On Fri, Jun 09, 2006 at 02:22:17PM +0200, Lele Gaifax wrote: > First of all, sorry for taking so long to answer! Secondly, the ML may
no problem! > be a better place for this kind of questions ;-) I'm subscribing to the ML right now ;) > I don't think so. I noticed those too, but no time to dig further. They > seem to be directories, that hg does not track. Most probably the hg > backend does not filter them out at commit time. correct, hg doesn't track directories. > Uhm, I don't understand this point. Do you mean that you get > "incompatible" hg repositories, branches where hg is not able to > exchange patches? yes. The changeset hashes changes quite randomly over time, it seems after some time I can't regenerate a pristine mergeable tree by using the same config file. It's not completely reproducible on demand but it happened more than once, and the trees I'm currently using have this problem (I can't regenerate them). > Isn't hg using some random bits to build up the changeset hash? May that > be the culprit? I thought not, if this was the case it would explain it obviously. But I tested it and it seems it's not generating random bits. I regenerated the tree twice, the second time after mv FormEncode FormEncode.orig; rm tailor.state. [EMAIL PROTECTED]:~/devel/python/network/x/FormEncode> hg pull ../FormEncode.orig/ pulling from ../FormEncode.orig/ searching for changes no changes found [EMAIL PROTECTED]:~/devel/python/network/x/FormEncode> _______________________________________________ Tailor mailing list [email protected] http://lists.zooko.com/mailman/listinfo/tailor
