=== vxnick is now known as vxnick-AFK [00:42] morning all [00:44] * fullermd_ waves at igc. [00:45] igc: Are you currently (in general, not this second) working on filtering stuff, or is that pushed down the stack? [00:47] Hi igc [00:47] morning #bzr [00:48] Hi jml [00:49] I'm going to release Bazaar 1.17 today [00:49] jml, you legend [00:49] i hope you had a hearty breakfast [00:49] hi fullermd: back on the content filter bugs today (after higher priority reviews) [00:49] poolie, I haven't eaten, in fact. [00:49] hi garyvdm! [00:50] i'm going to try to finish outstanding reviews and my assigned bugs [00:50] ie not spend all week on phone and email, as i did last week :/ [00:50] igc: Sweet. I don't have anything particular on it, was just curious where it fell. Looking forward to it! [00:51] It's amazing how much you can accomplish when you turn off the ringer ;) === fullermd_ is now known as fullermd [00:51] fullermd_: it's less important than making 2a stable [00:51] fullermd_: but I find it personally embarrassing and ... [00:51] Sure, but can't I have everything all at once? Immediately? And a pony? [00:52] I'd really like to ensure it's working properly by the 2.0 release [00:52] The sad part is that a key component of going from "works" to "usable" is per-branch config of it, and every time that comes up we all get eaten by a grue :| [00:53] fullermd: if you provide the patches (and parent ponies), I'll merge them immediately :-) [00:53] fullermd: so I guess that makes you the bottleneck :-) [00:54] That sounds messy... [00:59] jml, speaking of your stats on launchpad [00:59] i was thinking after La Mola that it might be interesting to do brief stats on our releases as to [00:59] what kind of bugs we fixed: plain bugs, new features, regression fixes, [01:00] if it was late, or how controlled the release process was in general and maybe why [01:00] this might not be worth it but it might be interesting to look back [01:00] poolie, well, my LP stats had the advantage of being really easy to get using bzr log + grep [01:00] nb i'm not suggesting adding any additional work to your plate for today [01:00] poolie, good good :) [01:00] yeah this would be at least partially manual [01:03] poolie, those would definitely be interesting stats to have. [01:03] although they'd all need strong not-a-metric qualifiers :) [01:03] you could get some kind of medium-term feedback on it perhaps... [01:03] s/on/from [01:04] i may try it one jetlagged morning [01:12] good idea [01:32] garyvdm: those pending merge screen snapshots are nice - well done! [01:32] Thanks [01:38] jelmer: a bzr-git branch is doing this: [01:38] SHA1KnitCorrupt: Knit corrupt: sha-1 of reconstructed text does not match expected sha-1. key ('git-v1:56599cc5b24f8a4e30423898c7cf5077c569b9ee',) expected sha d28a73fcdd473745fce7765c35fe419613e278e9 actual sha b3391cd0e3bc7a3c9a510c2055fd469ea5458821 [01:38] (i'm not actually sure which branch it is :() [01:38] jelmer: heard of this? [01:44] igc/garyvdm, is there already an option somewhere that does the equivalent of bzr commit --show-diff? [01:45] poolie: you mean on qcommit? [01:45] mm [01:45] preferably a gui option not a command line one [01:45] qcommit has a Diff button [01:45] of course [01:46] poolie: and if you're using explorer, the status view shows you what's changed [01:46] right [01:46] i suppose i meant to pop that up automatically [01:47] but that does sound pretty lazy doesn't it :) [01:47] mwhudson: hi [01:48] jelmer: hello [01:48] mwhudson: haven't seen that before [01:48] jelmer: ok, i'll try to find out where it's coming from [01:51] poolie: Sorry - I was away. Not quite sure what you are asking. [01:51] s'ok, igc answered it [01:51] ok [01:54] jml: I'm about to submit that patch for merging into 1.17 [01:54] jml: it's the same fix as Friday, but now I have a test :) [02:04] poolie: Is this ok: http://bazaarvcs.wordpress.com/?p=39&preview=true ? [02:14] mwhudson: My guess would be that we're calculating and inserting a wrong SHA1 