jmlmwhudson: I had a little trouble with it yesterday.00:00
mwhudsonjml: oh, ok00:00
mwhudsonjml: i thought you were clone('null:') ing00:00
mwhudsonor am i out of date00:00
jmlmwhudson: that was my second pass at a solution00:01
jmlmwhudson: after I couldn't find the repo format controls for create_branch, ISTR00:02
spivjml: tried sprout('null:')? :)00:02
mwhudsonspiv: blows up with looms and weaves00:02
spivOh, really?00:02
jmlspiv: clone('null:') works great except for looms & weaves. haven't tried sprout00:02
spivI'm not totally shocked by looms failing, weaves surprises me though.00:02
mwhudsonspiv: it might be a null:/None thing00:03
jmlspiv: weaves raises... RepositoryUpgradeRequired if you force no working trees...00:03
jmland NoSuchRevision if you pass 'null:' without caring about working trees.00:03
jmlNone works, but I didn't check that the new branch was empty00:04
jml(it might have just cloned the whole thing)00:04
jmlwhen I found out that looms failed I gave up and moved on to fixing bugs.00:05
jamspiv: Is poolie not around?00:05
jmlspiv: anyway, what were you going to suggest?00:06
spivjam: not that I've seen00:06
jamjml: "b = bzrlib.branch.Branch.open('foo'), b.clone(NULL_REVISION)" ?00:06
jmljam: rather than bzrdir.clone?00:07
jamjml: well, it is something to try00:07
jamLooms is a bug in looms00:07
jmljam: yeah.00:07
jamand I don't care about weaves :)00:07
jamjml: you can also "bzr uncommit -r 1"00:07
jamAnd I *think* "bzr uncommit -r 0"00:07
jambut I'm not sure about '0'00:08
jmljam: I'm not sure if I'm allowed to not care about weaves yet.00:08
jamjml: alternatively00:08
jamGo back to the raw Format objects00:08
jamand tell them to create something new00:09
jamIt may not work on pre-metadir, though.00:09
mwhudsonwith appropriate care for Remote*00:09
jamAnd I don't have much of an idea about looms00:09
jamspiv: ok, so no standup today? Was there one last night? I felt like I missed it, but didn't see any actual chatter00:09
jmljam, spiv: thanks. I'll have a play.00:10
jmlmaybe I'll end up patching loom or something.00:10
pooliehello jam, spiv, sorry i'm late00:10
spivpoolie: hello00:11
spivjam: not so far, but now poolie is here it'd be good to do one...00:11
spivpoolie: standup?00:12
pooliesounds good00:12
jamjml: though you are right, I don't think Branch5/6 (regardless of looms/weaves) truly supported cloning at the NULL_REVISION00:12
spivjam: So about landing my trivial -Dhpss fix on the 1.7 branch...00:13
spivjam: what should I do with the NEWS entry?  Start a bzr 1.7rc2 section?00:13
jamspiv: I'm trying to land it now00:13
jamlook at PQM00:13
jamPQM is backed up by a merge from vila00:14
spivjam: ah.00:14
jamwhich says "success" but is still stuck00:14
spivjam: ok, I'll wait for that to sort itself out then :)00:15
spivjam: thanks00:15
thumperis there a bzrtools coming soon that won't be tried to be removed with bzr 1.7rc from the PPA?00:33
abentleythumper: I'll release bzrtools 1.7 tonight.00:45
thumperabentley: cool00:46
Gioacchinogood night00:58
=== rocky1 is now known as rocky
jmlcan someone land my http smartserver tweaks patch please?01:47
=== kiko is now known as kiko-afk
spivjml: If someone unwedges pqm I will...01:58
pooliei don't see stubub yet02:00
pooliei'll ask steve-in-canberra02:00
jmlpoolie: Steve-in-London, I think you'll find02:03
YasumotoHey guys, is there a way to get just the files when you're making a new branch, rather than all the files inside a folder?04:39
RAOFYasumoto: I'm not entirely sure what you're asking for.  Are you saying that, rather than 'bzr branch foo bar' branching 'foo' into a 'bar' directory, you'd like to branch 'foo' into the current directory?04:40
YasumotoRAOF: actually, I may be asking about bzr branch foo bar04:44
Yasumoto(a friend and I are learning how to use bzr for an assignment together)04:45
Yasumotocool, thanks RAOF! :)04:46
RAOFHeh!  Remember, 'bzr help <command>' often contains useful information, and 'bzr help commands' will list all the commands :)04:49
Yasumotoooh, cool04:51
Yasumotothank you04:51
* jml writes a function that does get_stacked_on_url that return None when there isn't one.05:08
pooliejml, yeah, i'm not sure if that was the best api05:42
poolieto raise an error05:42
pooliespiv, how's it going?05:42
pooliei'm still trying to get my desktop back up :-/05:43
jmlpoolie: well, the advantage of the current API is that I *can* write such a wrapper.05:43
jmlpoolie: if it returned None and I wanted rich errors raised, I'd be in trouble.05:43
poolieif it returned None for more than one case, yes05:44
spivpoolie: pretty well.  I came across a vim window that reminded that I had a fix for a TooManyConcurrentRequests bug, so I finished that and fired it off.05:44
jmlright. So I'm returning None if it's NotStacked as well as if one of the formats doesn't support stacking.05:44
jmlbecause for most of what I'm doing the two are equivalent.05:45
poolieok, good05:45
* jml heads out05:45
pooliespiv, can you send your startup-time branch too (if you didn't already)05:45
* spiv re-reads what he actually did in that branch to make sure there's nothing outrageous...05:46
