[00:05] maxb: yeah. I'm trying to build it on LUCID [00:05] not "lucid with most of natty backported" [00:06] lamont: you exaggerate madly :-) [00:06] There's only two packages in bzr/builddeps for lucid [00:07] python is not one I'd get to backport [00:08] The modified package is the python metapackage, not the interpreter [00:09] wtf had to change in the metapackage? [00:09] very little [00:09] but specifically what? [00:10] because I can't see any sensible reason why the META package would need to change just so that your package can build [00:10] it strikes me as solving the problem in the completely wrong place [00:10] whatever the problem was [00:10] but I'm open to understanding why that might not be true [00:11] From memory, all I did in the python-defaults source was to bump the version number up to what dh_python2-using packages from sid expect as a builddep, and depended on python-backport-helper [00:11] From our perspective it's exactly the right place to be solving the problem, because it lets us solve it once, rather than once for each bzr plugin package we have in our PPA [00:12] and we have lots [00:12] This way we get to rebuild the Debian sid source unmodified in most cases [00:12] it conflates your packages into the whole system. [00:12] nope [00:13] how does installing the modified metapackage not potentially affect every build on the system? [00:13] It's in a separate PPA used only at build time for exactly that reason [00:13] if I just force the metapackage version down to lucid and add a build-dep on python-backport-helper, should it build? [00:14] yeah, separate ppa doesn't help me at al [00:14] l [00:14] Well, if you insist, it would be better to just make a few small changes so that python-backport-helper isn't needed [00:15] But I do wonder why you can't just use the PPA packages directly, what with PPAs being a controlled source-only build environment [00:15] one of us will get to do that before I can roll out 2.4 internally [00:16] that's a much longer discussion [00:16] Let me twiddle a few things and throw a source package into a different PPA [00:16] ta [00:17] historically, most of the stuff I have to backport fits through my "throw it against the wall and see if it sticks" script quite handily. bzr is the package that demonstrated that I don't want to deal with some backports [00:18] I sometimes suspect that you use the new features of python packaging just because they're new, while certain that I'm wrong in that thought [00:18] Have you got a preferred version string you'd like me to use? [00:18] I'll wind up mucking it about in the end anyway, so it doesn't much matter [00:19] for 2.4.0~beta2-2 it becomes 2.4.0~beta2-2~0.IS.10.04 for lucid (and ...0.IS.8.04 for hardy) [00:19] and yes, that means I throw away the craziness that is ~bazaar1~lucid1 [00:21] well, not craziness per se, but more than what I need to see in a version string [00:35] Uploaded to https://launchpad.net/~maxb/+archive/launchpad [00:35] Pushed to lp:~maxb/ubuntu/lucid/bzr/ppa-canonical [00:37] ta [00:37] I'll play with it after I get back from family time [00:37] getting dragged off now [00:38] gah [00:38] estimated start time "some time tomorrow" [00:38] *shrug* [00:38] You can rescore it if you care :-) [00:39] I'll just build it locally [00:45] it's running tests [00:46] export DEB_BUILD_OPTIONS=nocheck [00:47] Or, you could delete the entire block in debian/rules beginning with [00:47] ifeq (,$(filter nocheck,$(DEB_BUILD_OPTIONS))) [01:15] maxb: still there? [01:26] jelmer: hi [01:28] maxb: I'm considering whether to upload to 2.4b2 to sid. It seems like a good idea as 2.4.0 will be out before wheezy, it gives us fewer releases to worry about and it means we'll have the same version in oneiric and sid [01:29] maxb: what do you think? [01:31] I don't think I'm sufficiently visual on the state of bzr.dev to make a good call, but I'm not entirely convinced it's a good idea [01:32] I've actually dropped back to running 2.3.x locally after bzr.dev was a bit too unstable for my liking over the previous little while [01:37] Is there any particular advantage to having the same version in oneiric and sid *now*? [01:38] they will autosync? [01:39] looks like Jelmer's already uploaded 2.4b2 to oneiric anyway :-) [01:40] hm [01:41] and looks like we're going to be carrying an ubuntu-specific delta for the indefinite future [01:41] + Remove build dependencies not present in main: python-medusa, [01:41] python-subunit, python-meliae, python-lzma. [01:41] + Disable parallel testing. [01:43] yeah, we can't autosync because of those changes. I might have a look at getting subunit into main, but not sure about things like python-medusa, python-meliae [03:09] hi all === _thumper_ is now known as thumper [19:34] quick question: "bzr check" verifies the hashes of all files on the testaments? [20:33] presi: no [20:34] presi: testaments are not stored afaik [20:35] are dynamically generated with bzr testament? [20:35] yes, I think so [20:36] I mean "bzr testament --long" [20:37] yes, that's generated dynamically [20:37] the hash of the testament itself has to be stored [20:37] presi: the hash of the testament isn't stored in the repository, but eg. gpg signatures of revisions sign the testament [20:40] I'm using cat-signature branch to extract the testament signature [20:41] so the signture verify (or not) against dynamically-generated hash of the testament [20:41] that's correct? [20:43] yes - verifying it against a stored testament doesn't seem very useful as then you also need to test the validity of the stored testament [20:46] ok, I have a shell script that selects revisions and pipes cat-signature to gpg --verify if someone is interested I can send or upload it [21:34] hi guys. I want to start using Bazaar but I'm not totally sure how to set it up. [21:35] Do I select start a new project or do I select open a project and just select the folder I want to use === foocraft is now known as ewanas [23:22] jelmer: hi [23:22] jelmer: does File "/usr/lib/pymodules/python2.7/zeitgeist/datamodel.py", line 185, in __getattr__ [23:22] raise AttributeError("'%s' object has no attribute '%s'" %(self.__class__.__name__, name)) [23:22] AttributeError: 'Symbol' object has no attribute 'SOURCECODE' [23:22] ring any bells for you? [23:22] lifeless: hi [23:22] lifeless: vaguely, IIRC the zeitgeist folks changed the constants around a while ago [23:23] we updated bzr-dbus I think - where are you seeing this? [23:23] hi jelmer [23:23] jelmer: I've just pulled that in, will see if it addresses it [23:23] hi lifeless [23:23] g'morning poolie [23:23] jelmer: lp:~lifeless/subunit/bug-654474 [23:24] garr, I wished somebody updated x-chat to support lp: URLs [23:28] jelmer: https://code.launchpad.net/~lifeless/subunit/bug-654474/+merge/59627 [23:28] Ubuntu bug 59627 in linux-source-2.6.15 (Ubuntu) "Intel Core Duo 50% load with powernowd running" [Undecided,Invalid] [23:29] jelmer: if you feel up to a review [23:34] jelmer: and bug 755241 [23:34] Launchpad bug 755241 in subunit "subunit-filter ability to change fail to xfail based on external list" [Medium,In progress] https://launchpad.net/bugs/755241 [23:34] jelmer: are you doing further refactoring there ? or should it be fix committed [23:34] lifeless: Sure, I'd be happy to do a review. It'll probably be tomorrow though, as it's already past midnight here. [23:34] lifeless: I'm going to do more refactoring (hopefully making the stuff in Samba obsolete) [23:35] jelmer: kk [23:35] jelmer: I'd like to do a release of testtools and subunit midweek [23:44] jelmer: I'm going to land the patch; if you do a post landing review I'll act on it. [23:52] lifeless: it looks reasonable at a first glance [23:52] lifeless: wfm, I can do a bit more thorough review tomorrow