> & usr/src/pkgdefs/Makefile >> What is the stuff that webrev shows as deleted ? >> Does it need resyncing with the gate. > > Sorry, I'd updated the webrev based upon another feedback and forgot to > collapse the deltas. These still don't look right to me - its needs to be done against the original state of your ws bringover I think >> 2. usr/src/lib/logilab-astng/METADATA This needs to change now so it conforms to Norm's heads-up as in .. "http://wikis.sun.com/display/SFWNotes/METADATA"; >> 3. usr/src/lib/logilab-astng/Makefile.sfw >> In rule ... >> 9 $(VER)/setup.py >> do you need to 'touch' the 'setup.py' file after >> extracting the tarball > > I added this, but am Makefile illiterate. Can you explain why it's > needed? everything seemed to work the same without it. I actually meant after the line .. 40 gzip -dc $(TARBALL) | tar xmpf - rather than where you put it, so ... 40 gzip -dc $(TARBALL) | tar xmpf - touch $(VER)/setup.py but on second thought I'm not sure your makefile actually needs it. The idea was to ensure that the date/time on the setup.py file was later then other stuff. >> Line ... >> 29 REALPROTO=$(ROOT)/usr/lib/python2.4/vendor-packages >> a number of integrations are now using '/usr/lib/python2.4' so >> you might want to define it in Makefile.master and use that, >> for example something like ... >> In Makefile.master define >> PYTHONLIB=/usr/lib/python2.4 >> in Makefile.sfw change line above to .. >> REALPROTO=$(ROOT)/$(PYTHONLIB)/vendor-packages > > Back to my being Makefile illiterate, I'm reluctant to change the master > due to any potential impact to the other packages. We're already going > to have some conflict resolution during the merge, so I was hoping to > diminish this. Or is this simply ignorance on my part, and it would > improve things for everyone using the 2.4/vendor-packages? Its not essential, but it would save it having to be redefined everytime a new integration want to use that path. So if you are not happy doing it its okay with me. Line ... 48 ($(SRC)/tools/protofix --perm --pkg SUNWlogilab-astng) the pkg name here could be extracted from the METADATA, there are examples in the sfw gate. >> 5. usr/src/pkgdefs/SUNWlogilab-astng/prototype_com Question: should *.py and *.pyc be executable ? -- ---------------------------------------------------------------------- Paul Cunningham Software Engineer Tadpole Business Unit
