=== danilo-afk is now known as danilos === matsubara-afk is now known as matsubara [11:41] EdwinGrubbs, hi, do you have time to take a look at https://code.edge.launchpad.net/~danilo/launchpad/bug-474874/+merge/14450 === danilos changed the topic of #launchpad-reviews to: on call: EdwinGrubbs || reviewing: - || (r-c candidates skip the queue) queue: [danilo] || This channel is logged: http://irclogs.ubuntu.com === barry` is now known as barry [12:31] noodles775, al-maisan: I assume you are not doing any reviews today? [12:31] danilos: if need be I can review. [12:32] al-maisan, it's very short, just adding some DB permissions for a script: https://code.edge.launchpad.net/~danilo/launchpad/bug-474874/+merge/14450 [12:33] danilos: I'm not no, sorry. [12:33] danilos: OK .. give me 5 minutes and I'll review it. [12:33] al-maisan, excellent, thanks a lot [12:33] you are welcome [12:33] noodles775, no problem, al-maisan is taking care of it :) [12:39] danilos: this is supposed to go into 'devel' .. is that correct? [12:39] al-maisan, yes, but I am actually going to ask for this to be CPed (or perhaps not, GRANTs might be enough) [12:40] al-maisan, note that security.cfg changes do not need to go into db-devel [12:40] danilos: fine, just checking. === mrevell is now known as mrevell-lunch [12:54] danilos: looks good, r=me [12:54] al-maisan, thanks === mrevell-lunch is now known as mrevell === EdwinGrubbs changed the topic of #launchpad-reviews to: on call: - || reviewing: - || (r-c candidates skip the queue) queue: [danilo] || This channel is logged: http://irclogs.ubuntu.com === salgado is now known as salgado-lunch === danilos changed the topic of #launchpad-reviews to: on call: - || reviewing: - || (r-c candidates skip the queue) queue: [] || This channel is logged: http://irclogs.ubuntu.com [15:55] al-maisan: fancy a relatively small review? (or is already time for "Feierabend" for you?) [15:56] adeuring: a small one is OK. [15:57] al-maisan: thanks! https://code.edge.launchpad.net/~adeuring/launchpad/hwdb-udev-submissions-noise-reduction-2/+merge/14479 (130 lines; just two tuples changed) [15:57] beuno: ping [15:58] sinzui, hi [15:58] your suggest for the yes-no widget rocks. it hard to do [15:58] adeuring: OK [15:58] beuno: I'll work on it later today and see if I can make a universal dsolution [15:59] sinzui, because it uses db enums, yadayadayada? [15:59] beuno: no actually not [16:00] beuno: The launchpadformview custom_widget function does not work with macro-style widgets [16:00] beuno: making my widget work or fixing LaunchpadFormView is a big win for reuability [16:01] adeuring: so, this is a 'devel' branch, right? [16:01] al-maisan: yes [16:01] cool. [16:01] beuno: Your suggestion is *much* better that what I did. I spent more than an hour trying to craft a sentence that would work with yes/no. or true/false, or on/off. [16:02] * sinzui thinks all the boolean radio macros suck. [16:03] sinzui, so I don't think yes/no will work well, no matter how we craft the phrases, if this is an urgent branch, I'd be happy to trade an approve now for a fix before week 4 [16:05] beuno: no not urgent. I completely agree with your suggest [16:06] sinzui, so can we still be friends if I push back on this? [16:06] beuno: I think you misunderstand. I have no intention of landing this branch until that widget rocks. [16:06] adeuring: LP is still chewing on lp:~adeuring/launchpad/hwdb-udev-submissions-noise-reduction-2 [16:06] sinzui, ah! great! :) [16:06] I can't even merge it [16:07] al-maisan: did you try to simply branch it? [16:07] yes [16:07] sinzui, other branch approved, it's a great improvement, thank you [16:08] adeuring: this is very funny: [16:08] $ bzr merge lp:~adeuring/launchpad/hwdb-udev-submissions-noise-reduction-2 [16:08] bzr: ERROR: Not a branch: "bzr+ssh://bazaar.launchpad.net/~adeuring/launchpad/". [16:08] al-maisan: odd... Let me paste the diff [16:08] am I missing something? [16:08] adeuring: please do. [16:10] al-maisan: http://paste.ubuntu.com/310697/ === beuno is now known as beuno-lunch [16:10] * al-maisan looks [16:14] adeuring: r=me [16:15] al-maisan: thanks! [16:15] al-maisan: und schönen Feierabend! [16:16] adeuring: in 30 minutes or so :) but thanks anyway! === matsubara is now known as matsubara-lunch === salgado-lunch is now known as salgado [17:02] barry: I'd like your thoughts (consider it somewhere between a late pre-impl and a pre-review :-) ) on this: http://paste.ubuntu.com/310732/ It is supposed to be a fix for https://bugs.edge.launchpad.net/launchpad-foundations/+bug/475550 [17:02] Bug #475550: _pythonpath whitelist of clean_modules is too fragile [17:02] * barry looks [17:02] ty [17:05] barry: The old warning still has value, now that I think of it. Here's a variant: [17:06] If I find any of the namespace packages using the approach in the pastebin, then I go through sys.modules again. [17:07] if anything in sys.modules begins with ``${one_of_the_namespace_packages}.