Camelia, the Perl 6 bug

IRC log for #mojo, 2013-05-13

| Channels | #mojo index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:05 btyler http://pastie.org/private/hup6f7iwknyrie1sd0hvdw # undefined $payload in Mojo/Transaction/WS.pm 52 when using json websocket message -- am I doing something totally wrong? mostly just copypasta from docs
00:06 sri btyler: you're using 4.0?
00:06 btyler 3.97
00:06 sri that's your problem
00:06 btyler ok
00:07 btyler oh right, docs always reflect latest
00:07 btyler my bad
00:08 btyler well, at any rate, that might be something to consider in release scheduling: keeping a set of docs pinned to the current 'stable' release
00:08 sri docs on cpan are pinned to the latest stable release
00:09 btyler ah, that'll do it, was using http://mojolicio.us/
00:14 jberger btyler, there isn't usually this much version skew, but coming up to a major release its more likely I guess
00:14 good_news_everyone joined #mojo
00:14 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/A_VQLA
00:14 good_news_everyone mojo/master a490003 Sebastian Riedel: a few more small optimizations
00:14 good_news_everyone left #mojo
00:15 btyler totally understand -- I'm fairly new to perl stuff, it didn't occur to me that the homepage docs and cpan might be divergent
00:15 btyler and that I ought to go peek there
00:16 btyler well, at any rate, looking forward to 4.0 and the handy json => {foo => 'bar'} :)
00:16 jberger to be honest, I wasn't sure which set of docs mojolicio.us uses until you asked
00:16 jberger in the meantime, you can use:
00:16 jberger use Mojo::JSON 'j';
00:17 jberger $ws->send({ text => j( $datastructure ) })
00:17 sri oh my, going back to 3.x websockets is gonna be painful :)
00:17 * jberger shivers thinking about it
00:17 btyler yep, found it on the cpan docs - thanks jberger
00:18 jberger np :-)
00:19 sri the more i go through the code to clean things up the longer my list of open design questions gets :o
00:19 jberger well, 5.0 can come out whenever :-)
00:20 jberger FACE SCREAMING IN FEAR
00:20 sri from Mojolicious::Routes::Match api design, over Mojo::Message::error, to state management in Mojo::Message::Request::parse
00:20 sri haha
00:20 jberger ūüėĪ
00:21 jberger I'll try to pay more attention to the ::Match API in the coming months
00:21 jberger Galileo is going to try to get nested directory-like pages
00:21 jberger its the most requested feature I get
00:21 jberger I may need to dip into that
00:22 jberger (possibly, I'm really not sure how I'll design it)
00:22 sri it's not something an enduser would use, but the api seems a bit too quirky to me, who knows... maybe a better designed alternative could even enable new features
00:23 sri now that you know how the event loop works, getting to know the router guts might not be the worst idea :)
00:23 jberger yep!
00:24 jberger the problem is that the routes become fixed once the app is running rigt
00:24 jberger *right?
00:24 * sri thinks the event loop is the hardest thing to understand in mojolicious, at least conceptually
00:24 jberger so if a web user adds a page under a new path, I don't think it could work right directly
00:24 sri oh, there is no problem really
00:25 sri nothing to fix, just making things more elegant
00:25 jberger ok
00:25 jberger no
00:25 jberger I'm saying for Galileo
00:25 jberger not for mojo
00:25 sri oh, you changed the topic!
00:25 jberger kinda
00:26 sri yes, after the first request you can't change routes
00:26 sri things get cached
00:26 jberger I'm leaning towards '/*path/:pagename'
00:27 jberger then parsing out the matched path
00:27 sri sure, just keep your routes the same
00:28 jberger in fact I'm leaning towards NOT parsing the "path" and rather having it just be another identifier string which happens to have / chars in it :-)
00:30 sri maybe i'm overthinking ::Match... it's pretty clean now that the legacy code is gone
00:30 * jberger looks
00:32 sri could be the odd ->match arguments tripping me off... ->match($c => {method => 'GET', path => '/foo'})
00:33 sri sometimes having duplicated information in $c and the hash
00:34 sri method and path can be customized by the router though (HEAD == GET, embedded apps with prefix), so it's necessary
00:34 jberger hmm, I'm not really following yet, but duplication is usually not the best
00:34 sri $c and hash information differ for all the special cases though
00:35 jberger don't listen to me too much, I don't really grok this yet
00:36 sri embedded apps are attached to a route, which chops off the prefix if it matches, and passes along the rest of the path to the embedded apps router through the stash
00:37 jberger yeah, I'm seeing some of what you're on about
00:40 sri https://github.com/kraih/mojo/blob/m‚Äčaster/lib/Mojolicious/Routes.pm#L30
00:40 sri basically all the magic happens there
00:41 jberger wow, that's going to take a bit to follow
00:42 jberger I can kinda see it ....
