[02:18:01] ** bitmonk has left IRC () [03:06:46] ** bitmonk has joined us [03:09:45] ** bitmonk has left IRC (Client Quit) [03:30:17] ** bitmonk has joined us [04:02:52] [connected at Tue Jun 6 04:02:52 2006] [04:02:53] <> *** Looking up your hostname... [04:02:53] <> *** Checking ident [04:02:53] <> *** Found your hostname [04:03:24] <> *** No identd (auth) response [04:03:24] <> *** Your host is kornbluth.freenode.net[freenode.freenode.net/6667], running version hyperion-1.0.2 [04:03:24] [I have joined #peak] [11:08:32] ** erikrose has joined us [11:53:35] ** whit has joined us [11:56:58] ** whit has left us [12:19:27] ** pje has joined us [12:24:02] ** whit has joined us [12:24:32] ** whit has left IRC (Client Quit) [12:34:39] ** whit has joined us [12:38:41] ** nouri has joined us [12:38:55] pje: ayt? [12:39:10] yes? [12:39:12] oh well, IRC says you are :) [12:39:15] oh here you are [12:39:20] i got an annoying question for you [12:39:39] setuptools is not including my __init__.py's in the distribution [12:39:45] any idea? [12:39:55] I assume you have the packages listed? [12:40:05] packages=find_packages(exclude=['ez_setup']), [12:40:07] this is what i have [12:40:12] ** whit has left IRC () [12:40:14] let me show you the source of that for a moment [12:40:14] And that you're not asking about a bdist_rpm or other non-egg bdist format [12:40:22] and you're not using namespace packages [12:40:27] it's even in SOURCES.txt ! [12:40:31] i am using namespace [12:40:48] namespace package __init__.py are not included in bdist_rpm, bdist_wininst, etc. [12:41:07] because they would cause file conflicts [12:41:15] i.e., two RPMs can't install the same file [12:41:29] http://svn.plone.org/svn/collective/ZopeSkel/trunk/setup.py [12:41:37] i'm not doing any fancy distributions [12:41:40] just easy_install [12:42:24] SOURCES.txt has the __init__.py s, but when i install it, it doesn't include them [12:42:31] both in the namespace and in the subnamespace package [12:42:46] (i didn't intend to say easy_install is not fancy ;) [12:42:55] you [12:43:04] 're not doing namespace packages either, at least not in that setup.py [12:43:10] oh no sorry [12:43:11] i got confused [12:43:16] this is not namespace package [12:43:29] it contains a skeleton inside templates/ which has __init__.py's in it [12:43:43] but these it doesn't like to include [12:43:50] these are like package_data to me [12:43:56] and they are in SVN [12:44:12] i think that's why they correctly land in SOURCES.txt the moment i do setup.py egg_info [12:44:17] ah [12:44:28] I see [12:44:55] I think you may need to explicitly list them in setup or MANIFEST.in, because ISTR that include_package_data ignores .py files [12:45:08] oh [12:45:35] i thought if it's in SOURCES.txt it should be in the dist? [12:45:46] it will be, in the source dist. [12:45:54] Not in the binary, though. [12:46:05] hmm [12:46:09] * nouri thinks hard [12:46:31] OK, i'll just list them manually then [12:46:43] i guess there's no more elegant way to do what i want to do? [12:47:43] I'm looking at the code now [12:47:56] I suspect it may not work *even if* you list them manually. :( [12:48:19] well, i think PasteScript should have the same problem that I have [12:48:26] looking into that [12:48:44] Okay, you should be able to use a pattern like package_data = {'': 'templates/*.py'} [12:48:54] sounds good [12:48:54] it's only include_package_data that's skipping all .py files. [12:49:05] it won't be recursive, though. [12:49:35] ok, package_data can add to include_package_data, right? [12:49:44] i'll look into the docs [12:49:58] thanks much for your attention [12:50:31] yes, it can [12:50:36] cool [12:56:21] pje: mind explaining why it's excluding py files? so i can put it as a comment in there [12:56:45] it's to avoid duplication with the code that handles modules [12:57:21] I can probably get rid of it, but there's another problem, which is that it's probably going to create .pyc or .pyo files too. [12:57:49] i see [12:58:21] Somebody already complained about that one; I suggested using a different extension like .py.tmpl or something. [12:58:42] right, actually there's nothing wrong with using that in my case [12:58:53] _tmpl which pastescript takes care of [12:58:56] Or putting the templates