[06:56] dobey: ah, sorry about forgetting that dereferencing [10:23] Hi all [10:23] Is anyone still working on bzr-git? jelmer maybe? [10:24] I'm having to interact with a growing number of git repositories, and the pain of plain git is becoming a bit too much. === yofel_ is now known as yofel [12:54] Lo-lan-do: it works fine. Pushing back to Git is lossy, though [12:56] Yeah, I know about push/dpush, that's one of my concerns. [13:48] Lo-lan-do: I don't work on bzr-git anymore [13:49] jelmer: Sad. Does anyone else? [13:49] Lo-lan-do: not that I'm aware of, unfortunately [13:50] 'kay. Thanks for the information and the past work :-) [15:38] * fullermd waves at vila. [15:39] fullermd: o/ [15:39] fullermd: I noticed your bug about log but haven't dug it yet [15:40] hello, I've used bzr tag mytag-4.9 for a given revision [15:40] but now I would like to change it to a later revision [15:40] It's a lot of digging :| [15:40] fullermd: I had network issues and just got "something" back [15:40] it turns out i had to address a bug [15:40] fullermd: yeah, it's nightmarish [15:40] I added some extra comments yesterday trying to sum it up better. The main desc got very long and twisted while I was trying to isolate and characterize it. [15:40] fullermd: I'm not even sure there is a good fix [15:41] lduros: You probably want something like 'tag --force' (or possibly two-stepping tag --delete, then re-tag'ing, but it's the same thing in the end) [15:41] fullermd: yup, saw that too, will have look next week (when I'm back home with real network and real hardware to work with :-/) [15:41] vila: I'm sorry, I don't like that answer. Plz try again ;p [15:42] fullermd: the semantics for what log should consider to be a directory change are unclear [15:42] Well, possibly. There can be some discussion about what's "right", but what's happening now is certainly _wrong_. [15:43] The part where "taking out revs that don't touch anything change what shows up for `log dir`" I find particularly egregious... [15:43] there have been discussions in the past about whether they should include any change *below* the dir instead of just the changes to the dir itself [15:43] They do, and that was already changed some versions back. [15:43] morevover -v or not trigger different code paths :-/ [15:43] Taking out 3 of the "bzr ci --unchanged -m 'Nada'" commits makes all 4 expected revs show up. [15:43] fullermd: were you able to find a bzr version where the behavior is different ? [15:44] yeah, that one is scary [15:44] Nyet, I tried the heads of all the 2.x branches and .dev, and tried 2.0 with --pack-0.92 [15:44] It's something weird about that history shape that freaks things out. [15:44] ok, at least it's not a regression then [15:44] fullermd: thanks! [15:45] (I didn't try pre-2.0, but I'm a programmer, not an archaeologist :p) [15:45] fullermd: --force worked [15:45] lduros: Note that if you've push'd the tag anywhere, you'll probably have to resort to more big hammers to update it in those places. [15:45] And if anybody else has pull'd it, things get even more complex. [15:46] fullermd: hehe, yeah, fair enough, thanks for trying all 2.x though [15:46] fullermd: but now when I try: bzr push :parent it tells me there's a conflicting tg [15:46] tag [15:46] fullermd: yeh, you are right [15:46] Lo-lan-do: sure :-) I'm still hacking on Dulwich, and also still happy to provide help if anybody else is interested in hacking on bzr-git. [15:46] fullermd: just read your line :-) [15:47] lduros: I think push --force will update it. May be "safer" to tag --delete it from there directly, then try pushing (I can't remember if tag can do that remotely) [15:48] fullermd: it says no such option as --force [15:48] But that still leaves you with the troubles of other people who've already pulled it. Nothing you can do there but notify them to tag --delete it before they pull. [15:48] Oh, --overwrite I mean. [15:48] Of couse, they cna _pull_ just fine. THey'll just get a warning about the difference, and their local tag won't move. [15:48] fullermd: yay! Looks like it works! [15:48] fullermd: nobody else is using [15:48] :-) [15:49] Don't be so pessimistic 8-} [15:49] hehe [15:51] vila: (also, note the notes about the branch of the script where it makes a foo file rather than a foo/ dir, and how it does the same thing with -v. So we can't just blame the whole thing on dirs :| ) [15:52] I'm curious, what made Bazaar a GNU project? :-) Is it because of the fact that it is a fork of GNU Arch? [15:52] fullermd: ack, then somehow it's better as it means it could be easier to fix [15:52] It's not really a fork, it just inherited a few ideas and a lot of people who worked on arch before. [15:52] ah ok [15:52] :-) [15:53] I'm not sure what the background behind GNUification was, I just woke up one day and there was some antlered thing running around. [15:53] hehe [15:54] fullermd: well the reason I started using bzr is because it became gnu and thus became the vc tool of choice for gnu projects, so I guess it helped gain a few users! :-) [15:55] Ooh, maybe that explains why there are people who use emacs, too. I can't think of any _other_ reason why anybody would... [15:55] :-P [15:56] fullermd: I wasn't saying i'm using bzr just because of that, but it got me acquainted with it :-)