somewhere, or used to at some point. [02:15] mwhudson: possibly because we fail to update InventoryEntry.text_sha1 in a corner case. [02:16] mwhudson: when is it giving that error exactly? during "bzr pull" from git? [02:16] jelmer: no, in trying to generate a diff for a revision mail [02:17] mwhudson: ah, interesting [02:18] mwhudson: so there would potentially be more branches with this issue, it just hasn't been noticed yet because nobody is subscribed to diff emails? [02:18] jelmer: it's possible [02:19] * mwhudson files a bug "job system oops reports don't tell you which branch is being processed" [02:20] garyvdm: very nice, thanks [02:36] mwhudson: btw, did you see my reply to your bzr-svn + python2.4 bugreport? [02:36] jelmer: yes [02:36] jelmer: i'll try to reply sensibly soon [02:36] (it's been one of those mornings) [02:43] Hi there.. I just came across Bazaar... [02:44] I'm trying to set up a repository on my windows machine.. Something like the dropbox idea.. how is bazaar at handling regular windows kind of files? [02:44] it should handle them fine [02:44] I used to use svn but the repository was too big.. The local repo would be the same size as the actual fles [02:45] if your files are binary, you'll find that most VCS systems have that property, and those that don't are very slow ;) [02:46] ugh.. how do sites like dropbox do it then? [02:47] on their server they will have a copy of your files [02:47] and if your files are binary, it will be about the same size as your files ;) [02:47] Oh.. yeah but i don't mind that on the server [02:47] oh, you mean on your machine? [02:47] What I mean.. is with SVN it doubles the local storage [02:48] it does that because it has local data [02:48] copies of the files to do diff operations with [02:48] VCS systems like to make 'diff' operations quick [02:48] Does bazaar do that as well? [02:49] we keep a copy of all your history since you started the project [02:49] same as git, hg, darcs and monotone [02:49] hmm.. not what I want then [02:49] perhaps you don't want a VCS? [02:49] I want something like dropbox [02:49] do you need history? diffing? annotate? [02:50] I want a history and versioning. Not diffing or annotate [02:50] are you on a LAN, or working over the internet? [02:50] mostly LAN, sometimes internet [02:56] I need it to work over ssh, preferably [02:56] I want a dropbox idea but on my own server === timchen1` is now known as nasloc__ [02:59] well, you could use bzr with lightweight checkouts [02:59] they don't have a local cache [02:59] but the tradeoff is that its a lot slower to do many operations [03:02] igc: I'm fixing apply-inventory-delta branch [03:02] igc: are there any issues other than test failures now? [03:02] lifeless: I'm just looking over the code now [03:04] lifeless: I will look into that [03:13] lifeless: What is the windows client like for bazaar? Does it have something like TortoiseSVN? [03:14] There is a TortoiseBZR, yes [03:15] cool [03:15] bzr uses ssh? [03:15] malibu: it can, yes. [03:15] if you want it to, yes [03:16] ok [03:17] malibu: see http://bazaar-vcs.org/BzrExplorer as well [03:25] malibu: or iFolder, apparently it has an open source server component [03:25] So say I do light checkouts.. bzr will still try to do things efficiently right? For example, will it send a file if the last modification times haven't changed? [03:26] it does a sha1 check before uploading anything [03:26] Because as long as it's efficient as, say, rsync.. I'm fine with that [03:26] come to think of it, does it only send deltas of binary files? [03:27] that depends [03:27] :) [03:29] ok.. I'm just going to try it. [03:30] repo on my ubuntu machine, tortoisebzr on my windows laptop [03:31] With light checkouts [03:31] (rock on) [03:35] spiv, re the inventory-delta branch, you're going to put back IDS? [03:35] hm actually apparently you said you were going to do without it in the review... === _SamB_ is now known as SamB [03:47] poolie: I'm currently thinking I'll put it back for now, so that the other changes aren't blocked. [03:47] k [03:47] I still think we want to get rid of it :) [03:53] * igc lunch [03:57] spiv, lifeless: got any ideas on how I can trigger a logged exception onthe staging smartserver? [04:01] jml: what bzr version? 