| *** nitrogenycs has left #zope | 00:03 | |
| *** mcdonc_ has quit IRC | 00:06 | |
| *** mcdonc has joined #zope | 00:07 | |
| *** fredvd has joined #zope | 00:16 | |
| *** avoinea has left #zope | 00:18 | |
| *** alexpilz has quit IRC | 00:34 | |
| *** alga has quit IRC | 00:38 | |
| *** digitalmortician has joined #zope | 00:39 | |
| *** alexpilz has joined #zope | 00:40 | |
| *** waldi has quit IRC | 00:52 | |
| *** waldi has joined #zope | 00:52 | |
| *** runyaga has joined #zope | 00:59 | |
| *** Theuni2 has joined #zope | 01:01 | |
| *** TresEquis has quit IRC | 01:16 | |
| *** fredvd has quit IRC | 01:48 | |
| *** supton has quit IRC | 01:57 | |
| *** supton has joined #zope | 01:58 | |
| *** planetzopebot has quit IRC | 01:59 | |
| *** Theuni2 has quit IRC | 02:01 | |
| *** MrTango has quit IRC | 02:06 | |
| *** webmaven has quit IRC | 02:07 | |
| *** J1m has quit IRC | 02:29 | |
| *** webmaven has joined #zope | 02:41 | |
| kosh | betabug: you awake? | 02:47 |
|---|---|---|
| *** mr_jolly has quit IRC | 03:24 | |
| *** tiwula has quit IRC | 03:51 | |
| *** srichter has quit IRC | 03:57 | |
| *** planetzopebot has joined #zope | 04:09 | |
| *** tiwula has joined #zope | 04:34 | |
| *** metalfan_ has quit IRC | 04:39 | |
| *** jham has quit IRC | 05:01 | |
| *** _mup_ has quit IRC | 05:01 | |
| *** bowe has quit IRC | 05:02 | |
| *** kiorky has quit IRC | 05:02 | |
| *** rodgort` has quit IRC | 05:02 | |
| *** bowe has joined #zope | 05:03 | |
| *** kiorky has joined #zope | 05:03 | |
| *** rodgort` has joined #zope | 05:03 | |
| *** rodgort` has quit IRC | 05:03 | |
| *** jham has joined #zope | 05:03 | |
| *** _mup_ has joined #zope | 05:03 | |
| *** webmaven has quit IRC | 05:04 | |
| *** rodgort has joined #zope | 05:09 | |
| *** daMaestro has quit IRC | 05:38 | |
| *** gqlewis has quit IRC | 05:52 | |
| *** Arfrever has quit IRC | 06:05 | |
| *** planetzopebot has quit IRC | 06:05 | |
| *** planetzopebot has joined #zope | 06:09 | |
| *** slackrunner has joined #zope | 06:15 | |
| *** redir has quit IRC | 06:32 | |
| *** redir has joined #zope | 06:33 | |
| *** altipard has joined #zope | 06:43 | |
| *** altipard has left #zope | 06:43 | |
| *** sm has quit IRC | 06:45 | |
| *** sm has joined #zope | 06:53 | |
| *** slackrunner has quit IRC | 07:08 | |
| *** dayne has joined #zope | 07:19 | |
| *** sm has quit IRC | 07:27 | |
| *** hever has joined #zope | 07:41 | |
| *** hever has quit IRC | 07:44 | |
| *** hever has joined #zope | 07:45 | |
| *** slackrunner has joined #zope | 07:51 | |
| *** hever has quit IRC | 07:53 | |
| *** tiwula has quit IRC | 08:07 | |
| *** Theuni2 has joined #zope | 08:19 | |
| *** alga has joined #zope | 08:36 | |
| *** alga has quit IRC | 08:43 | |
| *** sm has joined #zope | 08:52 | |
| *** __mac__ has joined #zope | 08:58 | |
| *** wosc has joined #zope | 09:01 | |
| *** altipard has joined #zope | 09:08 | |
| *** zagy has joined #zope | 09:09 | |
| *** altipard has left #zope | 09:10 | |
| *** zagy has quit IRC | 09:15 | |
| *** zagy has joined #zope | 09:15 | |
| *** altipard has joined #zope | 09:16 | |
| *** digitalmortician has quit IRC | 09:17 | |
| *** agroszer has joined #zope | 09:22 | |
| *** Theuni2 has quit IRC | 09:34 | |
| *** hever has joined #zope | 09:39 | |
| *** hever has quit IRC | 09:41 | |
| *** hever has joined #zope | 09:41 | |
| *** ccomb has joined #zope | 09:56 | |
| *** Wu has joined #zope | 09:56 | |
| *** sm has quit IRC | 09:59 | |
| betabug | kosh: now I am :-) | 10:00 |
| *** Theuni2 has joined #zope | 10:00 | |
| kosh | betabug: ah cool I have a new product | 10:02 |
| *** alga has joined #zope | 10:02 | |
| betabug | ha! fun :-) | 10:02 |
| kosh | https://github.com/Immudzen/CacheVersionURL | 10:02 |
| *** yvl has quit IRC | 10:02 | |
| *** alexpilz has left #zope | 10:02 | |
| kosh | very very simple it generates unique urls with VERY long cache times for any object you hand it | 10:02 |
| betabug | hmmm, what? | 10:03 |
| betabug | ah, so the URL changes if the object changes? | 10:03 |
| kosh | so lets say you have an object called css you would create the url to it with thiscacheobject.version_absolute_url_path(path.to.your.css) and it will generate a url like thiscacheobject/ver_timestamp/path/to/your/css | 10:03 |
