00:00:00  * ircretaryquit (Remote host closed the connection)
00:00:08  * ircretaryjoined
00:00:37  <thlorenz>defunctzombie: so this is working: https://github.com/thlorenz/stack-mapper/blob/master/test/onefile.js#L25-L28
00:00:53  <thlorenz>gotta figure out how to do test this in the browser (need a vm)
00:01:23  <defunctzombie>hm
00:01:34  <defunctzombie>thlorenz: my suggestion:
00:01:46  <defunctzombie>create prepared stacktrace string
00:01:57  <defunctzombie>and a predetermined sourcemap file/json
00:02:00  <defunctzombie>and use those
00:02:04  <thlorenz>defunctzombie: good idea, so I do integration testing only with travis
00:02:05  <defunctzombie>to avoid __dirname use, etc
00:05:25  * tilgovijoined
00:10:10  * chapeljoined
00:22:03  * defunctzombiechanged nick to defunctzombie_zz
00:22:25  <thlorenz>defunctzombie: ok, all server side tests done -- the inline source of error origin can be quite useful for test asserts ;) https://github.com/thlorenz/stack-mapper/blob/master/test/threefiles-throw.js#L62-L68
00:24:04  <rowbit>Hourly usage stats: [developer: 0, free: 14]
00:24:26  * ralphtheninjaquit (Ping timeout: 252 seconds)
00:26:41  <thlorenz>so defunctzombie_zz made a test just to show how awesome this will be, with failing asserts showing up right in your stack trace
00:26:42  <thlorenz>https://github.com/thlorenz/stack-mapper/blob/master/test/threefiles-assert.js#L33
00:29:47  * thlorenzquit (Remote host closed the connection)
00:30:22  * thlorenzjoined
00:34:28  * thlorenzquit (Ping timeout: 246 seconds)
00:38:31  * wolfeidauquit (Read error: No buffer space available)
00:42:44  * yorickquit (Remote host closed the connection)
00:51:23  <substack>grncdr: so long as we can walk it, looks good!
00:59:05  <grncdr>crap, left recursion!
01:01:00  <grncdr>hm, I may just say you can't use concatenations for command names
01:01:37  <grncdr>like "ec"'ho'
01:01:56  <grncdr>or git-$subcommand
01:03:31  * Guest47861changed nick to prettyrobots_
01:10:18  <substack>can't you just fix that in a post-pass?
01:10:39  <substack>seems straight-forward to detect
01:18:46  * ferossjoined
01:21:17  <grncdr>fixed it a different way
01:21:28  <grncdr>makes the grammar a tiny bit more verbose, but it works
01:21:37  <grncdr>doing actual tests now as well :)
01:21:45  <grncdr>I should grab the ones from bashful...
01:24:04  <rowbit>Hourly usage stats: [developer: 0, free: 17]
02:05:04  * thlorenzjoined
02:24:04  <rowbit>Hourly usage stats: [developer: 0, free: 3]
02:31:57  * mikolalysenkojoined
02:36:03  <rowbit>substack, pkrumins: These encoders are STILL down:
02:49:03  <rowbit>substack, pkrumins: These encoders are STILL down:
02:51:38  * fronxjoined
02:56:43  * fronxquit (Ping timeout: 272 seconds)
03:11:40  * dominictarrjoined
03:22:41  <substack>dominictarr: for https://github.com/dominictarr/sha.js you can use https://npmjs.org/package/typedarray
03:22:51  <substack>works all the way down to ie6, all the same interfaces
03:23:12  <dominictarr>substack, oh, cool. I will do that.
03:23:38  <dominictarr>testling seems to be stalling,
03:23:53  <substack>yes we've been running out of RAM on the servers
03:24:00  <dominictarr>oh, bummer
03:24:01  <substack>since people have started running a lot more tests
03:24:04  <rowbit>Hourly usage stats: [developer: 0, free: 7]
03:24:07  <dominictarr>well, that is good!
03:24:24  * jcrugzzquit (Ping timeout: 246 seconds)
03:24:34  <substack>it's cool, pkrumins figured out a rackspace instance that costs the same that has the same stats but is SSD backed
03:24:43  <substack>should speed up everything a crazy amount
03:25:10  <substack>more ram, backed to ssd, faster cpu, same price
03:25:29  <dominictarr>digital ocean is pretty cheap too
03:25:32  <dominictarr>with, ssd
03:25:55  * contrahaxquit (Read error: Connection reset by peer)
03:26:26  <substack>might be hard to get windows servers running there though
03:26:45  * contrahaxjoined
03:31:42  <dominictarr>hmm, yeah it looks like they only have linux
03:34:51  <chapel>could always use microsofts hosting platform
03:40:01  * contrahaxquit (Quit: Sleeping)
03:40:55  <substack>isaacs: any idea what's wrong with this package? https://npmjs.org/package/parse-color
03:42:18  <substack>it was published during the dark times of great instability
03:51:42  * fronxjoined
03:52:14  <kenperkins>substack: that was me recommending to pkrumins to use our new SSD backed flavorz
03:52:22  <kenperkins>s/flavorz/flavors
03:56:37  * fronxquit (Ping timeout: 272 seconds)
03:56:57  <substack>oh cool!
03:57:20  <kenperkins>also, I'd love to know how else I can help with making your provisioning layer awesomer
03:57:26  * nickleeflyquit (Quit: Connection closed for inactivity)
03:59:10  <substack>kenperkins: these machines boot up in 2 minutes, so that completely changes what we can do with them
03:59:27  <kenperkins>have you thought about keeping 1 or 2 in a cache so it's instant?
03:59:42  <kenperkins>and then provisioning the next one whenever someone signs up/creates a session?
03:59:45  <kenperkins>etc
03:59:46  <substack>we've just been running a relatively fixed pool
03:59:48  <kenperkins>i.e. ref counting
04:00:13  <kenperkins>are you integrating via API? or all by hand
04:00:20  <substack>mostly manually
04:00:24  <kenperkins>gotcha
04:00:33  <kenperkins>I do all of the openstack and rackspace API sdks in pkgcloud
04:00:35  <substack>they booted up too slowly before to do anything useful with the API
04:00:38  <kenperkins>so I'd love to help however I can
04:00:48  <substack>cool!
04:01:01  * wolfeidaujoined
04:05:31  * mikolalysenkoquit (Ping timeout: 246 seconds)
04:17:05  <thlorenz>defunctzombie_zz: substack done and done: https://twitter.com/thlorenz/status/416784738238873600
04:18:48  <substack>great!