Yasumotohey guys, my friend's trying to push a branch onto launchpad (he just made an account!) but we're having issues with the windows client05:51
pooliehi Yasumoto05:51
pooliewhat specifically05:51
Yasumotowe found some bugs ( https://bugs.edge.launchpad.net/bzr/+bug/198092 ) about bzr launch-pad login05:52
ubottuLaunchpad bug 198092 in bzr "Should tell the user to run bzr launchpad-login before push lp:" [High,Fix released]05:52
Yasumotobut it's saying Autentication type (password) not permitted.05:52
Yasumotobzr: ERROR: Connection Error: Unable to authenticate to SSH host as user@bazaar.launchpad.net Bad authentication type (allowed_types=['publickey'])05:53
Yasumotowe're about to try it from cygwin..05:54
spivYasumoto: Launchpad's code hosting only allows publickey auth05:56
spivYasumoto: so you need to add an SSH public key to your Launchpad account05:57
spivYasumoto: https://launchpad.net/people/+me/+editsshkeys05:57
Yasumotospiv: yeah, we've got the public key on launchpad05:57
Yasumotohm, now in cygwin05:58
Yasumoto1 sec05:59
Yasumotohm, so I've got the private key in ~/.ssh/id_rsa and the public key in ~/.ssh/id_rsa.pub06:01
Yasumotoand it's giving us "bzr: ERROR: [Errno 0] Error"06:02
spivDoes "sft -v YOUR_LP_USER@bazaar.launchpad.net" show that it is using the key?06:02
spiv"sftp" rather than "sft"06:03
spivAlso, can you pastebin the traceback in your ~/.bzr.log file corresponding to the "Errno 0" error?06:03
Yasumotospiv: good call on the sftp, it looks like it's giving us permissions errors06:04
Yasumotoit's ignoring the private key06:04
Yasumotowe'll change those, 1 sec06:04
Yasumotoit's connecting now06:07
Yasumotohowever it's not accepting his passphrase06:08
Yasumotospiv: thanks for your help, by the way :)06:11
Yasumotohe's playing around with things a bit for now, I'll let you know if we make progress06:12
vilaGood morning bazaar (with a special thought for pqm)06:57
AfCIf you have a file added & committed to a branch, and later add a pattern to .bzrignore which happens to match it, it will still be governed as a versioned file, correct?07:14
vilaAfC: yes07:21
jmlhello everyone.07:21
vilahi jml07:21
jmlI would like to suggest that autopack speed become the #1 most important bzr issue.07:22
jmlvila: hello. how's things?07:22
vilajml: very very good :) Except for the fact that pqm is so proud of my last submission that it don't want to let it go (which is embarassing :)07:23
vilaespecially for jam and spiv who are trying to get *their* submissions accepted :)07:23
vilapqm: stop thinking, I don't want it to be merged in the 1.7 branch, go go go07:24
pooliehee hee07:26
poolieis it still stuck?07:26
vilapoolie: pqm unstuck and my submission was even accepted. Did anybody diagnose the root cause ?08:21
poolievila, stub killed a stuck ssh process08:24
pooliehe didn't tell me what process it was specifically08:24
vilapoolie: thks for feedback, I think pqm is just feeling anxious when it hears us talking about running the test suite on multiple platforms, someone should reassure it... I tried to be more gentle next time though08:26
spivTurns out it wasn't so hard to cut down the round trips of an empty push.08:59
poolieguilhembi, hello09:02
guilhembihello poolie09:04
guilhembivila: Bonjour; do you have time for a short question regarding "'bzr missing' does not show merged revisions" ?09:06
vilaguilhembi: sure09:06
guilhembivila: Thanks. You said, "fix committed", but, so, what does "bzr missing" show now? output similar to "bzr log" indented revisions?09:07
* guilhembi rushes to try that09:08
vilayou can try with the associated branch09:08
* guilhembi is evil, he tries with another branch09:08
vilaguilhembi: feedback welcome about response times since this makes missing O(history) when --include-merges is used (but work will continue to remove that)09:09
poolieguilhembi, when he said 'with that branch' he meant the fix is in the branch connected to that bug09:11
pooliei think it's not in trunk yet09:12
vilapoolie, guilhembi : poolie is right, tweaks were required, I'm working on it right now09:13
vilaguilhembi: sry, I misunderstood you, I thought you meant that you will merge my fix in your private bzr09:14
spivI think I just found a bug we really ought to fix before 1.7 :(09:14
vilaso 'fix committed' means that the fix is available somewhere (often in the branch associated with the bug), while 'fix released' means that the patch has been merged to bzr.dev and the series indicates in which released version the fix is available09:15
guilhembivila: but there is time between merge into bzr.dev and building of a release,09:16
guilhembiin this time window what is the status of the bug?09:16
vila'fix released'09:17
guilhembivila: isn't this misleading?09:17
guilhembi'fix released' though not yet in a release?09:17
vilayes, it is09:18
pooliethe bug statuses are not quite labelled right09:18
pooliespiv, want to call me?09:18
spivpoolie: just about to finish mailing the list with details09:20
spivpoolie: short answer: https://bugs.edge.launchpad.net/bzr/+bug/26921409:20
ubottuLaunchpad bug 269214 in bzr "Branching from shared repository via HPSS does not preserve repo format" [High,Confirmed]09:20
