Jeremy Katz wrote:
On Wed, 2007-05-16 at 20:53 -0400, James Bowes wrote:
On Wed, 16 May 2007 16:11:37 -0400, Jeremy Katz wrote:
On Wed, 2007-05-16 at 16:07 -0400, Jeremy Katz wrote:
On Wed, 2007-05-16 at 15:58 -0400, Jeremy Katz wrote:
In some (light at this point) testing, everything seems to work but if I
didn't break anything, I'll be shocked. Thoughts, comments, other
ideas?
I hate this bug.
+1 ;-)
+1, there the lazy initialization of self.repos is very smart, but in
this situation it make things very hard.
We _could_ make it so that ayum can be None and that doSetup can take an
ayum arg -- then, it wouldn't have to break API as old API users would
be guaranteed to be calling YumBase.doRepoSetup() which can do the right
thing; new callers would pass in the YumBase object to the constructor
and then it would be fine accessing later. But aiyee.
Doing a google search, the only outside caller of the RepoStorage
constructor is in yumex and Tim already made a change there... I think
if we did the second option above, yumex could go back to the code it
was using for yum 3.0 and not work.
It look ok, to me, i can't think of any better way to do it, maybe this
can be made better if we pull some callbacks into yumbase, as on
the todo list for post 3.2.
Tim
_______________________________________________
Yum-devel mailing list
[email protected]
https://lists.dulug.duke.edu/mailman/listinfo/yum-devel
