On Jan 19, 2008 11:44 PM, Eric Faurot <[EMAIL PROTECTED]> wrote: > I would suggest the attached diff instead. > > First, use VERSION and put it in SUBST_VARS so PLIST won't change > on each update because of this (rotten) egg stuff.
right! thanks for the change :) > > Second, I think the failing tests (due to missing modules mysql and pgdb) > should not be hidden, even if they are not really failures. It makes the issue > explicit even if non-critical, and it avoids hiding a real bug later. if you keep all db tests, regress simply fail because it can't test with postgres and mysql which is what I disabled in regress. I don't know if it's very usefull to keep tests that we know they will fail. More over in this case postgres, sqlite and mysql tests share the same code/object so it is fair to test only against sqlite amha, isn't it ? > > Thoughts? ok to commit? > > Eric. > change is ok for me. -- - benoît