pooliethanks for finding it09:20
vilaFor me, 'fix released' means 'this bug is available beginning with the milestone indicated'. If that milestone points to a release "in the work" it means the fix is available in bzr.dev only or in a release candidate if one exists at that point09:21
spivWell, John and someone on the list noticed it really, I just happened to put a couple of pieces together.09:21
guilhembivila: I tested --include-merges on a pair of MySQL branches and it looks like working09:22
spivpoolie: mail sent.  I know how to reproduce, but haven't tried to find the cause yet.09:23
vilaguilhembi: Good.09:23
pooliespiv, if you're still here, the failure in the 261315 test is (obviously really) there are tests for what rpcs are done by remote objects09:58
poolieand this adds a new one09:58
spivpoolie: I'm still here10:03
pooliei'm going to call it a night soon but can we talk it over first?10:04
spiv(And me too)10:04
lifelessvila: the LOSA's can fix wedged pqm's10:32
lifelessvila: I'm not here :P10:33
vilalifeless: Sure I can see that :) What means LOSA though :)10:33
vilalifeless: enjoy your week-end !10:34
lifelesspoolie: pqm still wedged?10:36
spivlifeless: no10:40
spivlifeless: so enjoy your weekend :)10:40
spivlifeless: (stub killed a stuck ssh process)10:41
* vila tries Launchpad Overlord System Administrators. Hmm, sounds good :)10:48
dereineis there something like codespaces for bazaar?11:11
lifelesswhat is a codespaces11:11
james_wdereine: you might want to look at launchpad.net11:13
dereinemh opensource...11:14
dereinei would like to publish it but...11:14
dereinebut thanks for the typ11:14
frk2can somebody help me out with the lp protocol please?12:19
frk2mine seems to skip the http_proxy variable completely12:20
frk2is there an alternate way i can access my launchpad bazaar repo12:20
james_wfrk2: "bzr launchpad-login <your lp id>" will make it use bzr+ssh://12:22
frk2im doing bzr push lp:~fkhan/+junk/test12:23
frk2seems like its throwing an exception on python xmlrpc lib12:23
james_wbug 186920 for the proxy issue I think12:24
ubottuLaunchpad bug 186920 in launchpad-bazaar "bzr launchpad does not handle proxy when used for name resolution " [Undecided,Confirmed] https://launchpad.net/bugs/18692012:24
james_wusing launchpad-login should make it avoid this12:24
frk2is there a work around?12:24
frk2my dns works fine12:25
frk2i have worked around it by using bzr+ssh://bazaar.launchpad.net12:37
frk2that works12:37
frk2however it says no such user on launchpad12:37
hsn_http://pastebin.ca/1200966 you know why shelve doesnt work,  i need to create some directory?12:37
frk2my local username is not the same as my lp username12:37
frk2am i being retarded here?12:41
james_wfrk2: run "bzr launchpad-login <your lp username>" and then continue using "lp:"12:41
james_whsn_: you'll need to put it somewhere other than pastebin.ca if you would like me to look at it, I can't access that site12:41
frk2james_w, connection refused - same problem12:42
james_wfrk2: "no such user"?12:42
frk2the exception happens in python xmlrpclib12:42
frk2no, the earlier problem with lp:12:43
frk2i am being a proxy12:43
james_wbut you've run "launchpad-login"?12:43
james_wdoes the end of the backtrace have "httplib.py"?12:43
frk2i read somewhere that using lp makes bzr do a xmlrpc lookup12:44
frk2which is not proxiable12:44
james_wah, true12:44
james_wuse bzr+ssh://username@bazaar.launchpad.net then12:44
frk2thats all i was looking for12:45
frk2done. just trying out bazaar- looks way cooler than svn which is what im using now12:46
hsn_is there way to do bzr up --dry-run?12:48
Peng_james_w: You can't access pastebin.ca?12:53
james_wPeng_: nope12:56
Peng_james_w: Hmm. It has AAAA records. Maybe you have messed up IPv6 support?12:58
james_wI don't know how I would have "messed it up"12:58
james_wIt resolves to an IP, and traceroute shows that packets get part way there, but just die after a few hops12:59
Peng_Oh. Never mind then.12:59
=== maw_ is now known as mw
vilahmmm, pqm... what's the problem with merging lp: branches ?14:09
* vila refuses to believe that pqm has anything against him :D14:10
vilaAnybody knows someone with a 'stub' nick ?14:11
vilapoolie said previously: vila, stub killed a stuck ssh process14:11
james_wstub doesn't seem to be around at the moment14:12
vilawhen it occured this morning (i.e. ~6 hours ago)14:12
vilajames_w: if you find him somewhere, tell him that pqm is wedged againg and will be again since I already have another submission in the queue14:13
james_whe may well be done for the day14:13
james_wJohn or Aaron will probably know who else has the power.14:13
viladamn, do *you* someone that can acces pqm ?14:13
vilajames_w: ok14:13
vilaabentley, jam: pqm needs more love ^14:14
abentleyvila: I don't have any particular power.14:14
vilaabentley: I know, but may be you know someone who has ?14:15
abentleyvila: The LOSAs have the power, but they don't have a detailed understanding of PQM.  What's the isue?14:16
vilaA submission blocked pqm this morning and stub killed a stuck ssh process14:17
vilaAn other submission is now blocking it and I have a third one in the queue14:17