| kosh | yeah it is based off of bobobase_modification_time | 10:03 |
| kosh | it can't detect changes to other objects that would cause the first one to change but that was not a problem I had to deal with for my usage at least | 10:04 |
| betabug | so you set the cache time really high, but when the object changes you still get the new version right away | 10:04 |
| kosh | but this makes is so that my js files, css files, images files etc all can have something like a 30 year cache time but still refresh instantly when changed | 10:04 |
| betabug | yeah, got it now | 10:05 |
| kosh | I just built this functionality into my regular product but I made this as a seperate product to work with any zope object | 10:05 |
| betabug | nice stuff | 10:05 |
| betabug | yeah, that's a good idea | 10:05 |
| kosh | one a test site in yslow I reduced the lookups needed to draw a page from 85 to 5 | 10:05 |
| kosh | it was very easy to do, I still need to do some more testing with it for virtual hosts etc but the product is tiny | 10:06 |
| kosh | all the work is in 2 methods | 10:06 |
| *** yvl has joined #zope | 10:06 | |
| *** digitalmortician has joined #zope | 10:07 | |
| betabug | yeah, I can see that... d/l now | 10:07 |
| *** planetzopebot has quit IRC | 10:08 | |
| betabug | du -sh * 0BCacheVersionURL.py | 10:09 |
| betabug | 4.0KREADME 0B__init__.py | 10:09 |
| *** planetzopebot has joined #zope | 10:09 | |
| betabug | something is wrong with that download | 10:09 |
| kosh | hmm let me check | 10:09 |
| betabug | files are empty | 10:09 |
| kosh | heh oops I need to try to commit that again | 10:09 |
| kosh | so hard to remember the git commands | 10:10 |
| kosh | now it is all there | 10:11 |
| betabug | I don't have git on this machine... I thought I had, but it's not here | 10:11 |
| betabug | ok, let me d/l again | 10:11 |
| betabug | kosh: yeah, now it's there | 10:12 |
| kosh | I added those files from the cli, normally I have kdevelop do it for me :) | 10:13 |
| betabug | kosh: looks very nice | 10:17 |
| betabug | only thing I saw is the fixed date of 2030.... you'll have to update that in 10 years or so :-) | 10:17 |
| kosh | yeah I should fix that at some point | 10:18 |
| kosh | just not high priority :) | 10:18 |
| betabug | and people will complain how your product is "unmaintained" HAHAHAHAHAHA | 10:18 |
| kosh | I figure by 2030 nobody will be using that anymore and someone can fix it | 10:18 |
| betabug | don't count on it | 10:18 |
| betabug | so much old software still runs after so many years | 10:18 |
| kosh | yeah I should fix it to just do the current year + 20 or something like that | 10:19 |
| betabug | no, leave it like that - it's one less computation | 10:20 |
| kosh | there is that, right now it does very little in the way of computation | 10:21 |
| kosh | but that means I only have about 18 years to update it ;) | 10:22 |
| betabug | you should make a project plan or something, with a gant chart and a timeline | 10:23 |
| kosh | hehe | 10:23 |
| kosh | it just seems like such a simple way to solve a problem | 10:23 |
| kosh | unique urls for the resources so they cache and expire nicely | 10:24 |
| betabug | yeah, it's been done before | 10:24 |
| betabug | but not as simple I guess | 10:24 |
| kosh | ah I had seen a suggestion by google to do something like that and they specifically said not to do stuff like foo/bar?ver=1231231 | 10:24 |
| kosh | since some browers under some situations won't cache that | 10:25 |
| kosh | so I figured I could just put the version number into the url easily and just use bobo_traverse to strip it | 10:25 |
| betabug | yes, it's not elegant | 10:25 |
| kosh | and since the time does not matter the bobo traverse to strip that element is very simple | 10:25 |
| kosh | so I don't have to actually verify the number | 10:26 |
| kosh | so far I have not thought of any problems with this but maybe you could think of some issues | 10:28 |
| CIA-83 | icemac * r121307 zc.recipe.macro/ (5 files in 2 dirs): Conform to repository policy. | 10:29 |
| CIA-83 | icemac * r121308 zc.recipe.rhrc/ (setup.py README.txt): Added not declared, but needed test dependency on `zope.testing`. | 10:29 |
| CIA-83 | icemac * r121309 zc.recipe.rhrc/ (4 files in 2 dirs): Conform to repository policy. | 10:29 |
