Options:
- # Session Start: Mon Aug 12 00:00:00 2013
- # Session Ident: #testing
- # [00:11] * Joins: gitbot (~gitbot@public.cloak)
- # [00:11] -gitbot:#testing- [web-platform-tests] heycam opened pull request #271: Remainder of my tests (master...submission/heycam/WebIDL-tests-1) https://github.com/w3c/web-platform-tests/pull/271
- # [00:11] * Parts: gitbot (~gitbot@public.cloak) (gitbot)
- # [00:23] * Joins: gitbot (~gitbot@public.cloak)
- # [00:23] -gitbot:#testing- [web-platform-tests] heycam synchronize pull request #271: Remainder of my tests (master...submission/heycam/WebIDL-tests-1) https://github.com/w3c/web-platform-tests/pull/271
- # [00:23] * Parts: gitbot (~gitbot@public.cloak) (gitbot)
- # [00:55] * Quits: timeless (~uid4015@public.cloak) ("Connection closed for inactivity")
- # [02:27] * heycam is now known as heycam|away
- # [03:19] * heycam|away is now known as heycam
- # [06:58] * Joins: shepazu (schepers@public.cloak)
- # [06:59] * Joins: zcorpan (~zcorpan@public.cloak)
- # [08:21] * heycam is now known as heycam|away
- # [09:44] * Joins: AutomatedTester (~AutomatedTester@public.cloak)
- # [09:46] * Quits: AutomatedTester (~AutomatedTester@public.cloak) (Client closed connection)
- # [10:01] * Joins: dom (dom@public.cloak)
- # [10:16] * Quits: Lachy (~Lachy@public.cloak) ("Computer has gone to sleep.")
- # [10:31] * Joins: Ms2ger (~Ms2ger@public.cloak)
- # [10:33] * Joins: darobin (rberjon@public.cloak)
- # [10:41] * Joins: Lachy (~Lachy@public.cloak)
- # [11:06] * heycam|away is now known as heycam
- # [11:32] * Joins: gitbot (~gitbot@public.cloak)
- # [11:32] -gitbot:#testing- [web-platform-tests] jgraham pushed 1 new commit to master: https://github.com/w3c/web-platform-tests/commit/4d2a2b6cb44dbdfa19d2f46dd848c8df0faab85c
- # [11:32] -gitbot:#testing- web-platform-tests/master 4d2a2b6 Aryeh Gregor: Add tests for reflection of content attributes in IDL attributes....
- # [11:32] * Parts: gitbot (~gitbot@public.cloak) (gitbot)
- # [11:34] * Joins: gitbot (~gitbot@public.cloak)
- # [11:34] -gitbot:#testing- [web-platform-tests] jgraham closed pull request #1: Aryeh's submission (master...submission/AryehGregor) https://github.com/w3c/web-platform-tests/pull/1
- # [11:34] * Parts: gitbot (~gitbot@public.cloak) (gitbot)
- # [12:22] * Joins: abarsto (~abarsto@public.cloak)
- # [12:22] * abarsto is now known as ArtB
- # [12:26] * Joins: tobie (tobie@public.cloak)
- # [13:09] * heycam is now known as heycam|away
- # [14:05] <jgraham> So, if I were writing a python server for testing, how would people like advanced (php-replacing) functionality to work?
- # [14:05] <jgraham> One option is that people write .py files with a number of predefined globals
- # [14:06] <jgraham> Like "request", "response" and "utils"
- # [14:06] <jgraham> Where "utils" might contain e.g. a function to take a file/string and return it piece by piece
- # [14:07] <jgraham> Or
- # [14:07] <jgraham> We could do something else
- # [14:07] <jgraham> Not sure what makes reusing the common things easiest
- # [14:08] <jgraham> Of course each funciton in utils could be exposed in some way as a url + query string
- # [14:10] <jgraham> so maybe you could do /utils/delayed?path=/foo/bar&parts=1d2b1d
- # [14:11] <wilhelm> Without answering your question: Note that the Mozilla Python HTTP server is already in use by the WebDriver tests.
- # [14:24] * Joins: jimevans (~jimevans@public.cloak)
- # [14:46] * Quits: jimevans (~jimevans@public.cloak) ("Colloquy for iPhone - http://colloquy.mobi")
- # [14:49] <tobie> jgraham: I see two kind of services: one that is standalone (e.g. an echo server)...