04:20:36  <thlorenz>was thinking of grabbing errors from a test output stream
04:21:19  <thlorenz>that way this could work with any test module, just pipe it through and have the errors mapped, kinda like faucet, but different purpose
04:24:05  <rowbit>Hourly usage stats: [developer: 0, free: 25]
04:29:33  * dominictarrquit (Ping timeout: 272 seconds)
04:30:50  * jcrugzzjoined
04:36:09  * jcrugzzquit (Ping timeout: 272 seconds)
04:40:25  * mikolalysenkojoined
04:42:41  * justinabrahmsquit (Excess Flood)
04:43:39  * justinabrahmsjoined
04:46:17  * contrahaxjoined
04:49:27  <grncdr>ugh, need a special case grammar rule for command lists in back ticks
04:49:36  <grncdr>y u so crazy bash
04:49:57  * contrahaxquit (Client Quit)
04:51:34  * fronxjoined
04:52:45  * anvaka_joined
04:55:50  * fronxquit (Ping timeout: 240 seconds)
05:14:17  * mikolalysenkoquit (Ping timeout: 252 seconds)
05:15:32  * thlorenzquit (Remote host closed the connection)
05:24:05  <rowbit>Hourly usage stats: [developer: 0, free: 28]
05:24:22  * jcrugzzjoined
05:29:21  * jcrugzzquit (Ping timeout: 272 seconds)
05:38:56  <anvaka_>it would be nice if plunker supported browserify
05:39:24  <anvaka_>just a random thought
05:39:31  <substack>what is plunker?
05:39:39  <anvaka_>http://plnkr.co/edit/
05:40:11  <anvaka_>very quick way to prototype and share browser code
05:40:35  <anvaka_>like jsfiddle, but more flexible
05:40:56  <substack>anvaka_: have you seen http://requirebin.com/ ?
05:41:15  <anvaka_>substack: no, i haven't
05:41:57  <anvaka_>wow.. this is awesome
05:42:37  <anvaka_>hm... how do i edit html here
05:51:36  * fronxjoined
05:51:43  <anvaka_>alright, I didn't know this existed. ogd: it looks pretty nice. Still couldn't find a way how to edit html/css
05:55:10  * jcrugzzjoined
05:55:56  * fronxquit (Ping timeout: 246 seconds)
06:02:55  * jcrugzzquit (Ping timeout: 272 seconds)
06:03:27  * anvaka_quit (Ping timeout: 240 seconds)
06:03:41  * anvakajoined
06:24:04  <rowbit>Hourly usage stats: [developer: 0, free: 14]
06:51:41  * fronxjoined
06:56:13  * fronxquit (Ping timeout: 252 seconds)
07:24:04  <rowbit>Hourly usage stats: [developer: 0, free: 7]
07:27:21  * AvianFlujoined
07:35:27  * tilgoviquit (Ping timeout: 240 seconds)
07:45:09  * shamaquit
07:51:16  * pfrazequit (Ping timeout: 246 seconds)
07:51:39  * fronxjoined
07:55:39  * AvianFluquit (Remote host closed the connection)
07:56:23  * fronxquit (Ping timeout: 260 seconds)
08:24:04  <rowbit>Hourly usage stats: [developer: 0, free: 4]
08:31:25  * peutetrejoined
08:33:13  * contrahaxjoined
08:36:03  <rowbit>substack, pkrumins: These encoders are STILL down:
08:49:03  <rowbit>substack, pkrumins: These encoders are STILL down:
09:06:33  <rowbit>substack, pkrumins: Encoders down: (dev-ie6-1)
09:11:13  * indexzerojoined
09:24:04  <rowbit>Hourly usage stats: [developer: 0, free: 10]
09:25:14  * jcrugzzjoined
09:29:26  * jcrugzzquit (Ping timeout: 240 seconds)
09:42:04  * nickleeflyjoined
09:45:57  <substack>isaacs: how can we access individual files inside of npm packages now that they're on manta?
09:53:13  * fronxjoined
09:57:02  * ralphtheninjajoined
09:59:20  * indexzeroquit (Quit: indexzero)
10:24:04  <rowbit>Hourly usage stats: [developer: 0, free: 14]
10:35:53  * nick__joined
10:36:34  * nick__quit (Client Quit)
10:38:32  * fronxquit (Remote host closed the connection)
11:10:30  <feross>substack, you awake?
11:24:04  <rowbit>Hourly usage stats: [developer: 0, free: 10]
11:27:42  <feross>substack: please merge https://github.com/substack/node-browserify/pull/573 when you get a chance
11:28:43  <feross>it updates native-buffer-browserify to 2.0.0, which adds IE6-8 support, improves Firefox perf, and reduces bundle size by 40%
11:29:39  <feross>(had to go with a different approach than `typedarray` to get true IE6-8 support, i can tell you about it later)
11:31:20  * peutetrequit (Quit: peutetre)
11:44:26  * ferossquit (Quit: feross)
12:24:04  <rowbit>Hourly usage stats: [developer: 0, free: 15]
12:25:23  * gildeanchanged nick to gildean_
12:25:32  * gildean_changed nick to gildean
13:24:04  <rowbit>Hourly usage stats: [developer: 0, free: 9]
13:31:04  <rowbit>substack, pkrumins: Encoders down: (dev-ie7-4)
13:38:15  * dominictarrjoined
13:45:22  * yorickjoined
14:21:48  * fronxjoined
14:24:04  <rowbit>Hourly usage stats: [developer: 0, free: 8]
14:34:30  * mikolalysenkojoined
15:01:35  <rowbit>substack, pkrumins: These encoders are STILL down:
15:04:42  * fronxquit (Remote host closed the connection)
15:07:55  * jcrugzzjoined
15:22:57  * dominictarrquit (Ping timeout: 252 seconds)
15:24:04  <rowbit>Hourly usage stats: [developer: 0, free: 14]
15:30:15  * pfrazejoined
15:48:53  * fronxjoined
16:01:23  * thlorenzjoined
16:01:40  * fronxquit (Remote host closed the connection)
16:05:29  * mikolalysenkoquit (Ping timeout: 252 seconds)
16:08:47  * fronxjoined
16:09:53  * thlorenzquit (Ping timeout: 252 seconds)
16:12:03  * fronxquit (Remote host closed the connection)
16:14:43  * defunctzombie_zzchanged nick to defunctzombie
16:20:08  * fronxjoined
16:24:04  <rowbit>Hourly usage stats: [developer: 0, free: 13]
16:41:20  * fronxquit (Remote host closed the connection)
16:46:04  * fronxjoined
17:02:22  * mikolalysenkojoined
17:07:17  * mikolalysenkoquit (Ping timeout: 272 seconds)
17:10:18  <defunctzombie>ircretary: tell thlorenz that stack-mapper doesn't work in ie < 9
17:10:19  <ircretary>defunctzombie: I'll be sure to tell thlorenz
17:15:27  * thlorenzjoined
17:16:58  <thlorenz>defunctzombie: what's not working in IE8 with stack mapper?