in the package metadata (under the .egg-info) instead [12:59:12] everything under there is treated as pure data. [12:59:24] how do i access it then? [12:59:32] i mean where does it live when installed [12:59:38] the metadata APIs - dist.get_metadata('path') [12:59:48] cool [12:59:52] where dist can be obtained via get_provider(__name__) [13:00:12] I'm starting to wonder if I shouldn't allow full string/stream/filename access to metadata. [13:00:47] I'm actually talking on another IRC channel right now with somebody who could use that, for a different application. :) [13:00:55] heh [13:01:07] well, your proposed solutions are certainly good enough for me right now [13:01:11] Can't really go in the 0.6 series, though, that'd be a new feature and I'm trying to finalize the 0.6 version. [13:01:16] right [13:02:55] you're doing a great job, keep it up! :) [13:44:55] ** whit has joined us [14:27:32] ** whit has left IRC () [14:32:58] ** whit has joined us [15:27:14] ** bitmonk has left IRC () [15:32:12] ** bitmvnch has joined us [16:07:38] ** bitmvnch has left IRC () [16:17:17] pje is now known as pje_mtg [16:27:47] erikrose is now known as ErikRose [16:27:47] ErikRose is now known as erikrose [16:59:53] pje_mtg is now known as pje [17:12:21] ** whit has left IRC () [17:12:49] ** erikrose has left IRC (Remote closed the connection) [17:13:29] ** ErikRose has joined us [17:19:23] ** ErikRose has left IRC (Remote closed the connection) [17:19:58] ** ErikRose has joined us [17:35:06] ** coderanger_ has joined us [17:35:32] Anyone on here who could tell me if I have the wrong idea as to how extras are supposed to work? [17:36:16] I had thought that the idea was you could mark an entry point as needed an extra, and then it wouldn't load that entry point unless that extra's deps were met [17:38:00] But when I try running setup.py develop with an entry point of the form 'wikirename.test = wikirename.test [extra]' I get errors [17:41:49] it's the other way around. [17:42:06] if you put an extra on an entry point, it means that when you load the entry point, it will require() the extra [17:42:37] hmm [17:42:37] ** ErikRose has left IRC (Remote closed the connection) [17:42:37] if you want to check an entry point for loadability, call its require() method and trap resolution errors. [17:43:14] ** ErikRose has joined us [17:43:40] i.e. 'except pkg_resources.ResolutionError:' [17:43:43] Yeah, I've got that all setup, it checks each extra for loadability (using working_set.resolve), and then only loads entry points with extras that it could find [17:44:02] okay, so I'm confused then. What was the question? [17:44:44] I have an entry point, 'wikirename.foobar = wikirename.foobar' [17:44:46] That works fine [17:45:03] If I make it 'wikirename.foobar = wikirename.foobar [blah]' it fails [17:45:09] define "fails" [17:45:21] with "Invalid module name 'wikirename.foobar '" [17:45:28] Ah! [17:45:30] okay [17:45:44] known bug, fixed in 0.6b1 if I recall correctly [17:45:52] it's a bug in the entry point parser code [17:46:00] should also be fixed in the 0.7 trunk [17:46:05] Ahh, okay, is that a new release? [17:46:14] 0.6b2 is the latest release on PyPI [17:46:14] I have 0.6a11 on here [17:46:23] 0.6b1 went out in mid-May [17:46:32] okay, I'll upgrade and try again [17:46:43] sorry about that, I wasn't clear on what kind of error you were having [17:46:58] yeah, neither was I :) [17:55:21] pje: Yep, works now [17:55:22] Thanky [17:56:16] np [18:00:00] If an egg is built on a machine with >0.6b1, and then used on a machine with a lower version, will this be an issue? [18:00:13] hrm. [18:00:15] Maybe. :) [18:00:41] If it is, make your egg require setuptools>=0.6b1 :) [18:01:56] ** bitmvnch has joined us [18:50:27] ** ErikRose has left IRC () [19:02:36] pje is now known as pje_afk [19:12:54] pje_afk is now known as pje [19:23:26] ** coderanger_ has left IRC ("Download Gaim: http://gaim.sourceforge.net/") [19:26:09] ** bitmvnch has left IRC () [20:08:26] ** pje has left IRC ("Client exiting") [20:35:03] ** bitmvnch has joined us [21:55:58] ** bitmvnch has left IRC () [22:24:02] ** whit has joined us [23:05:08] ** whit has left IRC () [23:17:47] ** bitmvnch has joined us [23:18:10] ** bitmvnch has left IRC (Remote closed the connection)