| CIA-83 | icemac * r121310 zc.recipe.wrapper/ (LICENSE.txt COPYRIGHT.txt setup.py): Conform to repository policy. | 10:29 |
| CIA-83 | icemac * r121311 zc.recipe.testrunner/ (4 files in 2 dirs): Conform to repository policy. | 10:29 |
| CIA-83 | icemac 1.0 * r121312 zc.recipe.testrunner/ (5 files in 2 dirs): Conform to repository policy. | 10:29 |
| CIA-83 | adamg * r121313 keas.build/ (src/keas/build/build.py CHANGES.txt src/keas/build/base.py): Fix RawConfigParser usage, it destroys options values by making them all lowercase. | 10:29 |
| CIA-83 | adamg * r121314 keas.build/ (CHANGES.txt setup.py): Preparing release 0.2.1 | 10:29 |
| CIA-83 | adamg * r121315 /keas.build/tags/0.2.1: Tagging 0.2.1 | 10:29 |
| CIA-83 | adamg * r121316 keas.build/ (CHANGES.txt setup.py): Back to development: 0.2.2 | 10:29 |
| *** mr_jolly has joined #zope | 10:38 | |
| *** mr_jolly has left #zope | 10:39 | |
| betabug | I think it's fine | 10:39 |
| kosh | cool | 10:40 |
| *** goschtl has joined #zope | 10:42 | |
| *** yvl has quit IRC | 10:47 | |
| *** alexpilz has joined #zope | 10:50 | |
| *** MrTango has joined #zope | 10:53 | |
| *** avoinea has joined #zope | 10:53 | |
| *** avoinea has quit IRC | 10:54 | |
| *** avoinea has joined #zope | 10:54 | |
| *** fredvd has joined #zope | 10:57 | |
| *** Wu has quit IRC | 11:05 | |
| *** vortec has joined #zope | 11:06 | |
| *** Wu has joined #zope | 11:06 | |
| *** eperez has joined #zope | 11:07 | |
| *** mitchell`off is now known as mitchell` | 11:08 | |
| *** gwik has joined #zope | 11:23 | |
| CIA-83 | ctheune python-3-2 * r121317 zc.buildout/ (5 files in 3 dirs): Snapshot: make 'pythonX.Y dev.py' work for 2.4, 2.5, 2.6, 2.7 and 3.2 | 11:29 |
| CIA-83 | ctheune python-3-2 * r121318 zc.buildout/src/zc/buildout/testing.py: Fix up test initialisation for Python 3.2 on linux. | 11:29 |
| CIA-83 | ctheune python-3-2 * r121319 zc.buildout/src/zc/buildout/testing.py: Normalize output for easy_install.txt to match shared library files on linux. | 11:29 |
| CIA-83 | ctheune python-3-2 * r121320 zc.buildout/src/zc/buildout/download.py: More API cover-ups for URLlib. | 11:29 |
| *** menesis has joined #zope | 11:37 | |
| *** evilbungle has joined #zope | 11:38 | |
| *** alexpilz1 has joined #zope | 11:42 | |
| *** alexpilz has quit IRC | 11:44 | |
| *** fredvd has quit IRC | 11:47 | |
| *** nlzh has joined #zope | 11:51 | |
| *** yvl has joined #zope | 12:01 | |
| *** J1m has joined #zope | 12:06 | |
| *** hever has quit IRC | 12:13 | |
| *** hever has joined #zope | 12:14 | |
| *** vortec has quit IRC | 12:27 | |
| *** vortec has joined #zope | 12:28 | |
| * Theuni2 waves towards J1m | 12:29 | |
| agroszer | hey Theuni2 | 12:30 |
| *** srichter has joined #zope | 12:31 | |
| Theuni2 | agroszer: hey | 12:31 |
| agroszer | how is it going? | 12:31 |
| * kosh set Theuni2 on fire | 12:31 | |
| Theuni2 | python 2-3 porting buildout is a nasty business | 12:31 |
| * kosh hands out marshmallows | 12:31 | |
| Theuni2 | kosh: thanks :) | 12:31 |
| *** tlotze has joined #zope | 12:31 | |
| * Theuni2 pushes kosh over the edge | 12:31 | |
| Theuni2 | good ol'e times | 12:31 |
| * Theuni2 burns happily | 12:32 | |
| *** teix has joined #zope | 12:32 | |
| agroszer | Theuni2, what about private packages, are going to tocuh that subject? | 12:34 |
| Theuni2 | agroszer: no, doesn't look like it | 12:34 |
| kosh | you can't push me over the edge | 12:34 |
| kosh | I went WAY over the edge long ago | 12:34 |
| agroszer | :-S | 12:34 |
| betabug | kosh: he's pushing you over the edge, but in the other direction | 12:34 |
| kosh | betabug: I will never let him push me towards sanity! :) | 12:35 |
| betabug | hehe | 12:35 |
| J1m | My IRC notifications seem to be hosed. | 12:35 |
| wosc | agroszer: what are private packages about? | 12:35 |
| betabug | wosc: nobody knows, they're private | 12:36 |
| Theuni2 | J1m: We started looking at the Python 2->3 branch yesterday and we deeply understand why you're annoyed with various aspects on the buildout code maintance. | 12:36 |
| J1m | I wrote buuldoutsftp a long time ago to deal with private packages. | 12:36 |
| J1m | Works pretty well for us. | 12:36 |
| J1m | Theuni2, heh | 12:36 |
| J1m | Dis you learn anything interesting about distutils2? | 12:38 |
