[00:20] can someone give me a hand with this mess: http://paste.ubuntu.com/19747/ [00:22] do you have a /.bzr dir or something? [00:23] bob2, where? [00:23] in / [00:24] of the branch...yes [00:25] this branch has been fine until tonight [00:25] no, in / [00:25] oh [00:25] nope [00:25] or does ~/Desktop/mythbuntu_branches/mythbuntu-apple-trailers contain etc? [00:26] it used to, I removed it [00:26] did you touch anything in ~/Desktop/mythbuntu_branches/mythbuntu-apple-trailers/.bzr? [00:27] nope [00:37] dunno, sorry - perhaps try the list? [00:46] bob2, I resolved it, I must have broken it somehow, I just pulled the lastest copy of the branch and it works now === mark1 is now known as markh [02:07] I often find myself wanting to switch to the top thread of a loom or to the bottom thread. [02:07] What would be a good command line syntax for this? [02:10] hmm [02:10] switch top: switch bottom: I guess [02:10] if you mean switch and not up-thread [02:13] I do mean switch. [02:26] k === mw is now known as mw|out [02:38] who do I speak to if I want to get my blog syndicated to Planet Bazaar? === thumper_laptop is now known as thumper [02:41] pooolie [02:42] Hey hey. [02:42] How do I exclude a particular directory from a "bzr diff"? [02:42] bzr diff | filterdiff -x dir [02:42] is what comes to mind [02:43] lifeless: Okay. [02:43] I assume that's part of patch-utils. [02:43] diffutils I think, but yeah [02:43] Ah, yeah. [02:45] lifeless: I tried diff-options="--exclude" once, but that didn't work. [02:46] probably need "--exclude old/PATH" [02:46] but we call diff individually, not sure it would honour it in that case [02:47] Ahh, hmm. [02:48] I'll try. [02:51] wow, http://andrew.puzzling.org/ is a little borked. or something [02:51] lifeless: filterdiff doesn't work. [02:52] mkanat: no? [02:52] mkanat: I find I usually need to tweak the pattern a little to make it happy then it plays nice [02:52] bzr diff -rancestor:bzr://bzr.everythingsolved.com/bugzilla/3.2 | filterdiff -x extensions/ > ~/diff [02:53] I tried it without the / too. [02:53] mkanat: also, please do file a bug, -x is reasonable to have [02:53] Does it need a * or something? [02:53] lifeless: Okay. :-) [02:53] mkanat: try new/extensions [02:53] lifeless: Okay. [02:53] (and variations). or old / - see thhe path in the diff header [02:53] === added file 'extensions/launchpad/code/install-before_final_checks.pl' [02:53] --- extensions/launchpad/code/install-before_final_checks.pl 1970-01-01 00:00:00 +0000 [02:53] +++ extensions/launchpad/code/install-before_final_checks.pl 2008-06-12 21:24:07 +0000 [02:53] lifeless: a.p.o -> displaying mashed up text snippets of the entries? [02:53] bob2: yeah [02:54] like apache had a bad day [02:54] spiv: ^ [02:54] mkanat: oh. emm, try "extensions/**" then I guess [02:54] lifeless: Ahh, good idea. :-) [02:54] lifeless: Ahh, yep. :-) [02:55] I did it with just * but that worked. [02:55] some weird firefox/backwards interaction [02:55] bob2: it works with other thing? [02:55] spiv: ping [03:04] reconcile on knits of launchpad -> pain [03:08] Hi all [03:11] hi [03:14] hi lifeless [03:14] i'm going to read "stackable branch ui 2" [03:14] I'm having an issue with bzr where it gives me this error message. bzr: ERROR: Path(s) are not versioned: These are the steps I took to get this error message: [03:14] 1) bzr init [03:14] 2) cd [03:14] 3) bzr branch [03:14] 4) bzr branch mybranch [03:14] 5) cp /files/I/needed/for/mybranch /path/to/mybranch [03:14] 6) cd /path/to/mybranch [03:14] 7) bzr commit /path/to/files/in/mybranch [03:14] whoa lag [03:14] cara: you need to bzr add the new files [03:15] ooo [03:15] cara, 'bzr add' in the root of the branch should be enough [03:15] hehe [03:15] for the files I copied over? [03:15] or do I copy them over that way? [03:15] lifeless: bug 53992 was already there. [03:15] Launchpad bug 53992 in bzr "diff should have an -x option (exclude certain files)" [Wishlist,Confirmed] https://launchpad.net/bugs/53992 [03:15] cool [03:15] after you've copied them in [03:16] do I still have to commit them? [03:18] yes [03:18] add tells bzr to track them [03:18] commit actually takes a snapshot [03:19] oh [03:19] ok [03:19] thanks poolie [03:19] * cara hugs poolie [03:19] you're such a sweetie [03:19] * cara would have been up all night trying to figure out what the heck was going on [03:20] I kinda just jumped into the bzr thing. [03:21] you're welcome [03:24] you just don't know how happy I am right now lol [03:24] i'm really happy to hear that then :) [03:24] please do ask if there's anything else confusing [03:45] oh yeah.. How do I remotely remove a branch I created? [03:46] would I have to literally log onto the server and delete the branch? I actually pushed the branch [03:53] cara: IIRC, bzr remove-tree protocol:/path/to/branch [03:54] mneptok: That's not actually going to remove the branch, though, just the working tree, right? [03:54] RAOF: yeah. [03:54] * mneptok pouts [03:58] * RAOF :P [04:06] can't you just use ssh to remove it ? [04:07] cara: it depends on what access you have to the remote machine [04:13] cara: is this on Launchpad? [04:15] launchpad? [04:15] thumper: I don't know [04:15] cara: where did you push the branch? [04:16] cara: and what command did you use? [04:16] bzr push bzr+ssh http://link.to.branch [04:16] yay tshirt [04:17] err [04:17] bzr push bzr+ssh http://link.to.branch cara or something [04:17] * cara is sleepy [04:17] lifeless: I got one too :) [04:18] cara: perhaps best to come back when not sleepy :) [04:18] cara: or do you really need to get rid of it now? [04:19] nah I will come back later thanks thumper [04:19] hey thumper do you use gentoo? [04:19] cara: nope, kubuntu [04:19] lifeless: yeah, mine arrived yesterday [04:20] hehe I use kubuntu too :) [04:20] so, is this inversely proportional to code written or something [04:20] I thought I seen you in a gentoo channel looooooooooooooooooooooooooooong time ago [04:20] i got mine yesterday too [04:21] been trying to figure out how set the time [04:21] everything is grayed out [04:22] cara: right click on the clock applet and click on "set date and time"? [04:23] cara: s/set/adjust/ === rockstar_ is now known as rockstar [04:29] lifeless: if it were inversely proportional, I would have got mine five years ago. [04:29] :) [04:40] jml: order not dates [04:41] loonch time [04:59] cara: you need to press the 'unlock' button in that window [04:59] it's a bit unobvious [05:00] it's meant to be like the mac idiom of showing things readonly until you authenticate [05:00] which is not a bad idea, but doing it just in one place in confusing [05:00] hth [05:12] got a the tshirt today too :) (evening) [05:15] oh great [05:15] thanks poolie [05:15] beuno: I've pushed the change I was mentioning yesterday [05:15] * poolie takes a deep breath [05:16] ok seriously into your collossal patch now [05:16] sorry :) [05:16] haha [05:16] bob2: 17K lines. [05:16] lifeless, ah, just pulled revision 29 [05:16] si it won't break on re-indexing now? [05:16] right [05:16] * beuno goes look at the code [05:16] should a second or two to index a new commit too [05:18] heh, net negative lines of code, tho [05:18] yah [05:19] lifeless, so. Get the current list of revids, and go one by one through the revids in the branch, and stop when you find a match? [05:20] get the current index (lazy load - no IO at that point), then yes. Oh, I forgot to stop the search, thats why its slow. [05:20] :) [05:20] I was going to ask how you did that part [05:21] $ time bzr index [05:21] real 0m1.257s [05:21] user 0m0.792s [05:21] sys 0m0.184s [05:21] incremental on bzr.dev [05:21] still too slow [05:22] is the ideal for it to be fast enough to run it automatically when the branch is