Mochabot log - CommonJS IRC channel: #commonjs on irc.freenode.net

2010-02-04:

[4:16] <deanlandolt_> ugh, i missed the [AE]JSGI convo and all the resultant gists are fail -- i can't tell if if angry unicorn or leaky octopus makes me more sad
[5:13] <tlrobinson_> deanlandolt_: got a git server i can push the gist too ;)
[5:13] <tlrobinson_> ?
[5:14] <tlrobinson_> today i realized the full potential of a git backed pastebin
[5:14] <tlrobinson_> isaacs forked the gist, then i merged it back into mine
[5:14] <isaacs> yes, gist > pastie, by orders of magnitude
[7:58] <ondras> guys
[7:58] <ondras> require("assert").throws, what is the purpose of 2nd argument?
[8:38] <ondras> ashb: ?
[8:48] <evilstreak> the type of error
[8:48] <evilstreak> (I believe)
[8:49] <evilstreak> assert.throws( function() { throw "one"; }, "two" ) // will not pass
[8:53] <ondras> ah
[8:53] <ondras> so the code must throw
[8:53] <ondras> and the error must be of the specified type
[8:53] <evilstreak> yeah. if you don't specify that argument then any throw is accepted as a pass
[8:53] <ondras> understood
[8:53] <evilstreak> at least, that's how I interpret it
[8:54] <ondras> any expected/actual values for assert.throws? :)
[8:54] <evilstreak> hmm?
[9:04] <ondras> well
[9:04] <ondras> AssertionError works with these two values
[9:05] <ondras> it is logical that all assert.* methods pass them to AssertinError...
[9:05] <evilstreak> ah
[9:15] <evilstreak> sorry, discussion happened
[9:15] <evilstreak> I think it would work something like this:
[9:16] <evilstreak> assert.throws( function(){} ) // -> actual: no throw, expected: throw
[9:17] <evilstreak> assert.throws( function(){ throw "foo" }, "bar" ) // -> action: foo, expected: bar
[9:17] <evilstreak> err. s/action/actual/
[9:58] <ondras> evilstreak: are you related to narwhal/flusspferd ?
[9:58] <ondras> there is one more thing regarding their implementation of assert.deepEqual
[10:00] <evilstreak> ondras: lightly related to flusspferd; I work on Juice which uses it, I don't do much work on flusspferd directly
[10:01] <ondras> okay
[12:54] <lelec_> hi, has someone used js-test-driven or same king of plugins for eclipse with the narwhal js test tools ?
[15:13] <mthorn> Hi, I'm trying to use v8cgi to open a file in append mode. I called .open('a'), and didn't receive an error, but when I write to it nothing gets put into the file
[15:13] <mthorn> No error at all
[15:14] <mthorn> Any sugggestions? is 'a' not the right mode?
[15:14] <ondras> hmm
[15:15] <ondras> have you closed it?
[15:15] <mthorn> ahhhh, perhaps no
[15:15] <mthorn> it has a buffer or something then?
[15:15] <ondras> at a system level
[15:15] <ondras> v8cgi does not handle this stuff itself
[15:16] * ondras just tried appending and it certainly works
[15:16] <ondras> afk 30 minutes, we can solve that then
[15:16] <mthorn> ondras: actually that was it. sorry, n00b mistake on my part
[15:42] <ondras> np
[18:36] <Dantman> ^_^ I just had some fun with git
[18:37] <Wes-> pics or it didn't happen
[18:39] <Dantman> I took a git repo. Created a new branch and reset the head to empty. Spliced the history of another git repo into the repo. Placed the new branch into that spliced history and reset it to the head. Then did a merge commit joining the two branches together. Deleted files I didn't want and moved stuff to locations I wanted them in. Then amended the merge commit. Resulting in a git repo containing two branches with two separate initial commits which join i
[18:39] <Dantman> n one commit that looks like a perfectly clean merge.
[18:39] <Dantman> ^_^ Two git repos merged into one
[18:43] <Dantman> Chrome relies far to much on Google.com
[18:44] <Dantman> Unlike firefox it can't even guess that when you type in something.localhost you intended to go to http://something.localhost/, not google search it while you're not even online.
[18:47] <Dantman> Wes-, http://f.imagehost.org/0859/Screenshot.png
[18:47] <Dantman> ;)
[18:47] <ashb> we gitk sucks
[18:47] <ashb> *ew
[18:48] <Dantman> Bah... it works well enough
[18:48] <ashb> gogitgui is nicer
[18:48] <ashb> s/go//
[18:48] <Dantman> `git gui`?
[18:49] <ashb> think so
[18:49] <ashb> tho its no where as nice as gitx
[18:50] <ashb> (thats OSX only)
[18:51] <Dantman> git gui uses gitk
[18:54] <ondras> ashb: ?
[18:54] <ondras> ashb: http://github.com/ruediger/flusspferd/blob/master/js/assert.js#L206
[18:54] <ondras> in my opinion, the spec is not clear here
[18:54] <ashb> ondras: ask krisk
[18:54] <ashb> i stole it wholesale from narwhall
[18:54] <ondras> hm
[18:54] <ashb> ondras: i agree
[18:54] <ondras> I see differences
[18:55] <ashb> i think the spec is *Wrong* there
[18:55] <ondras> from narwhal spec
[18:55] <ondras> *narwhal impl
[18:55] <ashb> there's certainly something that came up about just what deepEq should do
[18:55] <ondras> hm
[18:55] <ondras> imho there should be ||
[18:55] <ashb> http://wiki.commonjs.org/wiki/Unit_Testing
[18:55] <ondras> so 0 vs. null won't fall to the objEquiv function
[18:55] <ashb> theres 4 things on that page
[18:56] <ashb> add more if you think there are problems
[18:56] <ondras> hm
[18:56] <ondras> I am not sure if this is spec problem
[18:56] <ondras> or impl problem :)
[18:56] <ondras> Other pairs that do not both pass typeof value == "object"
[18:56] <ondras> what exactly does this mean?
[18:57] <ashb> pass
[18:57] <ondras> if any of them do not pass, use == ?
[18:57] <ondras> or "both must not pass" ?
[18:57] <ashb> so i'm using something closer to node's version, which was taken form narwhal without hte deps on narwhal's util module
[18:57] <ondras> *does not pass
[18:57] <ondras> the thing is that narwhal is actually different here
[18:57] <ashb> so who's ever problem it is its not mine :)
[18:57] <ashb> it might have been changed afterwards
[18:58] <ashb> (in which case i guess it is my problem)
[18:58] <ondras> fp tests for null/undefined in objEquiv, something not present (maybe) in narwhal
[18:58] <ashb> and the test suote for assert needs extending
[18:58] <ondras> so, who is the author of the original spec?
[18:59] <ondras> I need someone to explain how is 7.4 precisely meant
[18:59] <ondras> *7.3
[18:59] <ashb> yeah. that is one of the things on the wiki page
[18:59] <ashb> is there's ambiguity over how its meant to behave
[18:59] <ondras> imho "at least one is not object -> use =="
[18:59] <ashb> i.e. i had a different internal modles for the deepEqual etc
[19:00] <ondras> fine, will ask krisk
[19:01] * ondras spent some quality time rewriting all his unittests from (message, expected, actual) to inversed form...

 

 

Logs by date :