abentleyvila: blocking how?14:18
vilaDisplaying: success: merge lp:~vila/bzr/bzr.integration/ http://bazaar-vcs.org/bzr/bzr.dev14:18
vilai.e. the merge seems to have occured but nothing happens14:18
abentleyvila: Has the branch actually been updated?14:19
vilaNot after the submission14:19
vilaif you look at http://pqm.bazaar-vcs.org/ you'll see that I use two different branches, one for each submission14:20
abentleyvila: Okay.  Are you on irc.canonical.com?14:20
abentleyvila: No, I meant has bzr.dev been updated?14:20
vilaabentley: no. I can't14:20
vilaabentley: no, bzr.dev doesn't look updated14:20
vilaChecked. Not updated14:21
vilaThe only cause I can think of is that I use 'lp:' urls where others use 'http:' urls14:22
vilaBut I shoot in the dark...14:22
=== mvo__ is now known as mvo
abentleyvila, meet mthaddon.  He is working on unblocking PQM.14:25
mthaddonabentley, should be okay now14:25
vilaabentley: thanks a ton14:25
vilamthaddon: hi14:26
abentleymthaddon: Thanks.14:26
vilamthaddon: hmm, you're fast, its working again :D14:26
vilamthaddon: but I bet it will block again in about 1 hour14:26
vilathe current submission should be ok, but the next one should trigger the same problem14:27
abentleyvila: Don't be pessimistic.  Maybe it will block sooner :-)14:27
vilaabentley: lol14:27
abentleyAnyhow, it's worth confirming the behavior.14:27
vilaI hate it when I tickle a bug but have no idea where to look to isolate it...14:28
vilamthaddon: what did you do to unblock it ?14:29
abentleyvila: It looks like killing the process prevented your last submission from merging, as we feared.14:29
mthaddonvila, killed the process14:29
vilastrange, this morning killing it make it complete successfully14:29
abentleymthaddon: What version of bzr is running on that box?14:30
mthaddonabentley, we're using bzr 1.6 for that PQM instance14:30
vila1.6 or 1.6.1 ?14:31
abentleymthaddon: There were some performance regressions with
abentleySo this may be related.14:31
mthaddonah, interesting14:32
vilaabentley: and I were lucky this morning because we wait longer before killing it ?14:33
abentleyvila: yes, seems plausible.14:33
abentleymthaddon: Is it possible to update that box to 1.6.1?14:34
mthaddonabentley, lifeless has typically been responsible for that, and last time I did it I broke things, so I'd rather have him to that (as much to approve it as anything else)14:36
mthaddons/him to that/him do that/14:36
abentleymthaddon: Well, it probably blocks merging 'till Sunday, but I guess we can live with it.14:38
mthaddonbzr 1.6 has been on there for a while, I believe... any idea why it's suddenly become a blocker?14:39
vilamthaddon: If I knew I'll be working around it instead of annoying you :-/14:41
vilaMy last submissions have blocked pqm, I have no idea why14:42
vilaMy last submission *before* that problem was  around 2008-09-0114:43
vilaExcept upgrading bzr.dev, I fail to see what I changed14:43
grutte_piervila: remember the discussion about the webdav stuff (when I had nick Jakob, which apparently was registered).....14:47
grutte_pierwell.... webdav is on, but only for one directory where systemadmin has some svn repos14:47
vilaastro.rug.nl ?14:47
vilathat explains it I think14:47
grutte_pierso I have to think of something else :P14:47
vilatry the smart server instead, the performances ought to be better now and far better tomorrow :)14:48
grutte_pierbut the whole thing is that I want others to access the repo, with the smart server they need access to the system right?14:52
grutte_pierthe sysadmin will never setup accounts for people not physically working at the institute, I'm sure14:53
=== mvo__ is now known as mvo
=== abadger1999 is now known as abadger1999_afk
hsn_so its simple do: date(256) + XYZ year16:21
vilagrutte_pier: sry, my window was iconified :-/16:30
vilaIt you setup a smart server, you can then use .htaccess to define your own users/pass (if your server allows that)16:31
ftais there a way to cherry pick dis-joined revisions while merging ?16:31
ftasomething like -r 274..275,283..285 ?16:31
vilamthaddon: the pending submission passed, did you something ?16:33
vilawow, now that's weird, the one you killed have indeed been merged to bzr.dev16:33
vilamthaddon: ok, thanks for confirming and the help, I'll see with jam and lifeless16:33
vilaand will refrain from submitting until then :-D16:34
james_wvila: sorry, I obviously don't know my http error codes :-)16:58
vilajames_w: hehe, don't worry. as mentioned, I had to check too :)17:10
=== mvo__ is now known as mvo
fmhi i cannot get bzr 1.6.1 or 1.7rc1 to work with python 2.6b3. see https://bugzilla.novell.com/show_bug.cgi?id=425644 any idea?18:02
ubottubugzilla.novell.com bug 425644 in Development "bzr init fails: failed to load bzrlib.repofmt.pack_repo.RepositoryFormatKnitPack1: cannot import name U32" [Normal,New]18:03
fmthat is 1.5 the later fail with "TypeError: object.__init__() takes no parameters"18:03
james_whey fm18:26
james_wfm: please run "python -c 'from gzip import U32'"18:26
james_wfm: or python2.6 as necessary18:26
fmjames_w: ImportError: cannot import name U3218:30