altered? [05:22] I want a tip_change hook for it [05:22] then my branches can be indexed [05:22] spiffy [05:22] and I can forget its installed [05:23] that would be cool [05:23] and fairly easy I suppose [05:23] lsprof-file ftw [05:29] oh well, I guess it can fork() [05:29] 81% in bzrlib.index.GraphIndex.iter_entries() [05:29] I have to fix that, but not today. [05:30] oh heh. that explains why its slow. [05:30] I'm reconciling knit based launchpad in another window, and I forgot. [05:30] lol [05:30] still: just did a pull() on bzr.dev: [05:30] $ time bzr index [05:30] real 0m8.902s [05:30] user 0m6.540s [05:30] sys 0m0.312s [05:31] yeah, 8sec may be a too slow for a handful of revisions [05:31] thats with reconcile in background [05:31] I'm not unhappy :) [05:31] will try later and see [05:35] [05:35] real 2m26.834s [05:35] user 2m19.133s [05:35] sys 0m0.888s [05:35] new index on bzr.dev [05:35] :) [05:35] not bad [05:35] should be subsecond to run again [05:36] oh, let me push [05:36] and [05:36] beuno@beuno-laptop:~/bzr_devel/bzr.dev$ time bzr index [05:36] real 0m5.908s [05:36] user 0m5.756s [05:36] sys 0m0.160s [05:36] after pulling and re-indexing [05:36] ~3 seconds when there is nothing to index [05:36] you want revno 30, pushing up now [05:38] :) [05:38] 0.2sec [05:39] yeah [05:39] so that should be 3 seconds to index new revs above, not 6 [05:40] yeah, shaved off those 3 seconds to compare [05:40] stopping the iteration and only removing NULLs if there actually are any [05:40] this is shaping up nicely! === mwhudson_ is now known as mwhudson [05:40] hello again beuno [05:41] hi again mwhudson [05:41] beuno: is https://code.edge.launchpad.net/~beuno/loggerhead/zpt.cleaner_urls up to date? [05:42] let me check [05:42] mwhudson, yes. It's missing the navigation links fix, which is what I was starting now [05:42] ok [05:43] lifeless: should index lock the branch? [05:44] bob2: it read locks it [05:44] ah [05:44] bob2: oh, during the what-to-index lookup [05:44] beuno: i don't think you need to pass the revno into the revisioninfo template function [05:44] it does [05:45] beuno: doesn't the change parameter have a revno attribute? [05:45] * beuno checks [05:46] mwhudson, do you mean the revid? [05:46] ah uh one moment [05:46] the way loggerhead does urls makes me so sad :( [05:47] good, I like it when we hate the same things [05:47] makes it more probable to get rid of it [05:47] I had to add that to display the revid in the revision view [05:47] because we didn't show it before, which I think is dumb [05:47] isn't there change.revid then? [05:48] yes, that's an improvement [05:48] well, we don't pass change to revid AFAIK [05:48] ah, wait [05:48] maybe we do [05:48] ... of course, after we've got rid of the revision cache we should just be passing bzrlib Revision objects around ... [05:49] yes please :) [05:49] revision cache is my next target [05:49] sweet [05:49] i guess it will be a two phase thing [05:49] delete the cache [05:50] get rid of these bizaare change/entry things [05:50] though we may need to be slightly clever to carry revnos around [05:50] mwhudson, we don't pass change [05:51] -def revisioninfo(url, branch, entry, modified_file_link=None): [05:51] +def revisioninfo(url, branch, entry, revid, modified_file_link=None): [05:51] what is entry here then? [05:51] but well, ok :) [05:52] it doesn't have revid for some reason :/ [05:53] * beuno is confused now [05:53] ok [05:54] I *do* have change [05:55] the -history.get_revno(entry.revid)+entry.revno stuff is kinda funny [05:59] funny as in...? [05:59] it seemed unnecesary to go get that info when we already have it [06:00] funny as in "ha ha how bad is loggerhead" [06:01] heh, yes. There seem to be a few layers of randomness on it [06:01] fixed and pushed my revid nonsense [06:02] and I missed something, revision 245 coming up [06:08] hm, loggerhead is playing