- # [14:49] <tobie> and one which isn't (eg delaying a regular response)
- # [14:50] <tobie> The latter I'd add a params of a regular request
- # [14:50] <tobie> e.g.: /foo/bar.html returns /foo/bar.html
- # [14:51] <tobie> e.g.: /foo/bar.html?delay=2 returns /foo/bar.html after 2 seconds
- # [14:51] <tobie> etc.
- # [15:11] <gsnedders> Note that some of jgraham's tests I'm currently getting out use a script currently called slice-and-dice, which allows to delays at arbitrary byte offsets in the file.
- # [15:24] <tobie> gsnedders: pointers?
- # [15:27] <jgraham> gsnedders: I plan to port slice-and-dice
- # [15:27] <jgraham> being able to implement that is a requirement
- # [15:27] <jgraham> wilhelm: I investigated that and it's not enough
- # [15:27] <jgraham> It doesn't allow you to do things like override the headers
- # [15:27] <jgraham> +required
- # [15:28] <jgraham> So you can't even do things like mod_asis
- # [15:37] <gsnedders> tobie: There's a PHP rewrite in my repo, op_browsing branch
- # [15:37] <tobie> gsnedders: which is?
- # [15:38] <gsnedders> https://github.com/gsnedders/web-platform-tests/blob/4aed7c6d584beb610d6ab04b6e483528aa497e87/html/browsers/browsing-the-web/navigating-across-documents/slice-and-dice.php
- # [15:40] <tobie> ty
- # [15:40] <tobie> so that falls in the second category I was suggesting
- # [15:40] <gsnedders> Right
- # [15:41] <jgraham> I don't think there should be two categories
- # [15:41] <jgraham> I think there should be /utils
- # [15:41] <tobie> so the only difference between Jame's proposal and mine is his is backwards compatible and mine is nicer :D
- # [15:42] <jgraham> Where each file contains a function which is registered as a handler for a particular url string in a well defined way
- # [15:42] <jgraham> So if you had /utils/slice-and-dice it might have a signature like
- # [15:42] <tobie> jgraham: yeah, the problem with that is that is you can't combine them.
- # [15:43] <jgraham> def slice_and_dice(file=file, parts=parts):
- # [15:43] <jgraham> pass
- # [15:43] <jgraham> And then more complex behaviour would require that you write your own .py file
- # [15:43] <jgraham> Where utils would be avaliable
- # [15:43] <tobie> sure, you'd still want to think of them as middleware which you could combine
- # [15:44] <jgraham> so you could use utils.slice_and_dice("foo/bar.html", "1,2d")
- # [15:44] <tobie> yeah, I mean't in the url itself
- # [15:45] <jgraham> I don't know how you would compose them in a url
- # [15:45] <tobie> e.g. foo/bar.html?delay=3&responseStatus=200
- # [15:45] <jgraham> That would mean that you could only have one parameter per function or something
- # [15:45] <tobie> jgraham: ^
- # [15:45] <tobie> not necessarily.
- # [15:46] <jgraham> Or at least you would need unique naming
- # [15:46] <jgraham> And order isn't very obvious
- # [15:46] <jgraham> But sure, I guess it is possible to invent some scheme that you could write in a URL
- # [15:47] <tobie> these are valid concerns
- # [15:47] <jgraham> After a point, it's likely just easier to write a custom .py file
- # [15:47] <tobie> that too
- # [15:47] <jgraham> Which you need for lots of things anyway
- # [15:47] * jgraham needs to look at all the requirements that mozilla have in Mochitests
- # [15:47] <tobie> I just liked the liked the idea that you could abritraly modify the behaviour of a url by passing in params
- # [15:48] <tobie> feels very intuitive
- # [15:48] <jgraham> Well I don't disagree that something like that could work for simple cases
- # [15:48] <tobie> more than passing said url as a param to a function
- # [15:48] <tobie> yeah, a list of more complex requirements would be useful
- # [15:49] <tobie> (haven't dug much there)
- # [15:51] <jgraham> foo/bar?pipe=slice("1,d1,2000")|status(200) or something
- # [15:51] <jgraham> Seems like it might be nice or might be too complex
- # [15:54] <tobie> lets look at the reqs, then we see if we can come up with a nice dsl
- # [15:54] <tobie> else we'll go for something simpler
- # [15:54] <tobie> might be useful to be able to set http headers that way too
- # [15:54] <tobie> for one offs
- # [15:54] <jgraham> Well I'm 100% sure a nice dsl won't work for all use cases
- # [15:54] <jgraham> Because there are tests that need conditional logic
- # [15:55] <jgraham> (e.g. tests where things change after a resource has been loaded once)
- # [15:55] <jgraham> So far, for headers, I was considering a .headers file
- # [15:56] <tobie> Well I'm 100% sure a nice dsl won't work for all use cases -> agreed
- # [15:57] <tobie> So far, for headers, I was considering a .headers file -> yup, me too. Might still be useful to be able to override those through url params
- # [15:57] <jgraham> Fair enough
- # [15:57] <tobie> … let's do .headers first, then we can see for the rest
- # [15:57] <jgraham> Well first I will see if I can serve plain files :)
- # [15:57] <tobie> :)
- # [15:58] <jgraham> (*very* simple things work, so maybe I can)
- # [15:59] * Joins: jimevans (~jimevans@public.cloak)
- # [16:06] <tobie> Catching up on last week's threads. Decided to create a FAQ to answer some common questions. Help welcomed.
