[00:00] Mmm. Dig back into the windows (or mac?) binary installer package thingies and see what they had? [00:03] I'm updating the mac installer sources, so the latter is right out [00:03] (already done!) === spm is now known as stevemci === stevemci is now known as spm === spm is now known as stevemci === stevemci is now known as spm [02:05] fullermd: well, I've discovered that the installer fails to bundle testtools ;-) [03:13] Sweet! No test failures! ;p [06:03] fullermd: crashing in the test harness doesn't really count ;-P [06:05] That's the sort of detail that needs to be glossed over if we're going to declare success and go to the circus instead. [06:08] That won't get me a working & tested installer, though, now will it? [06:09] Well, no, but there's cotton candy and a contortionist. [06:09] Which is pretty much the same as dealing with installers, really. [06:11] Eh, it really seems to mostly be a matter of selecting the versions of packages that go in, and configuring and building them ... [06:14] and waiting for the bzr docs to crawl through pdflatex ... [06:16] Ah, I have this dream that someday I'll find a documentation toolchain that isn't a pile of misery. [06:16] I expect it will be delivered by my flying pony any day now. [06:26] it's the page numbers that get you [06:27] Nnooo, I'm pretty sure it's the "whole stupid thing" part :) [06:28] But you never know. There might be some bizarre concatenation of circumstances that causes somebody to actually start maintaining and improving [open]jade some day, for instance. [06:28] I mean, I'm pretty sure that's why pdflatex gets rerun so many times on each document, even when nothing has actually changed [06:28] * SamB_MacG5 was actually wishing someone would do that too [06:30] (maintain jade and friends) [06:30] Or somebody might dream up an XSLT/XSL-FO implementation that is none of "buggy and worthless", "written in java", "expensive proprietary non-cross-platform", or "multiple of the above". [06:30] so libxslt is buggy *and* worthless? [06:31] Well, I presume so just on GP. But it doesn't handle the XSL-FO side anyway. [06:31] oh, right, FO [06:32] * SamB_MacG5 isn't sure what's so great about FO [06:32] I think it stands for "f$*%ing owful". [06:33] But how else do you define output into a PDF? [06:33] I mean, I guess you could write an XSLT sheet that translates into manual Postscript. There are worse fates than having to do that. Though I have a little trouble thinking of them. [06:34] were there even any halfway decent implementations when the spec was frozen, though? [06:35] SamB_MacG5: hmm, testtools for 2.3... It's a shame we didn't track that explicitly, so [06:35] i'd guess 0.9.2 or 0.9.4 [06:36] I freely admit that, as a programming language, TeX seems to rival intercal... [06:36] and I'm not sure it has ever been bunbdled [06:39] Well, that's the flip side; given the work, that gives you great PDF-ish output. But the choices to get HTML suck. [06:39] I _have_ considered things like XSLT/DSSSL/programmatic transforms to turn $random_SGML_application into TeX for the PDF side. But I eventually sobered up. [06:41] vila: well, since I'm building my own installer *anyway*, it will just take a bit more work to add that ... [06:41] I s'pose one plausible conclusion from the state of things would be that nontrivial stuff isn't _meant_ to be documented. [06:41] fullermd: how does that differ from what jade does? [06:42] Largely, that I'd have to write it all, which... uh... [06:42] No. Just, no. [06:43] well, I mean, I suppose the actual "massage into TeX" part is hidden... [06:44] I'm not Don Knuth; I'm not about to put *my* stuff on hole for 30 years, so I can write a peripheral thing I'm not really interested in to support it ;p [06:44] s/hole/hold/ [06:44] I thought it was more like 10 years [06:45] Well, depends what you count as the end. We're still only a third (a fourth?) of the way to having Vol IV out... [06:46] fullermd: just be patient... [06:46] vila: Yeah, yeah, patience; how long will that take? [06:46] but weren't volumes A through E mostly finalized in a decade? [06:48] granted, only fools use Volume E directly rather than using the Type 1 conversion ... [06:49] SamB_MacG5: from some random notes: 0.9.8 should be ok with 2.4, so that should give you an upper bound [06:49] fullermd: not much longer ;) [06:50] vila: I was just going to try 0.9.15, since that's the final release supporting Python 2.4 and 2.5 [06:51] SamB_MacG5: hmm, let say the relationship between bzr and testtools has been a bumpy road so, 0.9.15 *may* include all the needed fixes, otherwise *some* precise older version will be required :) [06:52] SamB_MacG5: thanks for working on this though [06:53] SamB_MacG5: and don't forget to make a merge proposal against lp:bzr-mac-installers/2.3 when you succeed [06:53] vila: of course [06:53] cool, even more thanks then ;) [07:01] * SamB_MacG5 -> bed [08:02] morning! [08:12] morning mgz === dpb_ is now known as Guest14449 === joey[a] is now known as joey === idnaria is now known as idnar === ScottK2 is now known as ScottK === hazmat is now known as 45PAA1GWO === wgrant_ is now known as wgrant [14:56] Heyyy! anyone know if there are documented issues with Bzr interacting with Git? [14:56] could someone tell me aboot dem =P === mmrazik is now known as mmrazik|afk [14:57] delinquentme: how do you mean? [14:57] jelmer_, does anyone have both running on their system happily [14:57] delinquentme: sure [14:57] I thought you meant "is bzr-git perfect" [14:58] delinquentme: sure, they can be installed together without problems [14:58] ok cool [14:58] yeah im usually a git kid but I think I might