On Mon, Mar 2, 2009 at 6:36 AM, Thomas Moschny
<address@hidden> wrote:
Derek Scherger wrote:
> I've committed a fix on net.venge.monotone for this problem in
> b1cec3176fd56af29275c2b620f8766b4382eec8 which fixes two associated
> xfailed tests in the monotone testsuite (tests/attr_mtn_execute and
> tests/defecting_attriibutes).
Great! Had something similar, but not yet working, in my workspace for a
while. I factored out the code to set/get the x bit in f954b59c88.. .
Thanks, I had looked at that and then forgot to do it.
So in short, I think we should pass NIL to the lua hook when an
attribute is cleared, instead of "", because the latter could be a valid
value.
Yes, this would probably be better. I'm not sure how we would pass that through though? The current cpp hook api takes a std::string which can't really represent nil. Maybe we need two cpp api's (set/clear) that call a single lua hook?
As far as I know the revision (i.e. changeset) format doesn't have the
vocabulary to talk about "dormant" attributes, and imho it shouldn't
need to.
Good point, these are only known to the rosters. So maybe this is all fine the.
Thanks for the review, and if you get a chance to fix the set/clear api's go for it... I may not get to it for a day or two.