17:17:09  <defunctzombie>thlorenz: dunno yet.. installing a VM right now
17:17:11  <defunctzombie>to debug
17:17:14  <thlorenz>I can't test this since testling has no servers for it
17:17:18  <defunctzombie>thlorenz: I have it working in chrome tho :)
17:17:22  <defunctzombie>thlorenz: modern.ie
17:17:27  <defunctzombie>thlorenz: it is awesome
17:17:35  <defunctzombie>thlorenz: stack trace contains the original source location
17:17:45  <defunctzombie>in the mocha results
17:17:56  <thlorenz>awesome! :)
17:18:03  <thlorenz>we can figure out IE8, could be some function I'm using, like reduce or so
17:18:08  <defunctzombie>it was very easy thanks for convert-source-map and stack-mapper :)
17:18:09  <thlorenz>maybe have to shim it
17:18:13  <thlorenz>:)
17:18:14  <defunctzombie>probably
17:18:22  <defunctzombie>are you using reduce or something?
17:18:29  <thlorenz>yep, just remembered
17:18:31  <defunctzombie>the only issue with shimming it in zuul
17:18:38  <defunctzombie>is that all the user's tests will get the shims
17:18:43  <defunctzombie>which I don't think I want to happen
17:19:06  <thlorenz>or I'll rewrite it imperative style ;)
17:19:07  <defunctzombie>unless.. I shim it and then unshim it haha
17:19:12  <defunctzombie>or that haha
17:19:44  <thlorenz>ok, so using reduce and forEach in places
17:19:59  <defunctzombie>yea, stack-mapper is one of those special module
17:20:03  <defunctzombie>we want to work without shims
17:20:23  <thlorenz>ok, let me try to rewrite 90s style
17:20:47  <defunctzombie>:)
17:21:38  <thlorenz>does IE8 have string.split?
17:21:44  <thlorenz>I sure hope so
17:21:44  <defunctzombie>hm
17:21:49  <defunctzombie>I think so
17:22:17  <defunctzombie>oh shit.. it occured to me that I was also using JSON.parse
17:22:18  <defunctzombie>haha
17:22:22  <defunctzombie>need to grap a json module
17:22:25  <defunctzombie>for old IE
17:22:27  <defunctzombie>sigh
17:22:39  <defunctzombie>man I hate IE
17:24:04  <rowbit>Hourly usage stats: [developer: 0, free: 9]
17:25:52  <thlorenz>defunctzombie: wow me too, gotta get rid of the filter too :(
17:26:01  <defunctzombie>heh
17:26:07  <defunctzombie>thlorenz: use shims, inline them
17:26:33  <thlorenz>not sure if that is best solution
17:26:35  <rowbit>substack, pkrumins: Encoders down: (dev-ie8-2)
17:26:52  <thlorenz>the work involved in adding and then undoing them maybe more than to just rewrite some parts ;)
17:28:08  <defunctzombie>adding them inline is easy
17:28:12  <defunctzombie>just grap the code from MDN
17:28:16  <defunctzombie>anyhow.. up to you
17:28:23  <defunctzombie>I just need it to work in IE :)
17:28:25  <thlorenz>almost done with the imperative way
17:29:54  <defunctzombie>thlorenz: shit
17:30:06  <defunctzombie>thlorenz: you will get more issues :)
17:30:13  <defunctzombie>thlorenz: stack-mapper uses source-map
17:30:18  <defunctzombie>which internally uses Array.isArray
17:30:33  <thlorenz>I can tell 'em to go back in time and file them with Microsoft
17:30:45  <thlorenz>ok, I have a simple shim for that
17:30:55  <defunctzombie>:)
17:31:02  <defunctzombie>I hate this crap
17:31:08  <defunctzombie>was so simple and awesome on chrome
17:31:15  <defunctzombie>but full of so much BS in IE
17:31:20  <thlorenz>yay, my nodejsreaction got posted: http://nodejsreactions.tumblr.com/post/71426391619/when-the-discussion-turns-to-build-tools-or-front-end
17:31:49  <defunctzombie>hahaha
17:31:54  <defunctzombie>amazing
17:32:25  <thlorenz>:)
17:32:34  <defunctzombie>thlorenz: got a module with tape tests for frontend?
17:32:39  <defunctzombie>so I can test the tape stuff?
17:33:45  <thlorenz>yep, hold on
17:34:13  <thlorenz>here is one: https://github.com/thlorenz/get-style-property/tree/master/test
17:34:26  <defunctzombie>also.. what was that module substack mentioned to parse tap output and get events?
17:34:30  <defunctzombie>like passing/failing?
17:34:57  <thlorenz>he wrote faucet
17:35:13  <thlorenz>is that what you mean?
17:35:15  <defunctzombie>no
17:35:25  <defunctzombie>I need just something that turns a tap output
17:35:28  <defunctzombie>into events
17:35:34  <defunctzombie>faucet was console pretty output
17:35:40  <defunctzombie>tap-stream or something like that?
17:36:48  <thlorenz>defunctzombie: faucet uses https://github.com/substack/tap-parser
17:36:56  <defunctzombie>yes, that one
17:36:58  <defunctzombie>I think
17:36:59  <defunctzombie>ok
17:38:12  <defunctzombie>need to figure out how to get errors from that
17:43:33  <thlorenz>defunctzombie: you can use what's on master to test
17:43:42  <defunctzombie>thlorenz: k, will try that
17:43:44  <defunctzombie>thanks
17:43:58  <thlorenz>ugly but should work now: https://github.com/thlorenz/stack-mapper/commit/8006006a627fb0ac50cf8f15758db39a99955967
17:44:14  <defunctzombie>hopefully works in ie6 too :)
17:44:20  <thlorenz>hold on a sec
17:44:33  <thlorenz>I guess I gotta put that Array.is .. there still
17:46:28  * AvianFlujoined
17:49:08  <thlorenz>defunctzombie: shoot still failing in IE8, even with isarray shim
17:49:14  <defunctzombie>:)
17:50:30  <thlorenz>defunctzombie: it's amddefine using map
17:50:42  <defunctzombie>haha
17:50:45  <defunctzombie>amd rofl
17:50:53  * fronxquit (Remote host closed the connection)
17:50:54  <thlorenz>no clue why the sourcemap guys made things so complex
17:51:00  <thlorenz>ok, I'll shim map
17:52:11  <defunctzombie>wow
17:55:12  <defunctzombie>substack: how do I get the stacktraces from tape assertions?
17:55:18  <defunctzombie>via tap-parser or something
17:55:47  * mikolalysenkojoined
17:57:42  <defunctzombie>so the really annoying thing about tape in the browser is that it uses console.log for output
17:58:10  <defunctzombie>which means that if I output tape shit to console log for debugging and am using a parser on console log then it goes into a loop
17:58:50  <defunctzombie>bah
18:02:28  * thlorenzquit (Remote host closed the connection)
18:02:43  * thlorenzjoined
18:04:59  <thlorenz>defunctzombie: ok at this point I'm clue less: https://ci.testling.com/thlorenz/stack-mapper
18:05:22  <thlorenz>IE8 fails on different line than IE7 and IE6, but in all cases they point to empty line
18:05:29  <defunctzombie>heh
18:05:32  <defunctzombie>I have an IE 8 vm
18:05:36  <defunctzombie>I can try to run in that
18:06:27  <thlorenz>cool, lmk what the code is, it's all in amddefine btw ;)
18:06:46  <defunctzombie>can we just fork and remove that shit?