be doing some VC w bazaar [14:58] any recommended " go to " tutorials for the newbie? [14:59] delinquentme: http://doc.bazaar.canonical.com/latest/en/mini-tutorial/ [15:00] darnit, I should have thought things through before running ./build.py ... [15:00] now I'll have to wait for pdflatex :-( [15:00] ... or I could reorder the packages in config.py ... [15:03] * SamB_Mac_ puts bzr last [15:03] * SamB_Mac_ waits for pyqt to fail to build === SamB_Mac_ is now known as SamB_MacG5 [15:06] O_o;;; [15:06] im on ubuntu .. and it doesnt come preinstalled? [15:08] delinquentme: apt-get install bzr [15:08] yeah sudo apt-get install bzr [15:13] delinquentme: why would it come preinstalled? [15:14] I mean, unless you chose a "install a bunch of development tools and such" option during installation [15:14] OMggg it was sooo easy!!!!! [15:15] GNU/Linux: Bazaar is probably in your GNU/Linux distribution already. [15:15] now a more complex question [15:15] how to add / push to a remote repo via ssh [15:15] i guess i can google dis. [15:17] delinquentme: "bzr help urlspec" may be useful [15:43] some kind of log command? I added files to the repo ... and now I'd like to confirm that I only added the one file. [15:44] @_@ [15:44] http://stackoverflow.com/questions/3682817/can-i-edit-the-message-of-an-older-revision-in-bazaar [15:45] I cant change the commit message without crapping on the repo? [15:46] delinquentme: you might find it useful to follow [15:46] yeah I've actually got it open :D [15:49] on changing commit messages, the important thing to understand is that dvcses let you edit history, but it's socially discouraged on public branches, because everyone needs to agree on the history to share changes [15:49] so, if you've just committed something and made a tyop, you just uncommit, then commit again with the fix [15:51] but rather than fixing an error from several months ago by changing all the history since then, you should generally just accept that history is history (though there are ways to do it with everyone's cooperation if it really matters) [15:58] hi folks, any possibility of getting a quick fix to the importer failure for gtimelog? http://package-import.ubuntu.com/status/gtimelog.html#2012-07-04 19:41:22.610986 [16:00] jelmer_, jam, james_w`, vila, mgz, or maxb (not to leave out any of the usual suspects :) ^^ [16:00] oh, it's one of those [16:00] ;) [16:00] let's see what it looks like when branched locally [16:01] maxb is quicker on the draw :) [16:01] * maxb glares at someone who has introduced a bad launchpadlib version check in lp:udd [16:01] ImportError: Version of launchpadlib 1.10.2 less than 1.9.0 [16:02] oh no it isn't :-) [16:02] hm, that was probably me. [16:02] being lazy and not wanting to split and intify [16:02] I'll fix. [16:07] a remote repository is called what? [16:07] barry: So this looks pretty easy to fix, in terms of me knowing exactly what data needs to go into udd's sqlite db to make it accept Martin's revisions and move on, but I don't have a neat tool to do that, so I'll need to construct some manual SQL (not that much, 1 line for natty, oneiric, precise, quantal each) - probably take me half an hour to evade other things I'm busy with and get to it, though [16:08] maxb: thanks, that's fine. it's lunch time here in us/eastern. i *really* appreciate the fix! [16:08] maxb: that sounds painful, is this not something we can hack into a script? [16:09] sure, but not I'm not good enough with bzrlib to do it before I go out for the evening today [16:10] maxb: obviously do what ever is easiest for you, I'm just aware of the fact I'd be stuck if you weren't handling it so some record of what you needed to do would be grand [16:11] mgz: One-line summary - grab the head revision from each series branch, and insert its tag, revid and testament into the revids table === deryck is now known as deryck[lunch] [16:27] maxb: [16:30] mgz: It looks good to me, other than I have no clue what the behaviour is when you compare a string with an int [16:30] I'd have to test that out before I could offer an opinion [16:30] safe but not useful in python 2 [16:30] and udd won't be python3ing any time soon. [16:31] I think any str is less than any int in Python 2, which is actually about what we want here [16:32] so 1.0.0-beta < 1.0.0 if launchpadlib were insane enough to do that. [16:32] (as "0-beta" < 0) [16:33] heh [16:33] In that case, MP gets my vote, though I'm not in range of a decent browser to enter it in at the moment [16:34] why am I not surprised that easy_install is responsible for breaking my build? [16:34] ...didn't you approve an mp from a bus in the past? have you lost your fancy phone? :P [16:35] SamB_MacG5: that'll learn ya [16:35] * SamB_MacG5 really wishes there was a Python option to trace all sys.path changes [16:37] anyway, I commented out the bogus .pth line and it seems to be building now ... so maybe I'll be able to test an installer then ;-) [16:38] (It was sticking /Library/Python/2.5/site-packages before the stuff from PYTHONPATH) [16:38] you should perhaps wish for the death of .pth files instead [16:39] this one was easy_install's creation, though! [17:28] barry: Import completed. [17:28] maxb: thanks! === deryck[lunch] is now known as deryck === 45PAA1GWO is now known as hazmat === ajmitch_ is now known as ajmitch [21:06] * SamB_MacG5 wonders why the launchpadlib build didn't return nonzero, since it clearly failed ... [21:16] hmm [21:32] huh, this installer project seems to have stale version numbers in it ... === jordan_ is now known as jordan [22:01] * SamB_MacG5 grumbles about there not being a standard distutils target for running tests