| J1m | did | 12:38 |
| Theuni2 | yes | 12:38 |
| * J1m loves making Theune type. | 12:38 | |
| Theuni2 | on the plus side: the API for dealing with mirrors and downloading packages seems simple. | 12:38 |
| CIA-83 | ctheune python-3-2 * r121321 zc.buildout/src/zc/buildout/testing.py: | 12:38 |
| CIA-83 | Fix IO encoding on Python 3.2 for subprocess interaction to detect other | 12:38 |
| CIA-83 | installed Python versions. (Doesn't completely fix selecting-python.txt, | 12:38 |
| CIA-83 | though.) | 12:38 |
| CIA-83 | hannosch * r121322 Zope/buildout.cfg: Allow getting argparse from googlecode | 12:38 |
| CIA-83 | hannosch 2.13 * r121323 Zope/buildout.cfg: Backport c121322 from trunk | 12:38 |
| CIA-83 | tlotze python-3-2 * r121324 zc.buildout/src/zc/buildout/download.py: fixed recent import fixes | 12:38 |
| CIA-83 | tlotze python-3-2 * r121325 zc.buildout/src/zc/buildout/testing.py: made testing.py compatible with both python 2 and 3 | 12:38 |
| Theuni2 | and also seems like the right place to direct some effort for mirror-selection, fail-over, ... etc | 12:39 |
| J1m | cool | 12:39 |
| Theuni2 | on the negative side: it doesn't support the existing metadata format at all and we don't have an information yet whether it is supposed to | 12:39 |
| Theuni2 | s/an/any/ | 12:39 |
| Theuni2 | we'll contact tarek about that today | 12:39 |
| Theuni2 | martin has been working on a script to demonstrate downloading a (recursively resolved) set of packages that we can use to experiment and demonstrate performance improvements | 12:40 |
| J1m | cool | 12:40 |
| Theuni2 | thomas and i looked at the Python 2->3 branch and (as you probably noticed) did some changes already | 12:41 |
| Theuni2 | we hit a rabbit hole yesterday with bootstrapping of some cross-version tests | 12:41 |
| J1m | It would so obviously be a mistake not to support existing meta data that I can't believe it wouldn't be supported. | 12:41 |
| Theuni2 | yeah | 12:41 |
| Theuni2 | we're wondering whether the buildout tests really need to run other version interpreters | 12:41 |
| * J1m doesn't read checkins | 12:42 | |
| Theuni2 | J1m: fine too. so you know now. :) | 12:42 |
| J1m | I know I sorta should, but life is short. I'm virtuous in other ways. :) | 12:43 |
| Theuni2 | SAme here. So: how do you feel the cross-version interpreter calls in the test suite? | 12:43 |
| J1m | Well, it's really a question of what you feel about the feature of supporting multiple versions within a buildout. | 12:44 |
| J1m | If you think the feature is important, it's worth testing. | 12:44 |
| tlotze | At least it's very icky to have Python versions 2.5 and 2.6 hard-coded. | 12:45 |
| J1m | It's a feature I use, but I've discussed dropping it before. | 12:45 |
| Theuni2 | We're having a hard time figuring out the intention of the tests ... some seem like "this test needs to run under 2.5 so we'll just kick up a python 2.5" others might be about running different interpreter versions within a buildout. | 12:45 |
| J1m | Well, I don't object to changing that. | 12:45 |
| J1m | If you can think of something better. | 12:45 |
| J1m | Basically, the idea is that you need to *another* version of Python. | 12:46 |
| tlotze | sure | 12:46 |
| Theuni2 | Right, the specific issue is that under Python 3.2 calling the 2.5 interpreter ends up with having a distribute that is Python 3 compatible pre-injected due to the Python path. | 12:46 |
| *** menesis has quit IRC | 12:46 | |
| J1m | The approach so far has bee to pick the previous version. Of course that gets out of date. | 12:46 |
| Theuni2 | Python2.5 fails on that. which indicates that the multi-version support might have become harder when introducing Python 3. | 12:47 |
| J1m | Well, I suspect this branch doesn't work w Py2 atm. | 12:47 |
| Theuni2 | On that front we're making progress. | 12:47 |
| Theuni2 | We got dev.py working on all supported versions and the testrunner doing something. | 12:48 |
| J1m | so it might be best to skip the multiversion tests for now until we run on multiple versions. :) | 12:48 |
| *** strichter has joined #zope | 12:48 | |
| tlotze | Somehow talking about different Python versions within one buildout sounds strange when contrasted with, e.g., the restriction of not even allowing different versions of an egg in different parts. | 12:48 |