- # [16:06] <tobie> https://github.com/w3c/testtwf-website/pull/67/files
- # [16:09] * Quits: jimevans (~jimevans@public.cloak) ("Colloquy for iPhone - http://colloquy.mobi")
- # [17:10] <ArtB> some of the eventsource tests e.g. view-source:http://w3c-test.org/web-platform-tests/master/eventsource/eventsource-cross-origin.htm try to access crosssite.w3c-test.org
- # [17:13] <ArtB> darobin, mikesmith, dom - is crosssite.w3c-test.org down?
- # [17:15] <darobin> ArtB: has it ever existed?
- # [17:15] <jgraham> I don't remember that being one that the W3C had
- # [17:15] <darobin> we have www2 for that; I don't recall there being a crosssite.
- # [17:15] <ArtB> beats me
- # [17:16] <jgraham> Opera had it, so it is possible that it made it into some of our tests
- # [17:16] <ArtB> ok, thanks, I'll start a thread on the list
- # [17:21] <odinho> It's just changing it.
- # [17:21] * Quits: Lachy (~Lachy@public.cloak) ("Computer has gone to sleep.")
- # [17:21] <odinho> I had a support.js file with that URL as a variable.
- # [17:21] <odinho> Also, we got www2 internally so that we could rather use the W3C urls.
- # [17:23] <ArtB> odinho not sure what you you mean by "It's just changing it"
- # [17:24] <odinho> The tests
- # [17:24] <odinho> That has crosssite written in them. :-) Change that to "www2"
- # [17:25] <ArtB> So, s/crosssite/www2/ ?
- # [17:25] <odinho> Yeah, something like that :)
- # [17:26] <jgraham> You get to play with sed :)
- # [17:30] * Joins: gitbot (~gitbot@public.cloak)
- # [17:30] -gitbot:#testing- [web-platform-tests] AFBarstow opened pull request #272: Change crosssite host name to www2 (master...master) https://github.com/w3c/web-platform-tests/pull/272
- # [17:30] * Parts: gitbot (~gitbot@public.cloak) (gitbot)
- # [17:32] * Joins: gitbot (~gitbot@public.cloak)
- # [17:32] -gitbot:#testing- [web-platform-tests] Velmont closed pull request #272: Change crosssite host name to www2 (master...master) https://github.com/w3c/web-platform-tests/pull/272
- # [17:32] * Parts: gitbot (~gitbot@public.cloak) (gitbot)
- # [17:32] * Joins: gitbot (~gitbot@public.cloak)
- # [17:32] -gitbot:#testing- [web-platform-tests] Velmont pushed 2 new commits to master: https://github.com/w3c/web-platform-tests/compare/4d2a2b6cb44d...8bf21c32d628
- # [17:32] -gitbot:#testing- web-platform-tests/master ec1a757 Arthur Barstow: Change crosssite host name to www2
- # [17:32] -gitbot:#testing- web-platform-tests/master 8bf21c3 Odin Hørthe Omdal: Merge pull request #272 from AFBarstow/master...
- # [17:32] * Parts: gitbot (~gitbot@public.cloak) (gitbot)
- # [17:32] <jgraham> odinho: Bah, I did the review first
- # [17:37] <odinho> jgraham: lol, I used github's tool.
- # [17:37] <odinho> It was 3 lines
- # [17:37] * Quits: zcorpan (~zcorpan@public.cloak) (Client closed connection)
- # [17:37] <jgraham> there's no evidence that you did anything at all :)
- # [17:46] * Joins: jhammel (~jhammel@public.cloak)
- # [18:08] * Disconnected
- # [18:09] * Attempting to rejoin channel #testing
- # [18:09] * Rejoined channel #testing
- # [18:17] * Joins: rhauck (~Adium@public.cloak)
- # [18:18] <jgraham> Anyone object if I create github.com/w3c/wptserve and fork it to work on?