18:06:48  <defunctzombie>what is it for?
18:06:48  <thlorenz>in the worst case I'll pull out the SourcemapConsumer from source map and publish it separately
18:06:57  <defunctzombie>yea, why won't we just do that?
18:07:03  <thlorenz>it's cause they use amd all over the place
18:07:19  <defunctzombie>it will let us have a smaller bundle too
18:07:22  <defunctzombie>in zuul
18:07:24  <defunctzombie>which is good
18:07:32  <thlorenz>yep, but we won't get upstream fixes
18:07:32  <defunctzombie>cause this unused crap bloats the bundle
18:07:38  <defunctzombie>sure we will
18:07:50  <defunctzombie>when we check the project in a month
18:07:53  <defunctzombie>we can get any fix we want
18:07:54  <thlorenz>at least not easily since we'll restructure that code like crazy
18:08:00  <defunctzombie>hm
18:08:01  <defunctzombie>sigh
18:08:07  <defunctzombie>we can't just rip it out?
18:08:12  <defunctzombie>I guess I haven't looked at it yet
18:08:20  <defunctzombie>probably doesn't use commonjs or anything normal
18:08:21  <thlorenz>the consumer is pretty involved
18:08:59  <thlorenz>I once tried to rip parts of it out, but it ended up being buggy: https://github.com/thlorenz/parse-base64vlq-mappings
18:09:15  <thlorenz>since I didn't get upstream changes in easily
18:09:25  <thlorenz>still used in browerpack tests though ;)
18:09:32  <defunctzombie>ah
18:10:04  <defunctzombie>so why do they have this amd shit then?
18:10:10  <defunctzombie>cause that is what they use for module system?
18:10:12  * thlorenzshrugs
18:10:22  <thlorenz>yep
18:10:36  <thlorenz>so I can fork the entire thing and rip out amddefine
18:10:49  <thlorenz>but I won't separate consumer from producer -- too much work
18:11:17  <thlorenz>but we are only pulling the consumer anyhow since stackmapper gets it directly: https://github.com/thlorenz/stack-mapper/blob/master/lib/setup-consumer.js#L3
18:12:58  <defunctzombie>thlorenz: will that be enough?
18:13:07  <defunctzombie>does anything call run_factory
18:13:11  <defunctzombie>runFactory
18:13:14  <thlorenz>yep, all we need to get rid of is amddefine
18:13:15  <defunctzombie>has a .map call in it
18:13:18  <thlorenz>yep, amddefine ;)
18:13:21  <defunctzombie>jesus
18:13:25  <thlorenz>well .map is shimmed
18:13:25  <defunctzombie>what calls amddefine tho?
18:13:32  <thlorenz>sourcemap :)
18:13:43  <thlorenz>but I'm about to rip it out, forked already
18:13:51  <defunctzombie>why does it call it tho
18:13:55  <defunctzombie>is this like initial setup crap?
18:14:02  <defunctzombie>or does it actually do something?
18:19:05  <rowbit>substack, pkrumins: Encoders down: (free7)
18:19:52  <thlorenz>ok done
18:20:06  <thlorenz>defunctzombie: I replaced the dep in stack mapper with my fork for now
18:20:14  <thlorenz>lets see how that works
18:20:16  <defunctzombie>k
18:21:05  <defunctzombie>thlorenz: still not passing
18:21:18  <defunctzombie>thlorenz: you really should use zuul or something to test this before committing :)
18:21:23  <defunctzombie>this is the annoying thing with testling
18:21:28  <defunctzombie>you commit just to see if it works :p
18:22:03  <thlorenz>that's ok for me -- I don't have IE on my machine ;)
18:22:26  <thlorenz>and am not about to set up a VM and install windows
18:23:22  <defunctzombie>right.. that is why you use zuul so you don't have to :)
18:23:29  <defunctzombie>or just get free ones from modern.ie
18:23:32  <defunctzombie>that is what I do
18:23:35  <defunctzombie>takes 5 minutes
18:23:36  <defunctzombie>anyhow
18:23:39  <defunctzombie>base64 module
18:23:41  <defunctzombie>uses forEach
18:24:03  <defunctzombie>dunno what uses base64
18:24:04  <rowbit>Hourly usage stats: [developer: 0, free: 12]
18:24:36  <thlorenz>defunctzombie: ok I'll shim that, that's sourcemap
18:25:17  <defunctzombie>thlorenz: the what you shim is not good btw
18:25:21  <defunctzombie>you are shimming globally
18:25:27  <defunctzombie>this is exactly what I wanted to avoid
18:25:32  <thlorenz>how else am I gonna do it?
18:25:44  <thlorenz>every time you ask for a map?
18:25:47  <defunctzombie>you gotta unset the shim I think
18:25:48  <defunctzombie>yea
18:25:50  <thlorenz>ok
18:25:57  <defunctzombie>otherwise if a user has something with Array.isArray in their code
18:26:00  <defunctzombie>it will pass in zuul
18:26:04  <defunctzombie>but not work in actual usage
18:26:08  <thlorenz>got it
18:26:14  <thlorenz>let me fix that and add foreach
18:26:34  <defunctzombie>but only remove it if you had to shim it
18:26:38  <defunctzombie>and only shim it if you have to :)
18:31:23  <defunctzombie>thlorenz: we have more problems ahhaha
18:31:27  <defunctzombie>SourceMapConsumer
18:31:31  <defunctzombie>uses Object.defineProperty
18:31:32  <defunctzombie>rofl
18:31:44  <thlorenz>I'll rip that right out
18:31:54  <defunctzombie>haha
18:32:19  <defunctzombie>jesus... maybe we don't even care about old IE but damn it will make tracking down errors easier
18:32:20  <defunctzombie>in the future
18:32:21  <defunctzombie>in tests
18:32:29  <defunctzombie>for all those poor souls on IE
18:32:38  <defunctzombie>I don't even support IE 9 in my sutff
18:32:49  <thlorenz>don't worry we'll figure it out
18:34:51  <defunctzombie>out life is difficult
18:34:53  <defunctzombie>*our
18:40:23  * mikolalysenkoquit (Ping timeout: 272 seconds)
18:43:29  <thlorenz>wow not that easy to rip out the defineProperty stuff
18:43:38  <thlorenz>maybe we oughta just shim that as well
18:44:22  * mikolalysenkojoined
18:50:39  <jjjohnny_>return zulu
18:51:44  <defunctzombie>thlorenz: can't shim defineProperty
18:51:55  <thlorenz>yeah you could, but it's ugly
18:52:17  <thlorenz>so I'm gonna try to rip it out correctly, but looks like it's gonna be an after lunch affair ;)
18:52:34  <defunctzombie>haha kk
18:54:07  <thlorenz>shoot it will change the API though :(
18:54:23  <defunctzombie>does it matter?