1.16.1? [04:01] jml: what bzr version is it running? [04:01] jinx [04:01] jml: if so, call any unknown smart method... [04:01] jml: more creatively, issue a ('get', '../foo') [04:02] I'm not 100% sure, either 1.16.1 or 1.17rc1 [04:02] (lp:bzr-ping is a good, short plugin to crib from if you want inspiration for generating arbitrary simple requests) [04:03] spiv, thanks [04:04] although I get stopped on the client side when I try client.call('oprah') [04:05] * jml digs around [04:05] jelmer: what is this about? [04:05] Now on revision 3127. [04:05] Conflicting tags: [04:05] bzr-svn-0.6.3 [04:06] * SamB also wishes bzr would have told him what revision it started at at the end there ... [04:06] oh, just four revs but tons of pyflakes fixes? [04:07] * SamB wonders what pyflakes is [04:07] well, I mean, I can rather guess that it's some sort of lint tool ;-) [04:10] jelmer: oh, you tagged the wrong 0.6.2 as 0.6.3 :-P [04:11] er. s/the wrong// [04:12] When I try to do a checkout with tortoisebzr, I get error: TortoisebzrCommandError: Illegal command 'revert' [04:13] jml: the client side will see the error bit in the response and raise the response as ErrorFromSmartServer rather than returning it. [04:13] Sorry, unknown command 'revert' [04:13] spiv, I had to comment out this line -- #self._run_call_hooks(method, args, body, readv_body) -- for it to talk to the server at all [04:14] malibu: geeze, they really should allow selecting the text in those dialogs so you can copy/paste it! [04:14] jml: weird [04:15] yeah I know [04:15] jml: do you have some sort of weird plugin installed that installs a buggy hook? [04:17] jml: oh, I know. [04:17] jml: will fix, thanks for the report. [04:17] spiv, http://paste.ubuntu.com/222366/ [04:17] spiv, np [04:18] back home [04:18] jml: (it's a bug in the hook function installed by -Dhpss) [04:18] * SamB is beginning to think there's a *reason* it's called Bundle Buggy [04:28] Hmm, I'm pretty sure I type my gpg passphrase faster shortly after having a coffee. Pity that I seem to also mistype more too... [04:30] f [04:40] hmm, not bzr 1.18 milestone [04:40] * lifeless maksifies [04:41] hmm [04:42] * SamB hopes igc notices https://code.launchpad.net/~naesten/bzr-fastimport/400960-heredocs/+merge/9015 soon [04:44] * SamB got his wires crossed and requested some random unrelated person to review it :-( [04:52] spiv, i've wanted to write some program that tracks my typo rate [04:52] and typing rate [04:53] WTB review: https://code.edge.launchpad.net/~lifeless/bzr/bug-367632/+merge/8928 [04:53] want to buy? [04:53] yes [04:53] lol [04:53] lifeless: I took a quick look at that one this morning [04:53] mary's sig the other day said something about "Do real heros start their career squashing household pests?" [04:54] lifeless: I thought we ought to get the apply-inv-delta one landed first [04:54] igc: why? [04:54] igc: the bug 367632 bad deltas are already caught [04:54] Launchpad bug 367632 in bzr "bzr revert '.' can fail when there are added, missing, directories." [High,Fix committed] https://launchpad.net/bugs/367632 [04:54] given I think it depends on it doesn't it (wrt the corruption fix)? [04:54] this just fixes the cause [04:55] igc: bug 367632 was about the delta generation. bug 367633 was the corruption [04:55] Launchpad bug 367633 in bzr "errors on revert propogate to dirstate" [High,Fix committed] https://launchpad.net/bugs/367633 [04:56] hey is bazaar 1.17 not ready or something? I can't seem to use tortoisebzr with it [04:57] lifeless: ok, I'll look a bit deeper later today [04:57] igc: its very shallow [04:57] though it took me ages to dig through the layers to find it ;) [04:57] poolie: typo rate judged by % of backspace hits? [04:57] lifeless: maybe the bug comments