| Theuni2 | We're still trying to establish a base-line of what's broken on 3.2, fix the low-hanging fruits there and keep going back and forth to have (almost) the same set of tests fail on all of them. | 12:48 |
| tlotze | To me, things like this sound like an indicator to keep things simple and drop multi-version support, but there might of course be use cases I don't know about. | 12:49 |
| Theuni2 | J1m: i'm fine with that. actually understanding that its about a feature to support multiple versions makes more sense than the impression of just randomly picking specific interpreter versions. | 12:49 |
| J1m | As I tried to say yesterday, a reasonable sprint strategy would be to focus on the tests that pass w Python 3.2 and make them work with Py2.5-2.7 | 12:49 |
| tlotze | that's waht we're currently doing | 12:49 |
| J1m | Theuni2, like I said, feel free to come up with something better. | 12:49 |
| Theuni2 | ack. | 12:50 |
| Theuni2 | wait | 12:50 |
| Theuni2 | py2.5? | 12:50 |
| J1m | Testing is meta programming, so now you're into meta meta programming. | 12:50 |
| Theuni2 | yesterday the list included 2.4 | 12:50 |
| *** srichter has quit IRC | 12:50 | |
| J1m | It might be best to find some way to explicitly lst the versions you want to use. | 12:50 |
| tlotze | An interesting point about the feature of supporting multiple python versions is whether that should work across the 2/3 gap. Supporting it within either 2.x or 3.x should be a lot easier. | 12:51 |
| J1m | If we keep the featutre, I think it should work accross any supported versions. | 12:52 |
| tlotze | k | 12:52 |
| J1m | Including jython, etc. | 12:52 |
| J1m | But I can live with dropping the feature. | 12:52 |
| tlotze | sounds like fun testing | 12:52 |
| Theuni2 | Actually it sounds more like cleaning up the architecture and cleaning up between the distinction of the Python interpreter buildout is run with versus what the application is run with. | 12:53 |
| Theuni2 | Historically both versions were the same and I would guess clean abstraction would help with testing there. | 12:53 |
| J1m | Let's keep in mind that while buildout is useful, it's not terribly exciting to maintain, so doing things that lower the maintenance burden, even at the expense of features are well worth considering. | 12:53 |
| Theuni2 | ack | 12:54 |
| Theuni2 | so, other question, as that was dropped: do we support 2.4 or not? | 12:54 |
| Theuni2 | We thought we do. | 12:54 |
| J1m | There are *lots* of projects I'd much rather work on. | 12:54 |
| J1m | I'm OK w dropping Python 2.4 support. | 12:55 |
| J1m | On that topic, I think distutils: | 12:55 |
| J1m | - Ads support for Python version compat in meta data | 12:55 |
| tlotze | distutils2, you mean? | 12:56 |
| J1m | - And PyPy supports (or will support) getting meta data from PyPy wo downloading the package. | 12:56 |
| J1m | yes | 12:56 |
| tlotze | Yes, there's an API for that which works on new-style meta data. | 12:57 |
| J1m | so it would be great if buildout or the improved download library could take supported Python version into account when picking distros. | 12:57 |
| tlotze | ack | 12:57 |
| J1m | It would be especially great if we could leverage this somehow so as not to break existing Python 2.4 users. | 12:58 |
| tlotze | J1m: coming back to multiple python version support again, are there actual use cases for that feature? | 12:58 |
| J1m | Maybe releaese a Python 2.4-compatible version that knows not to try to upgrade to a Python 2.4-incompatible version. | 12:59 |
| J1m | tlotze, yes. | 12:59 |
| tlotze | k | 12:59 |
| J1m | We've used it. For example, we had buildouts for Zope 2 projects that needed old Python versions but included tools that used newer versions. | 13:00 |
| Theuni2 | J1m: so far we don't have trouble supporting Python 2.4 | 13:00 |
| J1m | I have a buildout in my home directory that builds various scripts for multiple versions of Python. | 13:01 |
| Theuni2 | so I'd be happy to keep it a supported version for now. | 13:01 |
| J1m | Cool. | 13:01 |
| *** slackrunner has quit IRC | 13:01 | |
| J1m | Again, I'm willing to drop the multi-version support if it becomes too painful. | 13:03 |
| *** slackrunner has joined #zope | 13:04 | |
