/srv/irclogs.ubuntu.com/2011/03/25/#bzr.txt

jxselfpoolie: Will do, thanks. There's more than a decade of history in CVS and I don't want to lose that.00:01
jelmerwgrant: Is it no longer possible for me to change Fix Released Ubuntu bugs back to any other status?00:06
jelmerI accidentally changed a bug to Fix Released, now I can't change it back.00:06
wgrantjelmer: You can't reopen a bug unless you are a bug supervisor or the reporter.00:07
wgrantjelmer: Which bug?00:07
jelmerwgrant: the ubuntu bit of https://bugs.launchpad.net/ubuntu/+source/loggerhead/+bug/58661100:08
ubot5Ubuntu bug 586611 in loggerhead "loggerhead needs a hard dependency on simplejson" [Medium,Fix committed]00:08
poolieis that new?00:08
wgrantpoolie: Yes :(00:08
wgrantMore misfeatures.00:08
pooliei don't suppose it's documented other than in the source?00:08
wgrantjelmer: What should it be?00:08
wgrantpoolie: lol00:08
jelmerwgrant: In Progress I guess (there's an upload going into Debian that fixes it that'll request a sync for)00:09
jelmers/that'll/that I'll/00:09
wgrantFixed.00:09
jelmerwgrant: thanks00:09
mr-russ-workwhat's the best windows bzr client?00:47
pooliebzr explorer00:48
mr-russ-workdoes it need bazaar installed?00:52
pooliemr-russ-work, yes01:08
pooliebut i think they're both included in the default windows installer?01:08
pooliehi spiv?01:16
=== r0bby is now known as robbyoconnor
knighthawkthanks poolie01:33
spivHi poolie01:40
spmpoolie: jubany now has: ii  bzr 2.3.1-0~0.IS.8.04; will update the RT shortly01:41
spivspm: thanks!01:41
spivI'll give those imports a kick then01:41
spmsweet01:42
pooliehey spiv01:51
pooliei missed you before :)01:51
spivHmm, the package importer is getting lots of “lazr.restfulclient.errors.HTTPError: HTTP Error 500: Internal Server Error”, e.g. http://package-import.ubuntu.com/status/ceph.html#2011-03-25%2002:08:19.18174102:21
pooliethere should be an oops header02:30
pooliegah02:30
pooliei filed a bug asking for lplib to pull that into the exception02:30
poolieyou could fix that02:30
lifelesstimeout bug I suspect02:39
lifelessfix is in trunk, pending deploy due to these security issues02:39
spivI've told the package-importer that it's a transient issue, so it'll automatically retry those failures after 24(?) hours I think02:40
lifelesswell02:40
lifelessthe problem is that filing a merge proposal loads /all/ the branch history for both sides into cache02:41
lifelessso if you want to retry02:41
lifelessdo it right away02:41
lifelesswhile you're still slightly hot02:41
achianghello, i noticed that debcommit has some magic to make a bzr branch automatically link up with a launchpad bug. how can i do that manually? bzr help commit doesn't really tell me much04:13
pooliebzr help mark-uploaded04:13
pooliei think04:13
lifelessmark-uploaded says when something has been uploaded to the archive04:14
poolieah gah04:14
pooliemisread04:14
poolieyou want bzr ci --fixes lp:BUGNUM04:14
achiangooh04:14
achiangok04:14
achianghuh, --fixes not documented in maverick's bzr?04:15
achiangyes it is04:15
achiangi just can't read04:15
achiangpoolie: thanks!04:15
pooliein 'bzr help launchpad' too04:15
poolieyou're welcome04:15
jderosehow do i print a summary of commits per person/ lines of change per person?  looked at bzr stats, but it doesn't seem able (or maybe is sort of broken on bzr 2.3.0/natty)06:46
pooliei thought it did that sort of thing06:53
achiangis there a way to do a bzr commit interactive? i only want to commit one hunk06:54
poolieinstall bzr-interactive i think06:54
achiangnothing shows up in apt-cache search?06:55
pooliejderose, what happens from 'bzr stats'?06:55
achiangoh, it's a plugin06:55
poolieyep, bzr branch lp:bzr-interactive ~/.bazaar/plugins/interactive06:56
achianginteresting, i'll try that06:56
pooliejderose, that seems to work for me06:56
jderosepoolie: turns out, something totally awesome! i've been up hacking too long at this point... was in the wrong branch! http://paste.ubuntu.com/585230/06:57
pooliehaha :)06:57
jderosepoolie: sorry for the dumb question, thanks for the smart answer :)06:58
poolienp, glad it worked06:58
pooliedon't know if it can count slocs06:58
pooliewhich obviously are highly useful in awarding bonuses06:58
jderoseah, gotcha... that would be cool06:58
jderosepoolie: np, i work for this crappy startup that doesn't pay anyone anything!06:59
* jderose is CEO :)06:59
pooliecongrats / good luck :)07:01
jderosepoolie: thanks. :) while i have you... out of curiosity, is there any timeline for addressing the submodules issue with bzr? - https://bugs.launchpad.net/bzr-git/+bug/40281407:03