confused me or something [04:57] yes, or ^w or alt-backspace or similar [04:58] obviously this is not perfect [04:58] Yeah. [04:58] It won't notice my mistyped passphrase, for example :) [04:58] But that's a bit of a special case anyhow. [04:58] SamB: I'll take a look [05:02] spiv, why not? [05:02] because you tend to type C-a and retype the whole thing? [05:03] also xchat ^w =close window sucks :) [05:05] poolie: because I am usually typing it into mutt, hitting enter, then getting the 'bad passphrase' message, so no ^A, ^W or backspace is involved. [05:05] Does bzr support nested repositories? [05:05] malibu: seems to be in-progress [05:05] oh i see [05:05] I really should futz about with making mutt and gnome-gpg or whatever is shiny these days play nice with each other. [05:06] spiv: it depends what GTK theme you have configured [05:06] But anything related to driving gpg very quickly drives me into a mini rage about how awful gpg... [05:07] 'how awful gpg is', that is. It's very unfriendly software. [05:07] it could be worse -- it could be x.509! [05:08] Heh. [05:09] My issue isn't with the underlying format/spec, it's just that the software to use and manipulate keys is so damn awful. [05:09] what's the trendy PPA for bzr nightlies nowadays? [05:09] google bzr nightly ppa [05:09] :) [05:09] spiv: well, okay, it has a UI that might have been nice when I was born ;-) [05:09] pretty sure that's right [05:10] for reference, I was born in '86 [05:10] poolie: okay ... that seems to be what I'm using ... [05:10] * SamB wishes those things had the date in the version somewhere [05:10] that would be nice [05:11] I don't have a clue what to report that wish against, though! [05:11] SamB: it has a UI that makes me pine for tla :P [05:11] spiv: no, it's not nearly *that* bad! [05:11] i hope that's not a pun :) [05:11] I mean, you don't have to use it that much [05:11] poolie: pretty sure it's not [05:11] I don't see how tla relates to an email client he doesn't even use [05:11] SamB: I dunno, tla made me want to drill holes in my head less than gpg. [05:12] spiv: when was this? [05:12] had you used a VCS with a half-decent UI at the time? [05:12] Stiil, I can believe it's a matter of taste about whether awful thing A is more or less awful than awful thing B :) [05:13] I mean, GPG may have a really rusty UI [05:13] but it doesn't have those %@%^#^ stupid paths! [05:13] or revision names or whatever the heck those things are [05:16] jelmer: bzr-svn needz moar -D ! [05:17] * SamB can't seem to svn-import from https://dosemu.svn.sourceforge.net/svnroot/dosemu today :-( [05:18] spiv, has the fix for bug 400535 landed? [05:18] Launchpad bug 400535 in bzr "ChrootServer/ChrootTransport not used by "bzr serve"" [Critical,Fix committed] https://launchpad.net/bugs/400535 [05:18] lifeless, bug 397705 is marked for 1.17 & not fix released -- what's missing? [05:18] Launchpad bug 397705 in bzr "inventory delta consistency checks are missing known problems and not universally applied." [Critical,Fix committed] https://launchpad.net/bugs/397705 [05:19] jml: it was a placeholder [05:19] jml: ignore it [05:19] jml: it will be in 1.18 [05:20] jml: (and you have the bit I wanted to get into 1.17 already) [05:20] oh, actually, svn can't connect either :-( [05:22] lifeless, cool. [05:22] lifeless, consider it ignored :) [05:24] So I created a repo on my ubuntu server, was able to check out the empty repo with tortoisebzr, but when I add files to the local working copy I don't get add or commit options in the context menu ? [05:25] Not sure what I am doing wrong [05:26] Oh ok now I get the commit option.. wierd! [05:27] jml: I sent it to PQM (for bzr.dev) a while ago. [05:27] jml: it's still playing, but on the second half of the test run by the looks. [05:28] spiv, cool. I guess I'll cherrypick it into 1.17 then merge back into bzr.dev [05:29] jml: sounds good. It should cherrypick neatly aside from the NEWS entry. [05:30] \o/ [05:35] grunk. I think I screwed up one of the commits to 