- # [18:19] * jgraham realises he doesn't really mind if anyone cares
- # [18:22] * Quits: rhauck (~Adium@public.cloak) ("Leaving.")
- # [18:22] * Joins: rhauck (~Adium@public.cloak)
- # [18:22] <jgraham> (done)
- # [18:23] * Quits: rhauck (~Adium@public.cloak) (Client closed connection)
- # [18:24] <jgraham> ArtB: I don't know what you are doing, but github seems to think that you keep pushing new commits to old PRs
- # [18:24] * Joins: rhauck (~Adium@public.cloak)
- # [18:24] <jgraham> Well, actually
- # [18:24] <jgraham> I think the problem is that you have used the master branch for multiple things
- # [18:25] <jgraham> and so when critic updates it, it gets a bunch of extra stuff
- # [18:25] <jgraham> Generally one branch per PR is a good idea
- # [18:28] <ArtB> (I don't know why I keep getting critic emails so I've just been deleting them ;-))
- # [18:28] <jgraham> Well there are two problems :)
- # [18:29] <jgraham> One is that critic doesn't seem to be getting notified by github when PRs are closed
- # [18:29] <jgraham> Or it is getting notified, but doesn't close the review, although the code to do that is clearly there
- # [18:29] <jgraham> the other problem is that you are creating reviews from the master branch
- # [18:29] <jgraham> Which isn't very hygenic
- # [18:38] <darobin> oooh, some jgraham hot action on the server side
- # [18:41] <darobin> I wish GitHub kept people from making PRs from their master branch
- # [18:41] <tobie> ^ +1
- # [18:45] <ArtB> sorry for my ignorance here re branches but why is this a (big?) problem?
- # [18:45] <jgraham> darobin: well feel free to look at the work so far, but there isn't much
- # [18:45] <jgraham> I expect ~100% of the code to be rewritten :)
- # [18:45] * jgraham away
- # [18:52] <darobin> jgraham: it looks pretty good so far, almost usable
- # [18:59] * Quits: darobin (rberjon@public.cloak) (Client closed connection)
- # [19:15] <tobie> ArtB: because every pull request you make also starts containing the other pull requests you made and so accepting one implies accepting the previous ones too.
- # [19:18] <tobie> ArtB: consider if you made commits 1,2, 3 into PR A, then continuing from the same branch you now make PR B with commit 4
- # [19:18] <tobie> PR B will infact reference commits 1,2,3 and 4.
- # [19:19] <tobie> So merging B also merges A. Except PR A still stays there.
- # [19:19] <tobie> No imagine, if in the meantime, you backed out of A or added commits 5 there. Everything quickly gets out of hand.
- # [19:19] <ArtB> oh, I see; thanks tobie; I'll try to do do-the-right-thing going fwd
- # [19:21] <tobie> ArtB: as a primer to our doc effort: https://github.com/w3c/testtwf-website/blob/gh-pages/docs/github-101.md
- # [19:40] * Joins: Lachy (~Lachy@public.cloak)
- # [19:42] * Quits: dom (dom@public.cloak) (Ping timeout: 180 seconds)
- # [19:43] * Joins: zcorpan_ (~zcorpan@public.cloak)
- # [19:44] * Quits: zcorpan (~zcorpan@public.cloak) (Client closed connection)
- # [20:08] * Joins: timeless (~uid4015@public.cloak)
- # [21:47] * Quits: tobie (tobie@public.cloak)
- # [22:04] * Joins: tobie (tobie@public.cloak)
- # [22:19] * Quits: Ms2ger (~Ms2ger@public.cloak) ("nn")
- # [22:27] * Quits: ArtB (~abarsto@public.cloak) (Ping timeout: 180 seconds)
- # [22:30] * Joins: abarsto (~abarsto@public.cloak)
- # [22:30] * abarsto is now known as ArtB
- # [22:43] * Quits: zcorpan_ (~zcorpan@public.cloak) (Client closed connection)
- # [22:43] * Joins: zcorpan (~zcorpan@public.cloak)
- # [22:51] * Quits: zcorpan (~zcorpan@public.cloak) (Ping timeout: 180 seconds)
- # [23:12] * heycam|away is now known as heycam
- # Session Close: Tue Aug 13 00:00:00 2013
The end :)