ubot5Ubuntu bug 402814 in Launchpad itself "Importing revisions with submodules is not supported" [Wishlist,Triaged]07:03
pooliehm it should go up from wishlist07:04
jderosepoolie: i keep telling everyone how awesome daily builds are, help them set them up, to find i can't import into bzr... i'm personally very interesting in see daily builds for gstreamer and pitivi07:04
poolieyeah, it's a big blocker for that07:04
poolieit's not close enough i could give a very reliable estimate but it's definitely in our target area07:05
jderosepoolie: i know there are always too many things to do and it's hard to pick priorities --- no judgments! -- but just wondering :P07:05
pooliemaybe in the next few months07:05
jderoseso this requires a pretty big change in bzr? new storage format big?07:06
jderosepoolie: anyway, bzr and launchpad are fantastic - that's for tools that have made my live easier! :)07:09
pooliejust a bug07:09
pooliefighting with all the others to be born07:09
jderosepoolie: gotcha.  thanks again!07:12
pooliehi jam!07:21
vilahi all !07:37
pooliehi there vial07:39
poolie*vila07:39
vilaI can't find back the reference, but someone mentioned that reviews can gain a 'Agreed' vote instead of an 'Approved' one07:39
pooliereally?07:39
vilaIt makes a lot of sense to me07:39
vilaIt gives a very different meaning and implies a different relationship which I think is far more appropriate07:40
pooliei just haven't heard of it07:40
vilaMost of the time I "agree" with the proposal as a peer instead of "approving" it as a semi-god :-D07:41
vilaha, got it: http://micknelson.wordpress.com/2011/03/22/code-reviews-as-relationship-builders/07:42
pooliehttps://lpstats.canonical.com/graphs/BzrProjectBugActivity/ -- everyone came back from holidays :)07:43
poolienice idea07:43
vilalol07:43
pooliepersonally i would rather get 'tweak'07:43
pooliehttp://en.wikipedia.org/wiki/I_approve_this_message07:44
vilayeah, it's needed more often07:44
pooliei think 'approval' is quite a positive sentiment though07:46
vilapoolie: granted there are positive values in approval but I think they are also there in agreement ;)08:00
pooliehaha :) indeed08:01
* vila back to losa hunting08:03
=== psynaptic is now known as psynaptic|afk
jammorning all09:09
vilamorning jam :)09:09
pooliejam, vila: should i add a mechanism to deprecate commands, or just delete them?09:12
pooliei guess the former09:12
jampoolie: so you get "bzr clone is going to be removed in the next release" messages?09:12
poolieyeah09:12
jamWe've loosened the deprecation stuff for api, but I think we probably want to do that for command  line.09:12
pooliesuppressable messages09:12
vilaconfigurable as we did for warnings09:13
jamside note: "safe and easy web browser from Mozilla" (firefox)09:14
poolie?09:14
jamjust very funny to read the "helpful descriptions" rather than the name of the program09:14
jamUpdate Manager09:14
poolieoh, yeah, that is strange09:14
jamThey also have a very strange (and inconsistent) Capitalization09:14
jampoolie: I'm tempted to go so many different ways.09:15
jamI would tend to say... if it is hard to do, punt09:15
jambut if it is easy, please deprecate09:15
poolieagree09:15
poolieit's not _that_ hard09:15
jamThat said09:15
pooliesome of the command framework stuff could do with some love09:15
jamif people are using "bzr clone" are they going to notice it tell them it will be changing?09:15
pooliewhy wouldn't they?09:18
jampoolie: because of all the other stuff that they ignore. They could.09:19
jamI just want to make sure that messages we send are relevant09:19
jamThat people who *want* to see them are doing so.09:20
poolieso i can think of two ways to do this09:20
poolieone is, to put it into the Command object09:20
pooliewe could just refuse it09:20
poolieand make them update09:20
poolietheir muscle memory or scripts09:21
pooliei guess some scripts may be hard to update09:21
jampoolie: There was a recent discussion in python-dev about deprecation09:27
jamIn py, they always do PendingDeprecation in X, Deprecation in X+1, Removed in X+209:27
poolieok09:27
jamPeople were complaining09:27
pooliewhat's pending deprecation?09:27
pooliejust in the docs?09:28
jampoolie: I think it is a by-default suppressed warning09:28
jamversus a vocal warning09:28
jamversus gone09:28
jamthe complaint was because they were trying to go from 2.X to 3, and running afoul of09:28
jamWell it was deprecated in 3.1, so removed in 3.209:28