| Theuni2 | ack | 13:04 |
| tlotze | OK, understood. | 13:04 |
| J1m | BTW, I'm interested in anything that makes maintaining buildout easier. | 13:04 |
| Theuni2 | yeah, we too after yestday's session. :) | 13:05 |
| J1m | The tests are rather painful. It's mostly necessary to run tests in sub-processes because buildout manipulates the Python environment so much when it runs. | 13:06 |
| J1m | I've wondered if, for some functionality, especially for error reporting, if there's a better way to run tests. Like maybe somehow leveraging multi-processing. That would require dropping support for older Python versions. | 13:07 |
| *** nitrogenycs has joined #zope | 13:07 | |
| J1m | I don't have a good anser, but I'm interested in ideas. | 13:07 |
| J1m | answer | 13:07 |
| tlotze | might be a good idea to look into multiprocessing | 13:07 |
| tlotze | but that's probably a good next step after porting | 13:08 |
| J1m | BTW, I just remembered that I used to deal with Python version dependencies by distributing buildout in egg , rather than in source-release form. | 13:09 |
| J1m | That's a really easy way to not break buildouts when dropping support for older Python versions. | 13:10 |
| J1m | At least until we can think of something better. | 13:10 |
| J1m | tlotze, yup | 13:10 |
| Theuni2 | something we see on the issue of maintenance is that a lot of the code is very mechanical | 13:11 |
| Theuni2 | that makes it hard to study and IMHO hard to maintain because intentions aren't clear | 13:11 |
| J1m | mechanical? | 13:11 |
| Theuni2 | i think that touches what you said yesterday about trying to make setuptools invocation "work" instead of adding an abstracting API | 13:12 |
| Theuni2 | J1m: the code does a lot of things that are incidental instead of inherent to solving the problem at hand | 13:13 |
| Theuni2 | that makes it hard to understand | 13:13 |
| J1m | Because of trhe underlying apis. | 13:13 |
| J1m | the | 13:13 |
| Theuni2 | ack | 13:13 |
| J1m | Yeah. | 13:13 |
| Theuni2 | one smell also is the size of files like easy_install | 13:14 |
| J1m | <shrug> | 13:14 |
| J1m | My main regret is the name :) | 13:14 |
| Theuni2 | heh | 13:14 |
| vortec | haha | 13:14 |
| Theuni2 | I'm just violently agreeing. I think that there's a good potential for improving maintainability. | 13:15 |
| J1m | That was originally a thin wrapper around the easy_install command, but It mostly isn't now. | 13:15 |
| J1m | yup | 13:15 |
| Theuni2 | here's another question about python version support: would you be fine to temporarily drop support for jython? | 13:15 |
| J1m | We really shouldn't every try to use the command directly, but should simple invoke setup.py scripts (or the new things). | 13:15 |
| Theuni2 | yeah | 13:16 |
| J1m | I really want to use jython, so I wouldn't want to drop support for long, although the existing support is tenuous. | 13:16 |
| Theuni2 | exactly. | 13:17 |
| Theuni2 | i wouldn't even think that it still works after some of the changes. | 13:17 |
| Theuni2 | i guess we should install jython over here and have a look | 13:17 |
| tlotze | indeed | 13:17 |
| Theuni2 | (i'm not referring to our changes but some other changes that happened after adding the support) | 13:17 |
| tlotze | both, probably | 13:17 |
| J1m | I'm planning a JVM implementation of ZEO servers to escape the GIL, and to get better performance by doing more in Java or Scala. Jython will be an important enabler of that. | 13:18 |
| J1m | Feel free to ignore jython for the sprint. | 13:18 |
| Theuni2 | k | 13:18 |
| Theuni2 | we'll try to not break it on purpose but we'll also not work hard for it then | 13:19 |
| Theuni2 | Our questions so far are covered. We'll spend a bit more time to make the test situation more even beetween 3 and 2 now. | 13:19 |
| J1m | cool | 13:19 |
| tlotze | J1m: thank you for your feedback! | 13:20 |
| Theuni2 | thanks for your time | 13:20 |
| J1m | Thanks for your efforts. Let me know if you want to discuss anything else. | 13:20 |
| tlotze | sure | 13:23 |
| strichter | J1m: how high on your stack is the Java ZEO implementation? :-) | 13:25 |
| J1m | strichter, not sure | 13:28 |
| J1m | Pretty high | 13:28 |
| strichter | cool! | 13:29 |
| J1m | Our ZEO server is one of our biggest bottle necks. | 13:29 |