fmjames_w: there are the following python packages on opensuse factory: http://pastebin.com/mc80b80818:34
fmjames_w: both my problems seem to be on windows too http://www.python.org/dev/buildbot/community/all/x86 Windows 2003 trunk/builds/66/step-bzr dev/018:38
james_wfm: it looks to me like there is something wrong with python 1.618:48
fmjames_w: how can I debug this?18:51
james_wfm: well, the failure you reported showed that it can be reproduced with just python18:51
james_wso I would find out what changed in python to make that go away, as it probably shouldn't have18:52
fmthe ""TypeError: object.__init__() takes no parameters" seems to be caused by http://bugs.python.org/file2323/new_init_strict.patch18:52
=== kiko-afk is now known as kiko-phone
pcc1how do I merge a merge-directive into a repository without a working tree?  currently getting the error message 'bzr: ERROR: No WorkingTree exists for "file:///..."'18:57
james_wpcc1: you can't, you need a tree to merge it with18:58
=== fm is now known as _fm_
_fm_james_w: u32 was removed http://svn.python.org/view/python/trunk/Lib/gzip.py?rev=61821&r1=61813&r2=6182119:25
james_w_fm_: that sucks, and may not be intentional19:26
james_w_fm_: you should find out if it is and request it to be restored if not19:27
_fm_i am just asking in #python19:28
beunoVerterok, ping19:32
Verterokbeuno: pong19:33
* mvo wonders what he can do about " TooManyConcurrentRequests: The medium '<bzrlib.smart.medium.SmartSSHClientMedium object at 0x893616c>' has reached its concurrent request limit. Be sure to finish_writing and finish_reading on the currently open request."19:50
james_whey mvo19:51
mvohey james_w19:51
james_wthat's normally a secondary error, indicating something else went wrong, and things went from bad to worse as it tried to unwind19:51
james_wwhat operation were you doing?19:51
mvoa commit19:51
james_wI'm not sure what that would be19:52
* mvo tries again19:52
james_ware you using a lightweight checkout?19:52
mvoI don't think so19:52
james_wyou could try "bzr unbind; bzr commit; bzr push; bzr bind; then19:53
AmanicAfwiw, Andrew send in a patch today about: "19:59
AmanicAThis small change fixes one of the remaining causes of TooManyConcurrentRequests19:59
AmanicAerrors.  "bzr pull" and "bzr merge"19:59
AmanicAso maybe this might be fixed in bzr.dev already, or this patch of andrew might help...19:59
AmanicAFor details, see: http://bundlebuggy.aaronbentley.com/project/bzr/request/%3C20080912022408.GA9671%40steerpike.home.puzzling.org%3E20:00
james_wAmanicA: thanks, but I don't think that code path is involved here20:10
AmanicAcool, it was just a stab in the dark..20:10
james_wyeah, I was hoping it was that one :-)20:11
ryanakcaIs there a bzr command to view what would be pulled before pulling it?20:41
Peng_ryanakca: "bzr missing"20:41
Peng_(Well, it'll also show anything that would be pushed if you pushed, but there are arguments to change that.)20:42
ryanakcaPeng_: ok, thanks20:42
=== mw is now known as mw|food
=== jelmer is now known as jelmer_
ryanakcaPeng_: hmm... could you make it show ``bzr diff'' for those files?21:27
Odd_Blokeryanakca: Look at revspecs, there'll be 'branch' or 'ancestor' or something which'll work.21:28
Odd_Blokebzr diff -r branch:<REMOTE> or something.21:28
ryanakcanevermind, I tihnk I found it, bzr diff --new=ARG where ARG is the new branch21:28
=== mw|food is now known as mw
=== jelmer_ is now known as jelmer
_amanica_does anybody know why serverside hooks work with bzr:// but not bzr+http://22:03
KilrooSo does the Bazaar community pretty much have everything -except- a forum?22:03
uwsa forum is for degenerate beings22:04
uwssorry, did i actually say that? ;)22:04
_amanica_whell theres answers..22:04
KilrooPft. I happen to like fora, er, Mathilda.22:05
uwsKilroo: who's Mathilda/ :)22:05
KilrooSorry, I was reading the quotes page. :D22:06
KilrooHm...thanks, _amanica_. That looks somewhat promising.22:06
KilrooFor some reason it still feels more like I'm intruding on people than I usually feel on a forum...22:07
Peng_That's true..bzr doesn't have a forum. Huh.22:07
Peng_I never noticed.22:07
KilrooI don't think I'm actually going to try to post a question until I've actually tried something, though.22:08
KilrooConsidering that what I'm thinking of trying to do is stupid anyhow.22:09
fullermdForum?  Oh, you mean those things like a mailing list with a horrible UI?   :p22:23
lifelessfora, those things that will kill you in australia22:23
=== fta_ is now known as fta
KilrooSee, to my way of thinking, a mailing list is sort of like a forum with a horrible UI.22:41
KilrooSimilar to those forums where you have the sort of tree view of the replies.22:42
Peng_Forums don't let you mark things as read and unread.22:42
KilrooNot at the post level, no. You have a point.22:43
KilrooEither you've seen that thread since anyone posted to it, or you haven't.22:43
Peng_Or your session has expired, so it assumes you've seen the thread.22:44
Peng_(Vanilla averts that, but it's not all that great otherwise.)22:44
KilrooI can't think of a forum that does that.22:44
Peng_phpBB, but maybe it's improved more recently.22:45