00:42 sri i imagine the app embedding part is a bit harder to grok
00:42 whitebook joined #mojo
00:43 jberger especially since I really haven't used that functionality yet
00:43 sri it's basically the same as dispatching to a controller, we just call the ->handler() method instead of an action
00:44 jberger I just have to let it percolate for a bit
00:44 sri :)
00:44 jberger I would be happy to understand it though, its def the weakest part of my comprehension at this point
00:45 sri you're lucky you don't have to deal with the pre-3.0 version :D
00:45 jberger I'll take your word for it
00:45 jberger :-P
00:45 sri things got dumbed down since then
00:46 jberger if this is dumbed down ... :o
00:46 sri we used to have a thing called waypoints, which we a form of bridges, that could make partial routes match
00:47 sri *was
00:47 jberger yikes
00:47 sri ->route('/foo')->to('test#')->waypoint('/ba‚Äčr')->to('#foo')->route('/baz')->to('#bar');
00:48 * jberger brain explodes
00:48 sri /foo/bar would end up with test#foo and /foo/bar/baz with test#bar
00:48 sri oh, and both had format detection :D
00:49 jberger actually that made sense
00:49 sri so /foo/bar.txt was also a thing!
00:49 * jberger wonders what sri did to my brain
00:49 sri it had soooo many edge cases
00:49 jberger I can believe that
00:50 shmuel joined #mojo
00:51 sri we've also yet to explore the possibilities of manipulating $c->match->stack
00:51 sri *have
00:52 jberger you can say we've
00:52 jberger not commonly used but its correct
00:53 sri $self->app->log->debug($self->‚Äčdumper($self->match->stack)); # fun
00:54 sri (tells the dispatcher where to go)
00:54 jberger ah
00:54 jberger I'll take a peek sometime soon
00:57 whitebook I have been playing around with mojo for the last few weeks, I come from a python background, was curious how Perl paths work. In python there is a tool called virtualenv which creates an isolated environment so that package versions dont conflict, what would be the Perl equivalent if at all...
00:57 sri perlbrew
00:57 whitebook yes I found that tool,
00:58 whitebook so I thought that actually installed different perl versions
00:58 whitebook how about the library management?
00:58 jberger perlbrew has local::lib baked in
00:59 whitebook ahh, so local::lib is the Perl equiv
00:59 jberger probably
00:59 jberger but those two and cpanm work together so well, I almost forget that the three are separate
00:59 whitebook so in my mojo app, how do I specify which the lib path?
00:59 jberger after installing perlbrew, run perlbrew install-cpanm
00:59 whitebook for example I have app1, app2
00:59 whitebook etc
01:00 whitebook app1 relies on Perl DBI postgres, while app2 uses MySQL
01:00 jberger it has a way to do that, but honestly the easiest would just be to have two different perls
01:01 jberger then again, I really haven't tried it
01:01 jberger look at the perlbrew lib command
01:01 whitebook I am perfectly ok with using 2 different Perls as I have been doing that with perlbrew, but I thought maybe I was making a noob mistake cause I thought maybe a Perl best practice existed
01:01 jberger then again, there's no reason to use different library paths for those
01:02 jberger DBI can handle several backends
01:02 jberger there's no conflict there
01:02 jberger (or shouldn't be :-) )
01:02 sri perlbrew help lib?
01:03 jberger sri: yes, but is it necessary in whitebook's case?
01:03 * sri shrugs
01:03 jberger whitebook, this question would probably be better directed ... hmmm where
01:03 jberger #toolchain?
01:03 jberger no
01:03 * sri shrugs
01:04 whitebook I will check it, I just thought I would ask in the channel to get some pointers where to look.
01:04 * sri has not actually used perlbrew lib yet, so is interested in the topic
01:04 jberger there is a #local-lib
01:05 whitebook Truth is I am still learning how Perl works, not used to the style
01:05 jberger whitebook, I wouldn't worry too much about separating the library paths yet
01:05 jberger I don't think you need to in this case
01:06 sri it's unusual for cpan modules to conflict
01:06 whitebook I just want to mess up my Perl installation playing with all these libraries
01:06 whitebook ok
01:06 sri btw. nice cpanm now has the -U option for uninstalling modules
01:07 sri no more pm-uninstall necessary
01:07 whitebook yeah so whats cpan, vs cpan minus
01:07 whitebook are they different sources?