| strichter | J1m: ok, that was my next question | 13:29 |
| J1m | We have a server that has has ~95 clients and runs at around 120% CPU (on a 4 cpu box), which is pretty good for Python. | 13:30 |
| strichter | it is really great to see new life coming to the ZODB in light of all the NoSQL hype | 13:30 |
| strichter | have you seen http://zodb.ws? | 13:30 |
| J1m | Yeah, although I haven't seen it do anything. | 13:30 |
| strichter | J1m: which means you are running already significant parts of the server in C space | 13:31 |
| J1m | Right, mostly the IO. | 13:31 |
| strichter | I have only played with it too; I have to find a small project to try it out; Python in the browser is also just very attractive to me | 13:31 |
| J1m | This is the result of the "thready zeo" work for 3.10. | 13:31 |
| J1m | Dude, come to terms with Javascript. Once you get past that, the world is a saner place. :) | 13:32 |
| strichter | J1m: I am, I got pretty good, just don't like it :-) | 13:33 |
| J1m | JS isn't that bad. | 13:33 |
| J1m | At least there's no JS 3 :) | 13:33 |
| strichter | LOL | 13:33 |
| CIA-83 | tlotze python-3-2 * r121326 zc.buildout/src/zc/buildout/ (testing.py tests.py): rewrote some print calls to give the same output in Python 2 and 3 | 13:33 |
| J1m | BTW, wrt to threadyness. You'd be amazed how long "non-blocking" IO can block. | 13:35 |
| J1m | We were seeing non-blocking IO calls taking up to a millisecond, at least according to strace. | 13:36 |
| J1m | Perhaps there's some issue with Python's socket implementation. <shrug> | 13:37 |
| * J1m suspects YUI3 will kill YUI. | 13:38 | |
| *** alga has quit IRC | 13:45 | |
| *** yvl has quit IRC | 13:46 | |
| *** regebro has joined #zope | 14:06 | |
| *** alga has joined #zope | 14:07 | |
| *** sunew has joined #zope | 14:12 | |
| *** purserj_ is now known as purserj | 14:20 | |
| *** menesis has joined #zope | 14:34 | |
| *** nitrogenycs has left #zope | 14:47 | |
| *** nitrogenycs has joined #zope | 14:48 | |
| *** fermigier has joined #zope | 14:57 | |
| *** fermigier has quit IRC | 14:58 | |
| *** Wu has quit IRC | 15:05 | |
| *** thetet has joined #zope | 15:07 | |
| *** yvl has joined #zope | 15:07 | |
| *** paddy_ has joined #zope | 15:17 | |
| *** _srichter has joined #zope | 15:21 | |
| *** strichter has quit IRC | 15:22 | |
| *** fredvd has joined #zope | 15:32 | |
| *** Arfrever has joined #zope | 15:33 | |
| *** m8 has joined #zope | 15:41 | |
| *** yvl has quit IRC | 15:53 | |
| *** menesis has quit IRC | 15:53 | |
| *** alexpilz1 has quit IRC | 15:55 | |
| *** _srichter has quit IRC | 15:55 | |
| *** menesis has joined #zope | 16:02 | |
| Theuni2 | tlotze: http://dpaste.com/529643/ | 16:18 |
| *** _srichter has joined #zope | 16:19 | |
| J1m | wow | 16:19 |
| J1m | Y'all are making progress. | 16:19 |
| Theuni2 | we're trying :) | 16:26 |
| *** m8 has quit IRC | 16:26 | |
| Theuni2 | J1m: that paste is the baseline i got on python 3.2 and that we try to get on the 2 series, t oo | 16:27 |
| *** wosc has quit IRC | 16:43 | |
| *** dayne has quit IRC | 16:44 | |
| tlotze | Theuni2: http://dpaste.com/529655/ (python-3-2 branch, Python 2.4) | 16:46 |
| *** slackrunner has quit IRC | 16:51 | |
| *** __mac__ has quit IRC | 17:01 | |
| *** slackrunner has joined #zope | 17:03 | |
| *** astoon has joined #zope | 17:05 | |
| *** Wu has joined #zope | 17:05 | |
| *** alecm has quit IRC | 17:15 | |
| *** alecm has joined #zope | 17:16 | |
| *** alecm has joined #zope | 17:16 | |
| *** alecm has joined #zope | 17:17 | |
| *** alecm has joined #zope | 17:17 | |
| *** sm has joined #zope | 17:31 | |
| *** slackrunner has quit IRC | 17:34 | |
| CIA-83 | tlotze python-3-2 * r121327 zc.buildout/src/zc/buildout/ (update.txt buildout.txt buildout.py): rewrote remaining print calls to give the same output in Python 2 and 3 | 17:37 |
| CIA-83 | tlotze python-3-2 * r121328 zc.buildout/src/zc/buildout/update.txt: fixed remaining call to subprocess.check_call | 17:37 |
| CIA-83 | tlotze python-3-2 * r121329 zc.buildout/src/zc/buildout/ (5 files): avoid representations of (unicode) text appearing in doctests | 17:37 |
| CIA-83 | tlotze python-3-2 * r121330 zc.buildout/src/zc/buildout/tests.py: fixed string substitution syntax broken by a recent print() fix | 17:37 |