18:54:26  <thlorenz>since you'll have to do consumer.sources() instead of consumer.sources
18:54:28  <defunctzombie>do you even use that api?
18:54:44  <defunctzombie>also.. I made zuul try/catch on require stack-mapper
18:54:55  <thlorenz>not to me, but yes I use some of it and if I fork and publish it'll have to work
18:54:56  <defunctzombie>so if something ever does fail in a browser it won't make all other tests not run
18:55:10  <defunctzombie>it will just not give sourcemapped results
18:55:15  <thlorenz>actually you should wrap stackmapp(map)
18:55:22  <thlorenz>and when you call sm.map on it
18:55:30  <defunctzombie>had to wrap the require actually
18:55:34  <thlorenz>but if you want I can add try/catch around it
18:55:42  <defunctzombie>na, that is ok
18:55:53  <thlorenz>ok, will continue after lunch
18:55:58  <defunctzombie>I am trying to figure out if IE8 even gives error stacks
19:00:00  * mikolalysenkoquit (Ping timeout: 245 seconds)
19:08:46  <defunctzombie>I think we can ignore old browsers actually cause we don't even get error.stack from them
19:08:57  * thlorenzquit (Remote host closed the connection)
19:11:09  <defunctzombie>actually I think I can get the filename and lineno
19:18:18  <substack>https://github.com/substack/node-browserify/issues/574
19:18:41  <substack>make sure to click the link, I mis-interpreted what was meant at first
19:19:37  <defunctzombie>I minify after I have the whole bundle
19:19:47  <defunctzombie>cause minification is better when you have the whole thing
19:19:52  <defunctzombie>and also I rip out sourcemap stuff
19:20:00  <defunctzombie>so I can get sourcemaps on compressed output
19:20:19  <substack>defunctzombie: click the uglify link
19:20:25  <substack>https://github.com/hughsk/uglifyify
19:20:27  <defunctzombie>I did
19:21:22  <defunctzombie>what about it?
19:21:39  <substack>it can prune required modules from the graph
19:22:19  <defunctzombie>you mean the fact that you can do that
19:22:24  <defunctzombie>not the actual fact that it minifies
19:22:37  <defunctzombie>cause the fact that it minifies I don't care about unless it can support sourcemaps
19:22:45  <defunctzombie>but pruning might be nice
19:22:51  <defunctzombie>but I think their examples are all bad
19:23:17  <defunctzombie>the first one is supper contrived
19:23:35  <defunctzombie>and the env one I dunno.. not sure I am a fan of that approach
19:23:54  <defunctzombie>maybe
19:24:04  <rowbit>Hourly usage stats: [developer: 0, free: 13]
19:25:35  <rowbit>substack, pkrumins: These encoders are STILL down:
19:26:46  <defunctzombie>why are there no good frontend testing libs/harnesses?
19:26:57  <defunctzombie>mocha doesn't give me a stacktrace with assertions for some reason
19:27:09  <defunctzombie>and I can't get stacks from tape output in any sensible way
19:32:56  <defunctzombie>substack: I think their examples for production stuff would only be used in top level app code anyway
19:33:07  <defunctzombie>which means your transform would be run on those files now even
19:33:20  <defunctzombie>at least that is my understanding
19:34:14  <substack>defunctzombie: tape gives the first non-itself line from the stack trace
19:34:30  <substack>I can add a configuration option to expand the entire trace
19:34:40  <defunctzombie>substack: how do I get that tho?
19:34:47  <defunctzombie>I tried using tap-parser
19:34:53  <defunctzombie>but maybe I need something higher level
19:35:09  <substack>defunctzombie: when you get an assertion that is !res.ok, just read the 'extra' events
19:35:14  <substack>that's what faucet does
19:35:31  <defunctzombie>I guess I need to make something that emits a "test" event
19:35:43  <defunctzombie>after it has read an assertion and any followup extra events for it
19:35:57  <defunctzombie>I would like it all in one simple event I can just use hehe
19:36:06  <defunctzombie>without having to further parse extra
19:36:20  <defunctzombie>since tape uses specific extra fields, but I think I can do that
19:36:30  <defunctzombie>also.. is there a way to run tape and give it an output stream?
19:36:33  <substack>if you want to hack that into an extra event for tap-parser I'm open to it
19:36:42  <defunctzombie>so it doesn't write to console.log?
19:36:51  <substack>yes you can pass in a custom stream as a parameter
19:36:56  <defunctzombie>like require('tape')(stream) or something?
19:37:34  * tilgovijoined
19:37:40  <defunctzombie>substack: didn't see that in the docs
19:37:48  <substack>var t = require('tape').createHarness(); t.createStream().pipe(wherever)
19:38:08  <defunctzombie>substack: but test files won't get that harness tho
19:38:15  <substack>or rather, t => test
19:38:21  <substack>test files?
19:38:26  <defunctzombie>within test files people do var test = require('tape');
19:38:28  <defunctzombie>right
19:38:33  <defunctzombie>which will have a different harness?
19:38:33  <substack>correct
19:38:41  <substack>so just capture stdout from the process
19:38:44  <substack>and pipe it somewhere
19:38:50  <defunctzombie>right.. I am saying I would like to not do that :)
19:39:20  <defunctzombie>but there is no easy way to not do that it seems?
19:39:25  <substack>correct
19:39:29  <substack>just spawn a process!
19:39:35  <substack>that is actually the easy way to do this
19:39:45  <defunctzombie>this is all in browser
19:39:50  <defunctzombie>this has nothing to do with processes
19:39:52  <substack>then do
19:40:00  <defunctzombie>the reason I don't want to write to console.log
19:40:10  <defunctzombie>is cause I just want to capture tap output separately
19:40:18  <defunctzombie>without having to filter it out of console.log
19:40:28  <substack>var original = console; console = { log: function (msg) { whatever(); original.log(msg) } }
19:41:26  <defunctzombie>also annoying
19:41:34  <defunctzombie>but I guess no one has a good answer to this
19:41:46  <substack>I'm not sure what you're after
19:42:25  <substack>why don't you just use the browser field to provide a custom 'tape' that sets up the harness with a stream?