poolie:)09:28
jambut 2.7 still had something09:28
jamso there was no way to support 2.7 *and* 3.309:29
jamor whatever09:29
jampoolie: but that also highlights something09:29
jampeople really care about Deprecation as a way to transition09:29
jamif we are just removing aliases to commands that already exist09:29
jamthey can switch to the existing command, and be supported across all versions09:29
poolieright09:30
poolieand, they should be able to use bzr aliases if for some reason they can't change09:30
jamI think people have trouble if it was "foo" in 2.0, and "bar" in 2.4, and they don't have an easy way to support the versions they want09:30
jampoolie: My experience in the past, was that DeprecationWarning wasn't helping people09:31
jamDevelopers tend to use crack-of-the-day09:31
jamand so never see them, because they are using the new apis09:31
jamUsers see them a lot, but don't have anything they can *do* about it09:31
poolieyes09:31
poolieand yes09:31
poolieshowing deprecationwarning to users is nuts09:31
jampoolie: I *think* we can do this for commands in a tasteful way09:32
jambecause there it is much more likely that a user is invoking the command that needs to be changed09:32
jamThere are scripts that don't apply09:32
jam(if I'm using, eg, etckeeper, then don't give me deprecations cause Upstream needs to fix it)09:32
pooliealso, it's probably unusual that someone chose to use these at all09:33
jampoolie: then my recommendation is to troll through bzr docs, especially ones not written by us09:33
jamI don't know of any particularly popular ones09:34
jambut I'm pretty sure I've seen wiki-like docs09:34
jamthat said "bzr clone ..."09:34
jamor maybe "bzr get ..."09:34
pooliemaybe09:34
pooliethose docs may have bigger problems09:34
poolieanyhow, fair enough09:34
jamanyway, that's the only place where I've seen recommendations for people to use non-official names09:34
poolieadding a command deprecation layer just for this does not feel like the most direct route possible09:34
pooliek09:34
pooliethanks09:35
jamAnd I think some people like "bzr get" because it was shorter to type09:35
pooliei'm happy to add 'bzr br' or similar09:35
poolieonce we sort this out09:35
spivjam: yes, I know of at least user that prefers "bzr get"09:37
spivI don't recall hearing anyone use or advocate using our current "bzr clone" alias.09:37
jamspiv: I think I've seen it in something like "bzr-for-git" discussion09:39
pooliespiv, bialix mentioned in the bug he prefers it09:41
poolieor was it someone else too?09:41
jam(because the command is more like 'git clone' than 'git branch')09:41
jampoolie: I'm pretty sure the one I remember is bialix09:41
pooliewell, they might say 'bzr clone' too then09:41
pooliebut it's not all that much like it09:41
spivI'm mainly thinking of glyph, but presumably these aren't the only people :)09:43
pooliewell, 'bzr alias get=branch' will restore it09:44
jampoolie: so I think you bring up a reasonable statement09:46
jamfail09:46
jambut give them advice on what to do09:46
jam"bzr get" --- "bzr get was removed in bzr 2.4. The recommended name is "bzr branch", to restore "bzr get" as an alias, use ..."09:46
jam(bzr get was removed as an alias in bzr 2.4)09:47
jampoolie: sort of inbetween "deprecated and I still work", and "where did the command go that I've been using for 3 years"09:47
poolieyep09:49
pooliei might implement that by adding a new command that prints the message, and binding them to that09:49
jampoolie: a few ways to do it. You could add "old_aliases = [...]" to Command objects09:49
jamif there are a lot of them09:49
pooliei was going to do that originally09:50
jamI think it depends how hard it is to get the appropriate text09:50
poolieit seems a bit inelegant to put it inline with the regular command stuff09:50
jambut I guess that is what https://code.launchpad.net/~mbp/bzr/506265-command-deprecation/+merge/54828 is about?09:50
ubot5Ubuntu bug 506265 in Bazaar "deprecate old command names" [High,In progress]09:50
jamSo the "cmd_I_was_removed_in_24" knows what it is aliased to now?09:50
poolieright09:53
pooliei think i might stop for today though09:55
jamspiv: You mentioned you could use "lp:~/project/branch" and we'll expand ~ to ~user. When you said that, I thought "really, since when".09:58
jamI just looked through the logs, and *I* implemented it09:58
jam(2.3b1)09:58
spivjam: :)10:36
lifelessspiv: should be able to retry those timeouts now10:58
jambbiab, lunch11:04
spivlifeless: I think they retry every hour actually12:02