1.17 :\ [05:35] quick! darcs rollback! ... [05:36] jml: tip looks odd [05:37] lifeless, yeah exactly. [05:37] what, did he lose the old one somehow ? [05:38] lifeless, I'm going to uncommit it and then check the news file to make sure that all of the bugs marked for 1.17 are still present [05:38] (not being able to think of a better way to sanity check) [05:39] you can do that? [05:39] (uncommit, I mean) [05:39] SamB: bzr help uncommit :) [05:39] SamB: (often followed by a push --overwrite...) [05:39] I meant in a "you won't get flayed alive for it" [05:39] kind of way [05:40] Well, that depends :) [05:40] I did actually know about the command [05:40] it's handy [05:40] oh, yeah, and I did think some branches disallowed push --overwrite? [05:43] I forget if the branch.conf option disables that, but certainly you could write a smart server hook fairly easily that would reject that sort of change. [05:50] jml: seems fine, get spm to uncommit the master for you [05:51] jml: ? [05:51] lifeless, I can do that myself, actually. [05:51] even better! :-) [05:51] jml: how? [05:51] lifeless, I have write permissions to the branch on launchpad. [05:51] jml: thats not the master [05:51] jml: you need access to balleny tho [05:52] lifeless, I thought PQM submitting the 1.17 branch to Launchpad these days? [05:52] jml: yes [05:52] jml: pqm pushes there [05:52] that doesn't make it the master [05:53] jml: btw, my fix for 400535 has landed. [05:53] lifeless, what makes a branch the master branch then? [05:53] spiv, thank you [05:54] jml: the place pqm reads the branch from, which is a secured directory on the pqm server [05:54] pqm can be configured to read from launchpad, but we haven't done that. [05:54] have you guys ever seen this happen? Its happened to me again.. no add/commit options on my working copy [05:56] lifeless, ahh ok. [05:56] malibu: no, never [05:57] igc: Its fun, finding this many existing, hidden bugs [05:57] * SamB has never even tried to use bzr on Windows [05:59] malibu: give bzr-explorer a spin - it's a bit better supported than TortoiseBzr is currently [06:00] lifeless: fun is one name for it :-) [06:00] samB: that fastimport change is good - merge to trunk now [06:00] SamB: ^^^ [06:01] igc: what, you think I'm case-sensitive or something? this is IRC ... [06:01] ... but how am *I* supposed to do the merge? [06:01] malibu: I hope to release bzr-explorer 0.5 later tonight [06:01] igc: found one test making a delta of [06:01] * spiv -> lunch [06:01] [('', '', None, InventoryDirectory(None, ...))] [06:01] SamB: I did the merge. Sorry typo above [06:01] spm, can you please uncommit the top revision of the 1.17 branch? [06:01] igc: that is, a root directory with file id None [06:02] lifeless: yuk [06:02] igc: oh good [06:02] that saves some awkward confusion ;-) [06:02] SamB: I always assumed IRC was case-sensitive [06:03] if not, you learn something new every day :-) [06:03] IGC: I suppose you do ;-) [06:04] at least, it's a pretty boring day if you don't [06:04] abentley: don't suppose you're around? [06:05] lifeless: maybe those groupcompress tests fail if make isn't run. I'll try running make and see if they go away [06:05] igc: if they fail, I suggest we let PQM decide [06:05] igc: I found another bug and wrote a test ... would you like the test without a fix? [06:06] SamB: feel free to push the branch [06:06] https://code.launchpad.net/~naesten/bzr-fastimport/401249-tag-from-ref [06:07] or lp:~naesten/bzr-fastimport/401249-tag-from-ref [06:07] samB: I can't merge it to trunk though until the fix is there because 'bzr selftest fastimport' would start failing then [06:07] igc: ah. [06:07] jml: to verify: uncommit+revert(?) revno 4531 in ~/archives/thelove/bzr/1.17 [06:08] spm, that's right. thanks. [06:09] igc: I think I could write an ugly fix pretty easily, but in the time that it took me to figure out how to write the test, I changed my mind about wanting to do it that way ;-) [06:10] igc: and \o/ found a form of corruption I'd overlooked [06:10] jml: bzr/1.17$ bzr revno ==> 4530 [06:10] spm, thank you [06:11] np [06:12] lifeless: those groupcompress tests pass after running make [06:14] that needs a bug then :) [06:16] igc: well, thanks for merging that [06:19] lifeless: yup, I'll raise one now [06:21] okay, how do I go back 4 revisions ? [06:21] bzr-svn is acting up :-( [06:22] lifeless: so the tests pass until your patch is applied: "bzr trunk bzr.dev xx; cd xx; bzr selftest groupcompress" works fine [06:22] lifeless: it's not an existing bug then I feel [06:22] igc: thats interesting; howver I haven't touched the VFs layer, so I have to assume it just uncovers a latent bug [06:23] lifeless: wouldn't be the first one with this patch :-) [06:26] ah ... bzr pull --overwrite -r 'last:5' . [06:36] what the heck ... now lp:bzr-fastimport hangs when I try to pull from it :-( [06:39] spiv, can you please sanity check: http://paste.ubuntu.com/222407/ [06:40] * SamB decides to hope it goes away by morning ... [06:43] gosh it's nice being able to develop in karmic, as opposed to hardy chroot. [06:54] jml: looks ok [06:54] spiv, thank you. [06:54] biab [07:06] igc: done! [07:11] lifeless: sweet. I'll take a look. [07:12] as for the gc tests [07:12] they have [07:12] self.requireFeature(CompiledGroupCompressFeature) [07:12] so they explicitly don't expect to run without the .so [07:13] igc: I'm done for the day; if you can let me know about both merges that would be great [07:14] lifeless: np [07:17] igc: the new changes in the apply-delta branch are, a new check for the thing that was corrupt in the deltas from tree transform and test smart, and fixes to those two tests to not create bogus deltas [07:18] lifeless: yup - double-clicking the entries in qlog is telling me exactly that :-) [07:20] lifeless: do we allow unicode file-ids? [07:20] no [07:20] dirstate will barf [07:21] we used to, but we did a sweep over the code base ages ago and redefined as utf8 strings [07:21] lifeless: thanks [07:21] theres a certain amount of 'whatever works [08:06] can we have bzr lolwhut too? [08:35] I just did this: commit, push, tag, push [08:35] but the final push says "No new revisions to push." [08:35] am I doomed to have this tag loiter eternally in my local branch? [08:35] jml: no [08:36] jml: the message is technically correct, but misleading [08:36] jml: no *revisions* were pushed [08:36] jml: the tags should have been, though. [08:36] jml: there's a bug about this UI flaw [08:36] spiv, how can I verify this? [08:36] jml: try "bzr tags -d URL" [08:37] "bzr tags -d :push", even! [08:37] http://paste.ubuntu.com/222489/ [08:37] I just got that error [08:37] jml: ooh, that looks like a bug! [08:38] jml: probably shallow, please file :( [08:38] * spiv needs to head off for the evening. [08:38] jml: happy releasing! [08:38] spiv, thanks. [08:40] lifeless: what is the actual point of subunit2pyunit?+ [08:40] to turn it back into the same output format produced by subunit? [08:43] poolie: to replay a test via pyunit; this effectively gets rid of all the subunit encoding, and will tend to hide passing tests etc, depening on the pyunit parameters [08:44] actually, i meant "in your example pipeline" [08:44] oh [08:44] so I can see it realtime [08:45] in a nicer format than just letting subunit flow to stdout? [08:45] yah [08:45] and as you say hiding passing tests [08:45] well the subunit-filter takes those out, in my example [08:45] after time reporting [08:45] my next main thing is progress data in the stream [08:46] once thats done I plan to do a subunit2bzrtest, or something like that, that will use bzr's progress bars to pretty-print streams === quicksil1er is now known as quicksilver [09:52] igc: hello [09:53] moin [10:18] hi bialix - great to have you back! [10:18] hi jelmer [10:18] igc: well, I'd like to be still on vacations, it was great but too short [10:22] igc: I've merged branch of amanica before I've read your objections. Should I uncommit? [10:22] bialix: no [10:23] I guess we can tweak it later [10:26]