19:42:44  <defunctzombie>substack: yea, that sounds like it could be a win
19:43:41  <defunctzombie>browser testing stuff is weird... I like browser stuff that lets me hook into the launching of the harness/suites
19:44:19  <defunctzombie>kinda different than server side where you just write to stdout, even there no reason that the testing lib has to know anything about how stuff is written out
19:44:24  <defunctzombie>it just has to provide a way to run it
19:44:34  <defunctzombie>anyhow.. I think I can add a 'test' event to tap-parser
19:44:36  <substack>console.log() works in both the browser and in node is why
19:44:39  <defunctzombie>that will solve some things
19:46:07  <defunctzombie>see there sorta is high coupling here
19:46:16  <defunctzombie>cause in order for tap-parser to parse extra for stuff like stack
19:46:21  <defunctzombie>or at
19:46:36  <defunctzombie>the tap output has to do that
19:48:38  <defunctzombie>and now parsing extra stuff becomes complicated
19:49:51  <defunctzombie>and when does a tap test start?
19:49:53  * jcrugzzquit (Ping timeout: 252 seconds)
19:53:24  <defunctzombie>substack: https://gist.github.com/defunctzombie/8163418
19:54:14  <defunctzombie>this is one of the reasons why I don't like just parsing console.log (and in browser) cause now for extra field stack capture, you have to do all sorts of detection to get all the meaningful info
19:54:21  <defunctzombie>which tape had in the first place
19:55:26  <defunctzombie>its great when you can just output to console.log for someone to read, but if you actually want that metadata, turning it into text form makes it harder to do some things
19:56:19  * mikolalysenkojoined
20:00:53  * mikolalysenkoquit (Ping timeout: 252 seconds)
20:06:11  <defunctzombie>ircretary: tell thlorenz that stack-mapper doesn't work on firefox stack traces :)
20:06:11  <ircretary>defunctzombie: I'll be sure to tell thlorenz
20:20:42  * thlorenzjoined
20:24:04  <rowbit>Hourly usage stats: [developer: 0, free: 24]
20:24:06  * kevino80joined
20:25:39  <thlorenz>defunctzombie: wow, thought stackmapper would be this one day kinda project
20:25:48  <defunctzombie>ahhaha
20:25:53  <defunctzombie>you should have known better
20:25:53  <thlorenz>can you file some issues with sample stacktraces for ff?
20:26:00  <defunctzombie>when dealing with client side error stacks
20:26:05  <defunctzombie>thlorenz: anyhow.. I am hacking around it
20:26:10  <thlorenz>we'll just tackle one problem at a time :)
20:26:14  <defunctzombie>by using tracekit
20:26:22  <defunctzombie>to get the stacktrace from the error object
20:26:25  <thlorenz>it'll be awesome in a month
20:26:31  <defunctzombie>and then making the string that stack-mapper needs
20:26:46  <defunctzombie>since it seems stack mapper is assuming v8 style stack string
20:26:51  <thlorenz>awesome then we'll create tests for each of those
20:27:00  <defunctzombie>I think we should have stack mapper take a js object
20:27:09  <defunctzombie>to not have to worry about string crop
20:27:10  <defunctzombie>*crap
20:27:16  <thlorenz>well it's what the underlying stack-trace module assumes
20:27:25  <defunctzombie>yea.. I think we can remove that module
20:27:30  <defunctzombie>and push that onto the user
20:27:37  * kevino80quit (Client Quit)
20:27:37  <defunctzombie>makes stack-mapper simpler
20:28:14  <thlorenz>ok, just lmk when you have something figured out -- btw taking string seemed to be the most generic ,not sure why taking Error is better
20:28:54  <defunctzombie>no no
20:28:57  <defunctzombie>don't take error
20:29:00  <defunctzombie>take an array
20:29:06  <defunctzombie>and we can define the array format
20:29:12  <defunctzombie>{ line:, column:, etc)
20:29:53  <defunctzombie>also.. what happens if I don't have a column number?
20:29:55  <defunctzombie>in the stack?
20:30:21  <thlorenz>ah, so take something similar to what stack-trace module provides
20:30:38  <thlorenz>and put other modules in the middle that take care of the parsing?
20:30:57  <thlorenz>if you don't have column we can make it 0
20:30:57  <defunctzombie>yep
20:31:05  <defunctzombie>I can make it 0
20:31:16  <defunctzombie>was just curious if it is needed, seems to be
20:31:17  <thlorenz>ok that'll work
20:31:21  <defunctzombie>and yea, remove stack-trace
20:31:28  <defunctzombie>and leave that up to user
20:31:38  <defunctzombie>that will be more cross platform friendly
20:31:38  <thlorenz>right now it is, but just cuz stack-trace spat that out at all times
20:32:13  <thlorenz>ok, and I guess then we'll write an overarching module that sniffs the UA and figures out what error parse module to use
20:33:09  <defunctzombie>yea.. tracekit kinda does that
20:33:23  <defunctzombie>so trying to use that right now to generate a string to feed into stack-mapper haha
20:33:30  <defunctzombie>until stack-mapper takes an array
20:33:36  <thlorenz>ok, cool in the meantime I'll hack to get the defineProperty gone
20:33:43  <defunctzombie>once this all works it will be quite impressive
20:33:52  <defunctzombie>thlorenz: I dunno if I would focus on old browsers
20:33:57  <defunctzombie>thlorenz: we don't get stacks with those even
20:34:06  <thlorenz>wut?
20:34:07  <defunctzombie>thlorenz: at best we get where the assertion happened
20:34:15  <defunctzombie>thlorenz: yea, ie error object doesn't have a stack
20:34:19  <defunctzombie>for old IE
20:34:25  <defunctzombie>so right now not even worth it
20:34:26  <thlorenz>well I'll make it work and pub, to call it done
20:34:35  <defunctzombie>to battle that crap
20:35:09  <thlorenz>it's just stupid work, but there are only 3 define props in places -- just piece meal work
20:35:14  <defunctzombie>heh
20:36:07  <defunctzombie>not having the right column seems to make bad things happens in stack-mapper
20:37:15  <defunctzombie>faking with column 1 for now... column 0 doesn't work
20:37:48  <thlorenz>huh? it shouldn't - will have to look at this, do you have `includeSource = true`?
20:38:25  <defunctzombie>no
20:38:28  <defunctzombie>I have that to false
20:38:41  <defunctzombie>should that be true?
20:38:44  <thlorenz>ok, then I don't know why it matters (maybe for the source map module)
20:38:56  <thlorenz>only if you want to add source to the first trace
20:39:29  <thlorenz>what really sucks about this source-map module, I can't just run one test file :(
20:39:44  <defunctzombie>?