Peng_Or, what do you mean?22:45
KilrooFar as I know, on phpBB, if you're not logged in it assumes you haven't read anything except maybe what's in session. If you are logged in it doesn't have to guess.22:45
Peng_Hmm... It seems Loggerhead's pages lack a DOCTYPE or charset declaration.22:47
KilrooMaybe I should give mailing lists a try. I'm sure there must be something out there that lets you read an entire thread of messages from a mailing list at once, in chronological order regardless of reply forking, and paginated. I'd be surprised if it can correct for screwy distributions of > characters in multi-quoted responses from people with different numbers of columns in their mail clients though.22:48
KilrooIt's all a matter of taste anyhow.22:49
fullermdQuite.  I'd hate to read in chronological order.22:49
KilrooIn some cases I like to have the option to toggle it, but I don't like being forced to have things show in tree order, and I absolutely despise having to read one post at a time.22:50
fullermdBut no forum will ever have an interface as customizable as my mail client, and couldn't come within 2 orders of magnitude of the responsiveness.22:51
KilrooTrue enough.22:51
fullermdHeck, I can dispose of a whole mailbox in the time it takes to load one forum thread.22:51
LarstiQKilroo: a problem with a forum for bzr would be that afaik not a lot of knowledgeable people would read it22:52
LarstiQKilroo: I for one wouldn't.22:52
fullermdKKKKK<center><space>jjjiKKKl<enter>  *done*22:52
KilrooIt does begin to sound that way.22:52
LarstiQKilroo: now, you could try gmane22:52
LarstiQKilroo: that may come closer to your tastes?22:53
KilrooNot particularly.22:53
KilrooThat is, it's better than many things I've seen.22:54
KilrooHowever, if gmane suited me, I'd be happy using a mail client.22:54
LarstiQKilroo: I acknowledge there are users (in general, not sure for bzr) who would benefit from a forum22:54
LarstiQbut we'd need to have some people to put effort in it to set it up22:55
LarstiQKilroo: if you're willing to learn about bzr and act as a bridge between the fora and the rest, you're welcome to set one up ;)22:56
KilrooAnd then it would only really be any help for people like me who would like a forum and don't like a mailing list, and even then only as much help as the people who would like a forum enough to post to it...and then the experts either have to spend time on both, or have to choose which one to follow.22:57
LarstiQso I feel starting up a forum none of us will use would be a disservice to users wanting to use it22:57
KilrooLong story short, a forum would probably be more trouble to the community than it's worth, and there just isn't a medium for asking the kinds of questions I have where I don't feel like I'm intruding on people's time.22:58
* LarstiQ nods at Kilroo 22:58
LarstiQKilroo: except, you're free to ask those questions22:58
KilrooI was about to say...why I feel less like I'm intruding by posting a forum thread than by starting a thread on a mailing list or posting a question to the answers thing, I do not know.22:58
fullermdIn high theory, there's no reason a forum and a mailing list can't 2-way gateway each other.  There are a number of impedance mismatches, though, on levels 6-8.22:58
LarstiQKilroo: how about asking on irc?22:59
KilrooI also have never figured out why I feel less like I'm intruding by having an off-topic conversation on IRC than by asking stupid questions.22:59
KilrooI think it's #ubuntu's fault.22:59
* fullermd riffles through the definion of IRC for any mention of "topic"...23:00
Peng_Posting on a mailing list feels more intrusive to me than a forum. I think.23:00
* LarstiQ nods23:00
Peng_I don't start many threads on forums.23:00
LarstiQbut that's the great thing about mailing lists :)23:00
Peng_So I'm not really sure.23:00
LarstiQPeng_: how would you compare the thresholds of irc and fora?23:00
KilrooI guess I think of a mailing list as being sort of like a forum where everyone who's registered is subscribed to the entire thing.23:01
KilrooAnd now I'm trying to remember which forum software it is where you could do that. Invision Power Board I think. Not sure of any others that support that.23:01
LarstiQKilroo: that is a difference. NNTP has a more granular subscription model.23:01
Peng_LarstiQ: I don't know. I usually lurk in IRC channels for 3 months before asking anything. Or I join, ask a question, everybody ignores me, and I leave. :)23:02
LarstiQKilroo: btw, do you actually have a question yet?23:02
KilrooOh yeah, the question.23:02
Peng_Kilroo: What do you mean by "subscription"? Most forums have email alerts and stuff. IPB's may be better though.23:02
Peng_Kilroo: Sorry, you've used up all your time today. Come back and ask tomorrow. ;)23:03
LarstiQPeng_: with a mailing list, you can't selectively watch a particular thread, you get everything23:03
Peng_LarstiQ: Oh, that's ture.23:03
LarstiQlukcily, my email client deals with that just fine, and I usually want to see everything.23:03
Peng_Which client?23:04