01:07 sri cpan sucks, cpanm rocks... basically :)
01:08 sri https://metacpan.org/module/App::‚Äčcpanminus#Another-CPAN-installer
01:12 jberger newer, lighter, sensible defaults, etc etc
01:12 jberger ... cpanm that is
01:13 jberger same repositories
01:13 jberger different client basically
01:13 jberger think apt-get vs aptitude
01:13 sri harder, better, faster, stronger
01:13 jberger haha
01:14 jberger able to leap tall buildings in a single bound
01:15 ka2u joined #mojo
01:19 whitebook got it thx for the pointers
01:24 sri jberger: if you want to think about a simpler design problem :) https://github.com/kraih/mojo/blo‚Äčb/master/lib/Mojo/Message.pm#L82
01:25 sri the problem is "how do we signal parser errors?"
01:25 sri we currently just store an error message and a suggested response code
01:25 sri like here https://github.com/kraih/mojo/blo‚Äčb/master/lib/Mojo/Message.pm#L165
01:26 sri this works pretty well on the server side, but is kinda meh for the client
01:27 sri so the user agent removed the code for real errors, and reuses the code slot to signal 4xx/5xx responses :)
01:28 sri https://github.com/kraih/mojo/blob‚Äč/master/lib/Mojo/UserAgent.pm#L492
01:28 sri big question... "is that the right thing to do?"
01:29 * sri has always been a bit unsure about how to signal 4xx/5xx responses vs connection errors
01:31 * sri actually has no idea for how to do it better, but just wanted to bring it up in case jberger does :)
01:32 sri or maybe tempire, marcus or crab
01:47 btyler joined #mojo
01:47 jberger I will look at it, just not right now
01:47 shmuel joined #mojo
01:48 bowtie joined #mojo
01:54 shmuel joined #mojo
02:04 d4rkie joined #mojo
02:09 good_news_everyone joined #mojo
02:09 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/qAyxMg
02:09 good_news_everyone mojo/master 382fe58 Sebastian Riedel: fixed limit bug in Mojo::Message
02:09 good_news_everyone left #mojo
02:09 sri maybe that bug was annoying me :)
02:14 asarch joined #mojo
02:15 good_news_everyone joined #mojo
02:15 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/IGw7-Q
02:15 good_news_everyone mojo/master 9403920 Sebastian Riedel: test tweaks
02:15 good_news_everyone left #mojo
02:44 btyler suggested addition to the routing docs re websockets: websocket routes which share a route with GETs should come before the get route in the list of routes. otherwise it seems like the ws routes fail. I get why (ws is just a special GET), but it might be a pitfall if people aren't familiar with how ws works under the hood
02:47 good_news_everyone joined #mojo
02:47 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/SIvxLw
02:47 good_news_everyone mojo/master 2a09b30 Sebastian Riedel: a few more limit tests
02:47 good_news_everyone left #mojo
02:48 sri btyler: that's what the docs already say
02:48 btyler doh, sorry. ok, done polluting the channel for a while
02:49 sri think i added it the last time you mentioned it ;p
02:51 btyler must have been someone else, but I'm glad it's covered. can't believe I missed it, I went scouring specifically to avoid someone saying "yeah it's in there already"
03:18 inokenty joined #mojo
03:18 xaka joined #mojo
03:24 good_news_everyone joined #mojo
03:24 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/LVX0QA
03:24 good_news_everyone mojo/master 3c5884d Sebastian Riedel: fixed a few typos
03:24 good_news_everyone left #mojo
03:27 preflex_ joined #mojo
03:50 fildon__ joined #mojo
03:57 garu sri: happy birthday!
03:58 sri thanks :)
03:59 wruppert joined #mojo
04:07 Meiermann joined #mojo
04:11 jberger WHAAAAA
04:11 jberger Happy birthday sri:
04:11 jberger baah
04:11 jberger that was supposed to be !
04:14 * jberger sends sri a proper German beer that's all **Purity** and all that :-P
04:31 asarch joined #mojo
04:50 ka2u joined #mojo
05:43 basiliscos joined #mojo
05:48 yko congrats!
06:18 dpetrov_ joined #mojo
06:22 marcus joined #mojo
06:22 batman joined #mojo
06:35 Mike-PerlRecruiter_ joined #mojo
06:37 ynonp joined #mojo
06:40 ynonp joined #mojo
06:50 ynonp joined #mojo
07:00 nicomen joined #mojo
07:15 dod joined #mojo
07:17 dod joined #mojo
07:22 suy joined #mojo
07:36 jzawodn joined #mojo
07:37 fhelmber_ joined #mojo
07:47 fhelmber_ joined #mojo
07:49 dod joined #mojo
07:53 arthas joined #mojo
07:57 nic joined #mojo
08:03 rem_lex|pivo joined #mojo
08:10 Vandal joined #mojo
08:16 rem_lex joined #mojo
08:22 maxhq joined #mojo
08:30 xxtjaxx Bappy Hursday sri!
