
scott at kitterman
Dec 18, 2006, 5:58 PM
Post #14 of 20
(3132 views)
Permalink
|
On Mon, 18 Dec 2006 19:47:54 -0500 (EST) "Stuart D. Gathman" <stuart [at] bmsi> wrote: >On Mon, 18 Dec 2006, Scott Kitterman wrote: > >> Since your preferred approach requires a change in the README for command >> line use, I'd appreciate it if you'd make those changes. > >So you'd like me to mention spfquery in README even though option 3 >is currently in CVS head? > >> I still prefer option 1, but prefer 2 over 3 eventually. 3 is a good way >> to go for the moment. > >I think the choice between 2 and 3 has to be permanent. We don't want >to rename the driver package down the road. Understand that now. The following comments are based on not doing option 1 (still my preference). >This is what I think I'm hearing from you: > > 1) change CVS to option 2 (make a branch tag?) Yes. With thr branch tag. > 2) update README to talk about spfquery For 2.1 and follow. > 3) 2.0.x will stay with single spf.py module Yes. > 4) 2.1.x will use option 2 (spf package) Yes. I'll (when I have time) make a Debian package for 2.1 and will maintain the 2.0 branch (I'll backport changes from 2.1 that look interesting/necessary). I'd suggest you set up 2.1 with spfquery so that an existing call to spy.py can be changed to spfquery.py with no syntax changes (that makes changing command line calls easy). Supporting the other spfquery interfaces would be nice if feasible, but lower priority. > >Does that match your current thinking? Yes (more detail than I'd thought through, but yes). >BTW, spfquery should have an option to run doctests for spf package, >since spf package won't do so anymore. I think this is mandatory. >Comment: IMO, spfquery, as currently consituted (to be compatible with >other spf packages), is a Royal Pain to use compared with spf.py. OK. I haven't looked at it. See my coment above. Scott K ------- To unsubscribe, change your address, or temporarily deactivate your subscription, please go to http://v2.listbox.com/member/?list_id=1007
|