`` (notice the dot) then blow u[ [17:07] up [17:07] complaining, as before, that _pythonpath should be imported earlier [17:09] gary_poster: what's an example of "fewer attributes" (or really different number of them)? [17:10] argh, my second idea doesn't work. OK, here's the idea. [17:11] but I can make it work. [17:12] barry: In this example, ``example`` is a "real" package with only an empty __init__.py [17:12] >>> import example [17:12] >>> dir(example) [17:12] ['__builtins__', '__doc__', '__file__', '__name__', '__path__'] [17:12] >>> import lazr [17:12] >>> dir(lazr) [17:12] ['__doc__', '__name__', '__path__'] [17:12] lazr is one of these namespace things [17:12] so, barry, I think what I can do instead is look for __file__ [17:13] no [17:13] that might not match some C modules, right? [17:13] right [17:13] argh [17:17] barry: But C modules would actually have something exposed *other* than just those three, so the approach as given works. What it *can't* do is actually recognize if there's a problem (e.g., lazr.uri was imported before) and complain, unless I switch to just hardcoding the namespace packages. [17:17] I'd rather not do that. a different kind of fragility [17:18] so inspect.getfile() actually tests for object.__file__ to see if it's a builtin module or not [17:18] right [17:19] btw, I discovered why our deb zope packages don't have this problem: they are installed in debian with the egg-style __init__.py, not this crazy .pth thing. [17:22] maybe... [17:23] 1. look for module.__file__; if it exists it's fine [17:23] 2. if not, look for module.__package__; if it exists it's a builtin; if not it's a pth namespace package [17:23] >>> import _socket [17:24] >>> _socket.__package__ [17:24] [17:24] huh [17:24] >>> for name in dir(_socket): [17:24] ... if name.startswith('_'): [17:24] ... print name [17:24] ... [17:24] __doc__ [17:24] __name__ [17:24] __package__ [17:24] === matsubara-lunch is now known as matsubara [17:25] barry, cool. That might work. To make sure I understand, I am right that the presence of __package__ in this case is kind of random--that is, an implementation detail that might very well change? [17:25] s/I am/am I/ [17:25] gary_poster: i'd have to look at the code to know that ;) [17:26] heh, ok [17:26] gary_poster: oh, let me also check py2.4 [17:26] For me on 2.4: [17:26] >>> import _socket [17:26] >>> _socket.__package__ [17:26] Traceback (most recent call last): [17:26] File "", line 1, in ? [17:26] AttributeError: 'module' object has no attribute '__package__' [17:26] fsk [17:26] otoh, it has a __file__ ;-) [17:27] yeah, so "implementation detail that might very well change" == True === beuno-lunch is now known as beuno [17:27] right [17:27] how strange! why did python 2.6 lose _socket.__file__?! [17:28] seems like that would still be pretty handy, and that _socket.__package__ is useless [17:28] yay for nothing in the NEWS file about this [17:30] barry: gary_poster: I have a small RC candidate for review: https://code.edge.launchpad.net/~sinzui/launchpad/noneable-bug-475433/+merge/14487 [17:32] sinzui: looking... [17:34] sinzui: mp has no diff yet [17:34] oh bugger [17:35] I think we are waiting for both the branch scanner and the mp cron [17:35] barry: do you want a pastebin? [17:35] sinzui: sure [17:36] http://pastebin.ubuntu.com/310757/ [17:37] sinzui: r-me [17:37] er, r=me [17:37] barry: I think this will work. [17:37] (1) if there's a __file__ leave it alone; it's fine. [17:37] (2) if there is a name in the dir that is not __*__ and the full dotted name is not in sys.modules, leave it alone, it's fine. [17:37] (3) Otherwise, we have a namespace package, and we can act on that. Moreover, if there *were* any names in sys.modules, that means that _pythonpath was imported too late. [17:37] barry: thanks [17:38] Going to sketch that [17:39] gary_poster: sounds reasonable. i'm going to grab some lunch now. feel free to paste anything for when i get back [17:39] barry: thank you, will do. [18:12] barry, fwiw: http://paste.ubuntu.com/310788/ [18:12] going to make an MP [18:22] flacoste, barry: https://code.edge.launchpad.net/~gary/launchpad/pythonpath/+merge/14490 [18:26] gary_poster: i'll leave barry do the actual code review [18:26] flacoste: cool [18:33] barry: I'm on this nick for a sec :-) [18:45] flacoste: I have a small RC candidate for review: https://code.edge.launchpad.net/~sinzui/launchpad/noneable-bug-475433/+merge/14487 [18:47] sinzui: approved, add it to CRB [18:47] thanks. [19:13] gary_poster: looking [19:13] barry: thanks. on call, so will have latency [19:18] gary_poster: r=me [19:19] barry, thank you [20:38] rockstar: I can has review? https://code.edge.launchpad.net/~abentley/launchpad/fix-rewrite/+merge/14498 [20:38] abentley, is this for RC? [20:38] rockstar: Yes. [20:39] abentley, I wonder if we should also fix the stdout flushing issue [20:39] abentley, also, r=me [20:39] rockstar: thumper wanted to fix it by invoking the python interpreter with -u, so I've gotten mbarnett to do that. [20:40] abentley, ah, okay. great. === salgado is now known as salgado-afk === matsubara is now known as matsubara-afk