spivOh, hmm, maybe not.12:03
spivI'll give it a kick anyway12:03
spivlifeless: still failing12:35
james_werk12:51
james_wI don't like the look of http://package-import.ubuntu.com/status/7ae30b44f1f904e04bbae6700b7181dd.html12:52
james_wand http://package-import.ubuntu.com/status/3321c5cfcfcf74249b92f963453f4701.html is pretty worrying12:52
james_was is http://package-import.ubuntu.com/status/ttf-arphic-gkai00mp.html12:53
james_whttp://package-import.ubuntu.com/status/ttf-arphic-uming.html12:53
james_whttp://package-import.ubuntu.com/status/ttf-arphic-ukai.html12:54
vilajames_w: I think that's what lifeless and spiv were just talking about12:54
james_wthough I guess those are transient with the upgrade of bzr?12:54
vilaeerk, the no module named revisiontree is different and indeed very worrying12:54
vilajames_w: 'no module ...' requeued12:56
vilajames_w: and they seem to be passing now, ttf-arphic-uming requeued too12:58
james_wvila, great, thanks12:59
james_wit would be great to have a packaging system that didn't break python on upgrades12:59
vilathanks for noticing, I think.. yeah :)12:59
vilajames_w: in the mean time, stopping the importer may help...12:59
james_wyeah12:59
=== mrevell is now known as mrevell-lunch
spivjames_w: in principle the packaging system could do "mv bzrlib .old.bzrlib; mv .new.bzrlib bzrlib; rm -r bzrlib" I suppose, but I can just imagine how that would actually be pretty hairy to actually implement13:26
spivjames_w: (and of course even that approach has a small window of brokenness)13:26
james_wspiv, my guess is that dpkg is fine, it's the Python-specific stuff on top that isn't, and hopefully that can go away in a few years when Barry's stuff is widely available13:27
spivAlthough I guess it's hard to avoid the risk of breaking running processes13:27
spivEven firefox wants a restart after its package is upgraded ;)13:28
spivjames_w: clearly bzr should not use lazy_import so that it gets all its imports done in as short a window as possible ;)13:29
james_wyeah :-)13:29
james_wif we can't even import bzrlib sometimes though... :-)13:29
=== psynaptic|afk is now known as psynaptic
spivOr just have the old processes keep running against an older btrfs snapshot of the system, or something handwavy like that...13:30
spivAnyway, zzz time!13:31
james_wnight13:34
=== mrevell-lunch is now known as mrevell
jamjelmer: you landed the bzr-hg change faster than Launchpad built the merge proposal for it. so there isn't a diff to read.14:41
jamalso, weren't we wanting to pair on some changes?14:41
jamI don't remember which ones off-hand14:41
jambut maybe we could do that early next week?14:41
jelmerjam: I've mostly been using merge proposals as a way to track individual changes, but self-reviewing as I always have14:42
jelmerjam: Yeah, some pair programming would be nice14:42
jamjelmer: sure. Just mentioning that tracking the changes isn't possible when you land faster than they get tracked :)14:42
jelmerjam: heh, fair enough :)14:43
=== deryck is now known as deryck[lunch]
=== psynaptic is now known as psynaptic|lunch
=== psynaptic|lunch is now known as psynaptic
=== deryck[lunch] is now known as deryck
=== Ursinha is now known as Ursinha-lunch
=== psynaptic is now known as psynaptic|afk
=== psynaptic|afk is now known as psynaptic
=== Ursinha-lunch is now known as Ursinha
=== psynaptic is now known as psynaptic|food
=== herb__ is now known as herb
=== mnepton is now known as mneptok
achiang_hello, how do i resolve a merge conflict? i get a foo.BASE and a foo.THIS, not sure which one is which21:32
achiang_but the base file foo doesn't seem to exist at all21:34
achiang_ah, i think that means my merge source rm'ed foo21:34
=== psynaptic|food is now known as psynaptic|afk
wgrantmaxb: Hi.23:18
ablmfHow to automatically "push" after every commit?23:34
bouncingzipuse a bound branch.23:35
ablmfbouncingzip: do u mean use a "checkout"?23:36
bouncingzipactually, depends what you're trying to do exactly23:40
bouncingzipmaybe relevant: http://stackoverflow.com/questions/460632/bazaar-bound-branch-commit-and-update23:40
bouncingzipif you want the tree somewhere remote, those are suggestions on how to do it, if you want a remote branch automatically mirrored each commit, binding to it (which yes, is the same as using checkout originally) is probably what you want.23:44
ablmfbouncingzip: I think the plugin helps, thx!23:47
=== psynaptic|afk is now known as psynaptic

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!