with my head [06:13] it does that [06:13] ok, no 245, loggerhead just has the habit of not dying when I tell it to [06:13] so *I* was right, not *it* [06:13] the proper thing to remember is that the answer to the question "is there a reason for it doing this particular thing this odd way" is usually "no" [06:14] heh, yes, I learned that the hard way [06:14] me too :) [06:14] I spent quite a few hours looking at code, thinking I was missing something [06:14] we should add that to the README :p [06:15] we seem to be making great strides on replacing the entire code base [06:15] which should alleviate the need for that :) [06:16] yes, that's a more optimistic way of looking at it [06:16] I'm just mad at it for making me go in circles for 10 minutes [06:16] now, let's get revnos into navigation links working properly... [06:17] beuno: at some point (probably not right now) we should have a think about what we _want_ loggerhead's urls to look like [06:18] mwhudson, absolutely. Have a plan rather than a mashup of things. I'll give it some thought [06:19] any kind of plan, no matter how insane, would probably be progress [06:19] I kinda like the idea of mimicking bzr commands [06:19] like log/revno [06:20] may be more intuitive for everyone, users and developers [06:20] lifeless: cstringio has some kind of limitation wrt binary data but i can't remember precisely what it is [06:20] do you know? [06:20] is it only if you give it unicode objects? [06:21] yes, don't give cStringIO unicode [06:21] very strange things happen if you do [06:21] it really should be 8 bit clean though :) [06:22] poolie: yes [06:22] poolie: StringIO(u'foo').getvalue() returns bytes such that decode('UCS4') works :{ [06:23] (e.g. the native internal representation of u'' strings is coerced into a plain string [06:23] mwhudson: they are not strange, very predictable [06:25] lifeless: well [06:25] i don't think strange and predictable are necessarily in opposition [06:25] :) [06:25] 'suprising' [06:26] sure [06:26] 'not inferrable until seen' [06:26] does 3K fix this [06:26] if not someone needs bitchslapped [06:28] beuno: brain too fuzzed to review the branch properly i think [06:28] i'll get to it over the weekend or monday morning [06:29] mwhudson, sure, no problem. I'll push the navigation changes in a while, so you can do a complete review [06:29] cool [06:30] And wrap up the theme tomorrow. At least the changelog and generic stuff [06:30] still working on the diff view to get the html/css *just* right [06:45] * mwhudson stops for the day [06:45] mwhudson, navigation links fix pushed [06:45] ah, well, have a good weekend then :) [06:45] you too! [06:45] * beuno is one day behind so still has friday :p [06:46] so that means I may still get bzr-search into an experimental branch :) [07:02] that's it for me today too [07:02] g'night [07:05] beuno: gnight [07:17] hey [07:18] hi [07:22] i'm trying to uncommit a revision, but i keep getting the message "Unable to obtain lock file:///var/bzr/blinds.com/.bzr/branch/lock [07:22] held by tsmith@blinds.com on host 5200MHz.xmule.ws [process #11135] [07:22] locked 7 hours, 25 minutes ago [07:22] " [07:22] well, is that process still alive ? [07:22] (if thatis your host, check with ps ax/top) [07:23] i guess i should check both hosts [07:23] 11135 is not an active process [07:23] if its not alive anymore, try bzr break-lock [07:24] is that the same thing w/ CVS locking up and me requiring a sourceforge ticket? [07:25] not really [07:25] cvs is broken-by-design in this area [07:25] ha [07:25] you can break your own lock [07:25] i read a fictional story today about dvcs changing a woman's life [07:25] we can't avoid stale locks completely - network interruptions/power failures etc [07:25] but i really do feel like bzr changed my life [07:26] cool! [07:26] <-- my wifi scrwed up [07:26] when the lock happened [07:26] yah [07:27] (i'm not crazy, but using BZR after 10 years of CVS or SVN (the last year) produces feelings in me akin to when i saw the grand canyon for the first time) [07:27] esp when i bzr switch branches ;o [07:27] :) [07:27] vertigo? [07:27] yeah exactly [07:28] now i'm just wondering if there's a place i can plunk down a few hundred to inspire someone to create bzr up -r functionality [07:29] I believe there is a crude patch [07:29] needs some love [07:29] i specifically need it to update files based on datetime diffs [07:29] like svn up -r {2008-06-10} [07:29] i find bugs all the time ni a codebase of 1.5 million lines [07:29] i have to do a LOT of sleuthing by date [07:30] ah. bisect doesn't help ? [07:30] hmm, this is the sort of case bzr-search _might_ help with [07:30] i do not see that command bzr help commands [07:30] bisect is a plugin [07:31] the company i work for, i'm the lead "fixit" guy [07:31] we have a 4-server dev cycle; code on your box, up to dev server, up to UAT, up to prod [07:31] the problem is, the lead of each site ups all the code to the prod server for htat site [07:31] they do this using svn [07:31] by hand [07:32] like svn diff -r100...101 > changes.diff; patch -p0 < changes.diff [07:32] there's no branching [07:32] there's no tagging [07:32] it's actually pretty sad; for every 3 bugs fixed, 5 more are created, 3 alone by this very merge process; and there's no way to revert [07:33] the main pusher for the main website in March, for instance, accidentally svn up'd in the root of the prod, and we spent 2 months cleaning up the damage [07:33] lol [07:34] ouch [07:34] so i'm in charge of www.blinds.ca [07:34] i am responsidble for all the pushes [07:34] what *I* do is i have a bzr repository w/ branches that mirror my local box, dev, uat, and prod [07:35] and i run rsync against the prod before i do a commit, so i can revert the commit if there are any problems [07:35] cuz every one uses svn and management is very hostile towards positive change (they love negative change, incidentally) [07:36] erg [07:36] you have my sympathy [07:37] nah man bzr saves me lol [07:43] lifeless: real 32m52.967s [07:43] mtaylor: thats significantly faster. woo [07:44] yup. woo indeed! [07:44] mtaylor: also, it will auto-update if you grab the latest code [07:44] the index will? [07:44] good [07:45] lifeless: bzr bisect is exactly what i need; awesome! [07:45] Tsmith: cool [07:46] gosh [07:46] i think some of the people in this channel need to give me thier paypal addresses (lloks @ lifeless) [07:46] Tsmith: lifeless's paypal address is jml@mumak.net [07:47] * jml whistles innocently [07:47] jml: thank you for proxying my due [07:47] (I don't use paypal) [07:47] Tsmith: mine is monty-paypal@inaugust.com [07:47] Tsmith: but I didn't write any of bzr [07:47] Tsmith: seriously though, there's no need - I am partly-paid, partly scratching-my-own-itch for what I do on bzr. [07:48] Tsmith: if we are ever in the same city, all I would say is 'buy me a beer' [07:48] ha what city are you in/ [07:48] Sydney for most of the year [07:48] sorry i dont think that's in my path ;p [07:49] but work has me travel frequently - I'm at UDS for example [07:49] We'll likely be in the US again late this year [07:59] Rebase creates new revisions with different parents and thus the potential for later merge conflicts, right? [07:59] Is `bzr merge [--force] ../other-branch/path/to/file` or `bzr merge -r 526..527 ../other-branch` any different? [08:01] i think rebase basically does bzr merge -r N..N+1 over and over again, doesn't it? [08:01] mwhudson_: {shrug} [08:02] I'm more interested about whether either preserves *any* information in super secret revision properties that might someday in the glorious future be useful to construct better histories / merges. If so, then that would be the technique to prefer, I would have thought. [08:03]