| CIA-83 | mauritsvanrees * r121331 ZODB/src/persistent/dict.py: Fixed typo in comment | 17:37 |
| CIA-83 | ctheune python-3-2 * r121332 zc.buildout/src/zc/buildout/ (compat23.py testing.py easy_install.py): | 17:37 |
| CIA-83 | Restructure calling external interpreters. | 17:37 |
| CIA-83 | * make IO handling consistent on Python 2 and 3 | 17:37 |
| CIA-83 | * simplify business code, extract into separate function | 17:37 |
| CIA-83 | faassen * r121333 hurry.workflow/src/hurry/workflow/interfaces.py: Interface wasn't complete. | 17:37 |
| *** FSCV has joined #zope | 17:41 | |
| *** digitalmortician has quit IRC | 17:43 | |
| *** sm_ has joined #zope | 17:43 | |
| *** sm has quit IRC | 17:46 | |
| *** sm_ is now known as sm | 17:46 | |
| *** tiwula has joined #zope | 17:51 | |
| *** sunew has quit IRC | 17:57 | |
| *** alexpilz has joined #zope | 17:57 | |
| *** zagy has quit IRC | 18:08 | |
| *** TomBlockley has joined #zope | 18:10 | |
| *** altipard has quit IRC | 18:10 | |
| *** sunew has joined #zope | 18:12 | |
| *** sunew has quit IRC | 18:12 | |
| *** daMaestro has joined #zope | 18:14 | |
| *** daMaestro has joined #zope | 18:14 | |
| *** giampaolo has joined #zope | 18:14 | |
| *** TomBlockley has quit IRC | 18:15 | |
| *** ccomb has quit IRC | 18:19 | |
| *** agroszer has quit IRC | 18:46 | |
| *** cpf_ has quit IRC | 18:48 | |
| *** goschtl has quit IRC | 18:48 | |
| *** paddy_ has quit IRC | 18:52 | |
| *** avoinea has quit IRC | 18:57 | |
| *** fredvd has quit IRC | 18:59 | |
| *** alga has quit IRC | 19:11 | |
| *** shastry_ has joined #zope | 19:15 | |
| *** shastry has quit IRC | 19:19 | |
| *** thetet has quit IRC | 19:21 | |
| *** mitchell` is now known as mitchell`off | 19:27 | |
| *** nlzh has quit IRC | 19:32 | |
| *** evilbungle_ has joined #zope | 19:32 | |
| *** evilbungle_ has quit IRC | 19:34 | |
| *** menesis has quit IRC | 19:36 | |
| *** evilbungle has quit IRC | 19:37 | |
| *** bowe has quit IRC | 19:43 | |
| *** zagy has joined #zope | 19:47 | |
| *** nitrogenycs has quit IRC | 19:49 | |
| *** giampaolo has quit IRC | 20:01 | |
| *** _srichter has quit IRC | 20:03 | |
| *** bowe has joined #zope | 20:03 | |
| *** eperez has quit IRC | 20:05 | |
| *** astoon has quit IRC | 20:25 | |
| *** Wu has quit IRC | 20:29 | |
| *** vortec has left #zope | 20:29 | |
| CIA-83 | jim * r121334 /ZODB/branches/jim-jvm-zeo: Experimental | 20:34 |
| *** mcdonc has quit IRC | 20:44 | |
| *** _mup_ has quit IRC | 20:46 | |
| *** Theuni2 has quit IRC | 20:47 | |
| *** _mup_ has joined #zope | 20:47 | |
| *** menesis has joined #zope | 20:49 | |
| *** m8 has joined #zope | 20:53 | |
| *** J1m has quit IRC | 20:58 | |
| *** hever has quit IRC | 20:58 | |
| *** webmaven has joined #zope | 21:01 | |
| *** hever has joined #zope | 21:04 | |
| *** hever has quit IRC | 21:07 | |
| *** hever has joined #zope | 21:09 | |
| *** __mac__ has joined #zope | 21:15 | |
| *** mcdonc has joined #zope | 21:16 | |
| *** __mac__ has quit IRC | 21:38 | |
| *** hever has quit IRC | 21:39 | |
| *** __mac__ has joined #zope | 21:43 | |
| *** zagy has quit IRC | 21:45 | |
| *** teix has quit IRC | 21:46 | |
| *** alga has joined #zope | 22:09 | |
| *** m8 has quit IRC | 22:14 | |
| *** slackrunner has joined #zope | 22:33 | |
| *** webmaven has quit IRC | 22:41 | |
| *** donri has joined #zope | 23:00 | |
| donri | How are persistent.dict and mapping different? | 23:00 |
| donri | API docs suggest the former inherits IterableUserDict but mapping is iterable too … | 23:01 |
| *** planetzopebot has quit IRC | 23:01 | |
| *** __mac__ has quit IRC | 23:02 | |
| donri | Nevermind, looked at the source | 23:04 |
| *** avoinea has joined #zope | 23:04 | |
| *** avoinea has left #zope | 23:04 | |
| *** _srichter has joined #zope | 23:08 | |
| *** planetzopebot has joined #zope | 23:09 | |
| *** shastry_ is now known as shastry | 23:10 | |
| *** RichardBarrell has joined #zope | 23:15 | |
| *** runyaga_ has joined #zope | 23:15 | |
| *** slackrunner has quit IRC | 23:25 | |
| *** alga has quit IRC | 23:27 | |
| *** kosh has quit IRC | 23:36 | |
| *** kosh has joined #zope | 23:38 | |
| *** alga has joined #zope | 23:38 | |
| *** webmaven has joined #zope | 23:43 | |
Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!