20:39:46  <defunctzombie>why not
20:40:10  <defunctzombie>thlorenz: https://github.com/defunctzombie/zuul/tree/sourcemap
20:40:19  <defunctzombie>thlorenz: that is my hacking so far today
20:40:28  <defunctzombie>use it on a mocha test for now
20:40:58  <defunctzombie>oh yea.. and you will need to update browser-pack in browserify to my branch with the prelude sourcemap
20:41:11  <defunctzombie>https://github.com/defunctzombie/browser-pack/tree/prelude-source-map
20:41:18  <defunctzombie>so this will let you test some stuff end to end
20:41:21  <defunctzombie>it is pretty cool when it works
20:41:32  <defunctzombie>cause you get stacktrace to the proper place in the original source file
20:41:40  <defunctzombie>regardless if browser supports sourcemaps
20:42:52  <thlorenz>cool, can't see in zuul where you call .map on the mapper
20:43:13  <thlorenz>nm found it
20:44:01  <defunctzombie>it is in the mocha shit for now
20:44:11  <defunctzombie>will be separate file so all the frameworks can use it eventually
20:44:12  <thlorenz>so yeah that should work - we'll have to figure out what's wrong with the source maps and why column 0 doesn't work
20:44:15  <thlorenz>cool
20:44:33  <defunctzombie>column 0 doesn't work, column 1 kinda makes it work
20:44:38  <defunctzombie>I hate browsers
20:44:57  <defunctzombie>and the better kicker is that the assert module
20:45:00  <thlorenz>oh, but looks like you are pulling stack-mapper from npm, should get the latest master
20:45:03  <defunctzombie>has no stack trace AT ALL on firefox
20:45:09  <thlorenz>wow
20:45:09  <defunctzombie>thlorenz: I had it linked locally
20:45:13  <thlorenz>ok
20:45:26  <defunctzombie>yea, because I use captureStackTrace
20:45:39  <defunctzombie>thlorenz: https://github.com/defunctzombie/commonjs-assert/blob/master/assert.js#L58
20:45:45  <defunctzombie>and I have no clue what to do about this
20:45:55  <defunctzombie>substack: any tips?
20:46:40  <thlorenz>defunctzombie: that may only be supported by some browsers: https://code.google.com/p/v8/wiki/JavaScriptStackTraceApi#Customizing_stack_traces
20:47:02  <thlorenz>maybe not on FF?
20:47:38  <defunctzombie>thlorenz: it is v8 only actually
20:47:49  <thlorenz>:(
20:48:28  <thlorenz>defunctzombie: sad: http://stackoverflow.com/questions/10484123/how-can-i-get-a-javascript-stack-trace-in-firefox
20:49:18  <thlorenz>defunctzombie: maybe this would help? https://github.com/stacktracejs/stacktrace.js
20:49:54  <defunctzombie>sigh
20:50:05  <thlorenz>yep, this all kinda sucks
20:50:06  <defunctzombie>I just dunno if I want to throw that into commonjs assert
20:50:33  <thlorenz>prob. not, but maybe you should allow extending it with an alternate way to get stack trace
20:56:21  * jcrugzzjoined
20:59:00  <defunctzombie>well, I kinda have something that could be a workaround for browsers without captureStackTrace
21:01:01  * jcrugzzquit (Ping timeout: 252 seconds)
21:01:36  <rowbit>substack, pkrumins: These encoders are STILL down:
21:07:04  <thlorenz>defunctzombie: so the source-map I have on master now works without properties :)
21:07:13  <defunctzombie>thlorenz: nice
21:07:27  <defunctzombie>thlorenz: also.. seems that stack is a read only property on safari
21:07:34  <defunctzombie>I want this pain to be over
21:07:45  <thlorenz>well you don't have to rewrite it right?
21:07:54  * mikolalysenkojoined
21:08:17  <thlorenz>I give you an error and you give me the traces w/out doing anything to the error
21:08:18  <defunctzombie>thlorenz: I was re-writing it so that the stack you get on the browser from mocha looked correct
21:08:24  <defunctzombie>but whatever, fuck that shit for now
21:08:42  <defunctzombie>the re-write was cool cause it made the output in the document the useful stack
21:08:44  <thlorenz>we just need to get a minimal working thing going
21:08:47  <defunctzombie>yea
21:08:57  <defunctzombie>so next is to have source-mapper take an array
21:09:06  <defunctzombie>and no longer rely on string of shit
21:09:16  <defunctzombie>that is a technical term
21:09:27  <thlorenz>so I'll pub the source-map as source-map-cjs so we can use it directly from npm -- no way the'll pull my changes ;)
21:09:48  <thlorenz>ok then make it rely on an array of shit
21:09:53  <thlorenz>also technical term
21:15:48  <defunctzombie>excellent
21:21:42  * daleharvey_changed nick to daleharvey
21:23:32  <substack>defunctzombie: your changes in browser-pack are live in 3.16.0
21:24:00  <defunctzombie>substack: oh boiii
21:24:04  <rowbit>Hourly usage stats: [developer: 0, free: 23]
21:24:25  * jcrugzzjoined
21:24:29  <defunctzombie>substack: we are doing some crazy science which will probably be useful for testling too
21:24:58  <substack>great!
21:25:14  <substack>I'm revamping browser-launcher so it isn't so opaque
21:25:30  <substack>and then I'm going to set it up so you can plug into the testling backend
21:25:45  <substack>so we can use that in the testling module
21:26:08  <defunctzombie>we need to find a way to fix commonjs-assert module
21:26:16  <defunctzombie>I have a hack right now for it
21:26:25  <defunctzombie>but basically, anything not v8 gives us no stacktrace
21:26:29  <defunctzombie>this is sadness
21:27:19  <jesusabdullah>I just bought "Hobo With A Shotgun"
21:28:30  <thlorenz>defunctzombie: ok we can use this now: https://npmjs.org/package/source-map-cjs
21:28:34  <grncdr>whoa, bash can redirect to TCP and UDP natively.
21:28:51  <defunctzombie>thlorenz: hahaha
21:28:56  <defunctzombie>I love the strikeouts on the readme
21:28:58  <defunctzombie>amazing
21:28:59  <thlorenz>:)
21:29:11  <grncdr>like echo "some stuff" > /dev/tcp/myhost/1337
21:29:42  <grncdr>jesusabdullah: is that a movie?
21:29:48  <jesusabdullah>sure is!
21:30:49  <grncdr>hey it's canadian!
21:30:55  <grncdr>and Rutger Hauer!
21:31:14  <thlorenz>defunctzombie: I'll update stack-mapper to use it, any changes you made to it that you want me to pull in?
21:31:23  <substack>grncdr: how do you get the responses back?