KilrooActually I'm probably going to come back next week. See, I have a (very vague) idea of what I want to do, but I haven't tried anything yet to see if it'll work, and I don't like asking "hey, can I do this?" without "Here's what I tried so far..."23:04
fullermdWell, you already asked _half_ the question...23:05
Peng_Kilroo: Well now I'm curious.23:05
Peng_LarstiQ: O23:05
LarstiQKilroo: I'll take your reluctance for that as credit and won't be bothered by you not having tried anything yet :)23:05
* Peng_ can't type right now.23:05
Kilrooas an aside...Peng_: If I'm  remembering right, with IPB you can subscribe to the thread, the subforum, or the entire board. With most other packages I don't think I've ever seen a way to subscribe to anything but the thread.23:06
KilrooOk, you've convinced me. The basic gist is this: I'm working with a team that currently has no source control at all...nor even a real development server. I'm trying to figure out whether there's a way of using bzr with the production server as a dumb server that doesn't require something ridiculously complicated, but yields more benefit than just having everyone keep history on their local working copies.23:08
KilrooAnd I'm having trouble even grasping what I mean by that because the only thing I've ever used before was an old version of Sourcesafe.23:09
LarstiQbzr doesn't require any extra software on machines you publish your changes on, just file (sftp/ftp/etc) access will do23:09
KilrooIt would be relatively simple if not for the fact that I cannot guarantee adoption of bzr across the board23:10
Kilroowhich means that for example people in other departments might be making changes and pushing them directly via FTP.23:10
LarstiQKilroo: would adoption be hindered by wanting another solution, or just general aversion to version control?23:10
Kilroolarstiq: There are three obstacles. One, I'm not even sure how to get the word out to everyone who currently has FTP access. Two, I suspect the powers that be would want to sit on it for a while before making it a standard, even if everyone who would need to switch understood what was happening. Three, we're actually supposed to be moving to Subversion...but for some odd reason doing this is taking months. Lots of them.23:13
LarstiQKilroo: I see.23:13
* fullermd suggests electroshock.23:14
LarstiQfullermd: tsk :)23:14
fullermdWhat?  Hot pokers aren't called for until week 2...23:14
LarstiQKilroo: so multiple people are collaborating on the same source now without any version control, across departements?23:14
KilrooWell, I take that back.23:15
fullermdSee?  Electroshock would be more merciful   :p23:15
KilrooSome of them are using lock files with Dreamweaver.23:15
LarstiQKilroo: aah, web work?23:15
KilrooBut not all of them are, I don't think anyone hesitates to override the lock, and those of us who use Eclipse instead tend not to notice the lock files at all.23:15
Kilroolarstiq, yep...mostly php.23:16
LarstiQKilroo: you don't need to get everyone onboard at the same time23:16
LarstiQKilroo: granted, things will be harder untill you do23:17
* fullermd uses bzr to manage piles of PHP code, and occasionally has to deal with Dreamweaver droppings.23:17
LarstiQKilroo: do you work directly on the server, or is there a syncing step to your workstation before you make a change, make a change, and then upload it again?23:18
KilrooLarstiq: Some people work directly on the server. On occasion, I actually have, myself, for quick fixes. Usually, though, what I do is download the remote files if I think there have been any changes to them, make changes to my local files, and then upload.23:19
LarstiQKilroo: right, so if someone doesn't sync, there is the possibility for overriding others work?23:20
KilrooLarstiq: Yes.23:20
LarstiQI see.23:20
LarstiQKilroo: what you could do, is start locally using it. 'bzr init; bzr add; bzr ci' your local copy, and then use bzr as you would normally. When you sync from remote, make sure you have no uncommitted changes, sync, and do a commit (possibly add first for new files) as bzr ci -m 'changes made by others'23:22
LarstiQKilroo: now, that will only give you versioning locally, and ideally you'd like to collaborate with others23:22
LarstiQbut it at least gives you some of the benefits23:23
LarstiQand no one else has to know that you're using bzr23:23
LarstiQfrom there on, you could expand it to whomever is interested23:24
Kilroothat's pretty much what I had in mind, actually, except I was having trouble wrapping my head around the entire process.23:24
fullermdI think an important step in the process is to never, ever let the idea works its way into your head that a VCS is a deployment mechanism...23:24
LarstiQfullermd: right, but we have `bzr upload` for that23:24
LarstiQKilroo: which is a plugin btw, you can find it on http://bazaar-vcs.org/BzrPlugins23:25
KilrooI've seen it. It confused me a little.23:25
LarstiQKilroo: I need to go to bed now, we can talk about how to use bzr with others another time, or someone else can help you.23:25
fullermdI stick with my belief that it's a bad idea in general, and imposes uncomfortable constraints on you even when it technically works perfectly.23:25
KilrooThanks for your help, LarstiQ23:26