21:31:43  <defunctzombie>thlorenz: nope
21:31:55  <defunctzombie>thlorenz: all of the browser stuff was in sourcemap
21:32:02  <grncdr>substack: I have no idea! going to play with it
21:32:03  <thlorenz>ok I'll optimize a bit wrt to adding sources to the trace
21:32:25  <grncdr>was just looking up a reference for what kind of file redirections are allowed and bam! http://www.gnu.org/software/bash/manual/html_node/Redirections.html
21:32:29  <defunctzombie>I am still trying to figure out the good way to have stack traces on assertions in firefox
21:32:32  * mikolalysenkoquit (Ping timeout: 246 seconds)
21:34:45  <substack>grncdr: oftentimes the best way to learn something is to reimplement it from scratch
21:35:27  <grncdr>well, it would probably be mostly trivial to add to the interpreter
21:35:48  <grncdr>not sure whether the parser actually needs to do anything special there, probably not
21:35:54  <grncdr>but neat all the same
21:36:08  <grncdr>so far I am able to cause ECONNRESET errors :P
21:51:01  <thlorenz>defunctzombie: I improved shimming, but OMG IE
21:51:07  <defunctzombie>hahaha
21:51:09  <defunctzombie>I know
21:51:10  <thlorenz>6 and 7 still fail on an empty line
21:51:16  <defunctzombie>IE is a lost cause
21:51:17  <defunctzombie>hahaha
21:51:52  <thlorenz>and 8 seems to have problems with at string: 'ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/' -> "Error: Object doesn't support this property or method on line 5534"
21:52:04  <thlorenz>that's just sad
21:52:44  <thlorenz>defunctzombie: it may become simple if you decide to not use stack-trace, then we can just focus on making sourcemap module IdiotE compliant
21:53:00  <defunctzombie>haha
21:53:06  <defunctzombie>yea.. I saw that error
21:53:10  <defunctzombie>I thought it was the forEach
21:53:12  <defunctzombie>on the next line
21:53:21  <thlorenz>I'm shimming that now though
21:53:28  <defunctzombie>I dunno what to tell you
21:53:40  <defunctzombie>:)
21:53:45  <thlorenz>and lineno points to the string
21:53:48  <defunctzombie>yep
21:54:05  <thlorenz>oh well we'll just publish it and wait for Domenic_ to come and fix it ;P
21:54:11  <defunctzombie>mabye can't have newline before method call? I have no idea
21:54:17  <defunctzombie>this IE shit is out of control
21:54:27  <defunctzombie>like I said we don't even get stacktraces for IE
21:54:34  <thlorenz>defunctzombie: that could be actually, I can fix that
22:06:17  * jcrugzzquit (Ping timeout: 252 seconds)
22:06:24  <grncdr>> >(tr a-z A-Z) echo what
22:07:05  <grncdr>so you can have IO redirection in any position, which is neat
22:07:10  <grncdr>but also crazy
22:10:05  * anvakaquit (Remote host closed the connection)
22:14:52  <defunctzombie>thlorenz: ok, I have a hack fix for stacktraces for the assertion stuff too
22:21:48  <thlorenz>awesome!
22:22:09  <thlorenz>defunctzombie: giving up on the explorers of this world, lets focus on browsers for now
22:22:39  <defunctzombie>yes
22:22:40  <defunctzombie>agreed
22:23:06  <thlorenz>it keeps pointing to .split although it should be there for IE8, would need to debug in IE but testling has no server for it
22:23:31  <defunctzombie>thlorenz: I have a VM up for it
22:23:40  <thlorenz>I improved the shimming define/undefine and optimized some things, so should be faster now and take less memory
22:23:45  <defunctzombie>but debugging that is kinda crazy
22:24:00  <thlorenz>defunctzombie: I know it's ok for now -- we'll figure that out later
22:24:04  <rowbit>Hourly usage stats: [developer: 0, free: 9]
22:25:16  <thlorenz>defunctzombie: so gonna go back to libuv for now, lmk if you need anything
22:29:05  * mikolalysenkojoined
23:03:54  <defunctzombie>thlorenz: array stuff for stack-mapper
23:03:57  <defunctzombie>thlorenz: want me to PR that?
23:04:36  <thlorenz>sure, I'll also add you to the repo
23:05:26  <defunctzombie>woooo
23:07:26  <substack>thlorenz: I'm looking into global transforms right now
23:07:40  <substack>one problem is the ordering
23:08:03  <thlorenz>well only the root package should be allowed to have global transforms
23:08:07  <substack>yes
23:08:09  <thlorenz>others should be ignored
23:08:28  <thlorenz>then run the local transforms of each packages first and finally the global ones in the order they were specified
23:08:39  <thlorenz>wouldn't that be deterministic?
23:09:53  <substack>but sometimes you might want to run a transform before everything else
23:10:14  <thlorenz>I wouldn't allow that since there is no way of telling what may happen
23:10:24  <substack>like for https://github.com/substack/node-browserify/issues/574
23:10:35  <thlorenz>i.e. local transforms would break horribly cuz they get different stuff than they expect
23:10:48  <substack>oh actually post global transforms would work for that use-case still I think
23:11:10  <thlorenz>yeah, he could uglify afterwards right?
23:11:25  <thlorenz>also he should just minimize the bundle itself
23:12:35  * jcrugzzjoined
23:12:39  <substack>that use-case just uses the minification step to prune requires
23:12:50  <substack>thlorenz: https://github.com/hughsk/uglifyify
23:13:07  <robertkowalski>mmalecki: that bitswipe thingy you build... is that usable for payments? would like to connect my app to it
23:13:40  <thlorenz>substack: but if that runs first, browserify couldn't follow the requires ??
23:13:53  <thlorenz>unless that is what is desired
23:14:21  <thlorenz>weird edgecase I'd think, maybe allow both then 'before
23:14:26  <thlorenz>and after local
23:14:35  <thlorenz>use at your own risk ;)
23:14:43  <substack>it can run at the end of the transform list
23:14:54  <thlorenz>the local list?
23:15:21  <thlorenz>but how is that different from a global transform that gets called after all local ones ran?
23:16:36  <thlorenz>substack: I think we should go for always running all local txs first and only if there are real use cases that don't work otherwise should we consider global txs that run before the locals
23:17:47  * jcrugzzquit (Ping timeout: 252 seconds)
23:18:57  <thlorenz>I'm thinking of: 1.resolve file 2. run local transforms on file 3. pass result to global transforms if present 4. look for added requires in the file 5. start over
23:20:21  * zz_runningskullquit (Read error: Operation timed out)
23:21:06  <rowbit>substack, pkrumins: These encoders are STILL down:
23:24:03  <rowbit>Daily usage stats: [developer: 0, free: 312]
23:24:04  <rowbit>Hourly usage stats: [developer: 0, free: 10]
23:25:47  * zz_runningskulljoined
23:29:57  <hughsk>substack, thlorenz: +1. Means I could implement stuff like http://ghub.io/disc as transforms in the future :)
23:30:54  <thlorenz>hughsk: but also in your case the global transforms could run after all the locals for the file ran right?
23:33:07  <hughsk>thlorenz: Yeah, locals->globals->resolve fits with all of the cases I can think of
23:44:31  <thlorenz>defunctzombie: I think I figured it out: https://github.com/thlorenz/source-map-cjs/blob/master/lib/source-map/base64.js#L11-L21 was running before any of the shims got applied
23:44:37  <thlorenz>since it happens on require time
23:44:47  <thlorenz>as you can see I dumbed it down for IE now :)
23:46:12  <thlorenz>defunctzombie: so progress, now the actual tests fail for IE8 https://ci.testling.com/thlorenz/stack-mapper