LarstiQKilroo: how did it confuse you?23:26
LarstiQfullermd: you think `bzr upload` is a bad idea?23:26
KilrooBasically, I'm not sure I understand how bzr-upload differs from making a lightweight checkout and putting it on the production server.23:27
LarstiQfullermd: it only works if deployment means having a copy of your tree somewhere23:27
LarstiQKilroo: the difference is in where the updating happens23:27
LarstiQKilroo: with the lightweight checkout, you would need to update it on the server everytime you want it to, well, be up to date.23:27
LarstiQKilroo: and then you'd need to have bzr on the server23:28
fullermdA capability is rarely a bad idea.  Only the use of it   :)23:28
LarstiQKilroo: with bzr-upload, you're just uploading the files that are in your working tree, no bzr required on the server, no cron, no nothing23:28
LarstiQfullermd: don't evade my question :P23:29
KilrooWell, actually what I meant was making a lightweight checkout and then uploading the result. But I see what you mean.23:29
LarstiQKilroo: convenience.23:29
fullermdI don't think I was evading it.  The biggest drawback of having the tool available is that people are more likely to use it, and I think that mechanism is a tool whose value increases the less it's used.23:30
KilrooAnyhow, thanks again for all your help. I need to go eat now. I think I have a good notion of what to do now, though.23:31
LarstiQI see, fair enough.23:31
fullermdIt ties the deployment process into the committing process, and the versioned tree shape to the deployed tree shape, and I think both of those are bad things.23:31
LarstiQfullermd: aah, good arguments.23:31
* Peng_ uses a VCS as a deployment tool. Cough.23:31
KilrooThat makes sense to me, fullermd23:31
fullermdAnd I'm crankier with it than it deserves, because I can't understand how so many people immediately jump to the assumption that it's a good idea.23:32
Kilroobut then, so does being able to deploy FROM version control.23:32
fullermdI mean, I used a VCS for over a year before the thought first occured to me, and after a half hour of setting it up once, I figured out it was a bazooka pointed at my foot   :p23:32
LarstiQfullermd: it's lightweight.23:32
Peng_Yeah, that's why I do it. I don't have any complex needs, and it means I don't have to learn any new tools.23:33
Peng_I only have one server anyway. (Well, two, but they don't share anything.)23:33
LarstiQfullermd: and I've been using it the last couple of days to deploy an oracle install script, for that it worked nicely.23:33
KilrooMy ideal notion is to have VCS, and everyone works with VCS, and then pushing to production consists of taking the current production branch in VCS and putting it on the production server. Hopefully by means of some scripted process.23:34
fullermdMy deployment process is generally a stack of Makefiles.  And my deployed tree shapes _never_ mirror the tree in VCS.23:34
fullermdAt the least, the deployed tree is a subset of what's in the VCS (e.g., things like database manglement scripts, don't go in a live environment).  And there's usually more differences.23:34
* LarstiQ nods at fullermd 23:34
KilrooI'm not sure you can really do that so much in web development.23:35
LarstiQfullermd: you use a heavier deployment process, bzr-upload doesn't add anything there.23:35
KilrooActually I take that back.23:35
* fullermd does web development :p23:35
LarstiQKilroo: zc.buildout!23:35
KilrooBecause I immediately thought of four ways that I could.23:35
fullermd('course, I don't do it, by royal decree, when I don't have real access to the server.  I changed double and a half time any time I have to touch FTP)23:36
LarstiQKilroo: an aside, do you use sftp or ftp?23:36
LarstiQbzr is much happier with sftp23:37
fullermdI guess it boils down to I don't think bzr-upload is evil or bad or broken, I just think it [the general workflow, not the specific tool] is used too often by reflex or whim, rather than really thinking through how the workflow should best be.23:37
LarstiQfullermd: right, I can agree to that.23:37
fullermdOf course, that doesn't put it in rareified company   :p23:37
KilrooLet me rephrase: I don't think it would be possibly for the company where I work to have the deployed tree be a different shape from the development tree without massive changes to the way most things are done.23:38
KilrooGranted, some of those changes would probably be beneficial. I personally don't like having any php file that isn't supposed to be accessed directly by the web live in a folder underneath the document root.23:39
KilrooLarstiq: ftp, unfortunately.23:39
KilrooI'm not sure what that business with the document root has to do with what I was talking about in the previous sentence, either.23:40
KilrooI think I'm too hungry to think.23:40
KilrooI'm going to go eat dinner. Thank you again for all your help.23:40
fullermdFood is not the answer.23:40
fullermdFood is the question.  "Yes" is the answer.23:40
KilrooAmen to that.23:40
LarstiQand I'll detach, ciao23:42
fullermdMwahaha.  And I'll stage a coup while you're all distracted.23:43

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