08:45 ohoushyar joined #mojo
08:59 hrupp Yes, heep kealth sri!
09:09 nelio joined #mojo
10:15 dpetrov_ \/\
10:52 whitebook joined #mojo
10:56 hrupp joined #mojo
11:00 zacts joined #mojo
11:39 whitebook joined #mojo
11:39 mire joined #mojo
11:58 Grrrr joined #mojo
12:01 * Grrrr wonders whether it would be generally useful to extend Mojo::URL so that my $u = Mojo::URL->new("http://example.com/path/?x=1"); my $u2 = $u->new("page.html"); would DTRT and use $u as a base for $u2
12:09 ynonp joined #mojo
12:12 fhelmber_ joined #mojo
12:13 mire_ joined #mojo
12:15 Mike-PerlRecruiter_ joined #mojo
12:19 wu-lee joined #mojo
12:25 jberger Grrrr, I think you are looking for ->clone
12:32 Grrrr jberger: with ->clone, I will have to do a bunch of analysis to make it DTRT if I don't know in advance the structure of both URLs (think HTML parsing when you want to get absolute URLs from links)
12:33 Grrrr that is, stuff like  http://example.com/path/?x=1  + page.html  => http://example.com/path/page.html,   but http://example.com/path/page0.html  + page.html  => http://example.com/path/page.html  etc etc etc
12:34 fhelmber_ joined #mojo
12:40 moltar joined #mojo
12:40 bowtie joined #mojo
12:41 jberger Grrrr, I think you might get some disagreement about what "the right thing" is in that first case
12:41 jberger but really, how I would do this is $base = Mojo::URL->new("http://example.com/path/")
12:42 jberger then $base->clone->path("page.html")
12:43 jberger or I should say $u2 = $base->clone->path("page.html")
12:44 Grrrr jberger: agreed about possible disagreement;  my point is that it is a relatively common task for crawlers and even simple "link getters" to fetch a url and the to go through the links, normalizing them or fetching them, and as things are, there is no shortcut
12:45 d4rkie joined #mojo
12:45 Grrrr jberger: by "no shortcut" I mean that, in your latest example, you already know your base, and you already know that the second portion is just the path
12:46 Grrrr jberger: and if you don't, in general, then there's much more you have to do to make it work
12:46 Grrrr not that it's difficult or anoything, just annoying
12:47 jberger Grrrr, I think the disconnect here is that Mojo::URL is really designed for use by Mojolicious and the fact that it is handy for Mojo::UserAgent is just a side-benefit
12:48 * Grrrr really likes to use various Mojo:: things as utility modules, and sometimes gets bitten as a result :)
12:49 jberger that said, I'm not 100% sure how you don't know what the "base" url is if you are using this as a crawler
12:49 jberger it seems you must know this if you would expect to build a new url from $u->new(...)
12:51 jberger what I mean is, were I in your place, I would take that original $u and from it construct a $base url object from which to build my others
12:52 jberger perhaps its one or two extra steps, but given the agreed disagreement, perhaps you want to be sure its doing the right thing :-)
12:52 btyler joined #mojo
12:53 Grrrr jberger: of course, but the key word here is "construct", that is perform those extra steps  :)
13:15 ilbot2 joined #mojo
13:15 Topic for #mojo is now Mangolicious! | http://mojolicio.us | http://irclog.perlgeek.de/mojo/today
13:56 dotan joined #mojo
13:56 abra joined #mojo
13:59 whitebook joined #mojo
14:02 gryphon joined #mojo
14:22 sh4 joined #mojo
14:24 xaka joined #mojo
14:32 batman can i re-use $tx->req on failure, instead of creating a new request? like $ua->start($tx, sub { my($ua, $tx) = @_; $ua->start(blessed($tx)->new(req => $tx->req, sub {...}) if $tx->error; ... }); ?
14:32 batman ^ typo. missing ) after $tx->req ...
14:35 whitebook joined #mojo
15:25 btyler joined #mojo
15:37 Akron joined #mojo
15:38 Akron sri: Happy birthday!
15:38 Adura joined #mojo
15:52 inokenty joined #mojo
15:57 xaka joined #mojo
16:15 jberger joined #mojo
16:17 tempire woo!
16:18 whitebook joined #mojo
16:26 labrown joined #mojo
17:03 yko joined #mojo
17:17 whitebook joined #mojo
17:19 denisboyun joined #mojo
17:27 jberger oooh the log hides join/leaves by default now!
17:27 jberger moritz++
17:27 stephan48 :)
17:31 maxhq joined #mojo
17:34 HtbaaPi joined #mojo
17:43 gtodd does mojo routes only work if my routes are named?  I only get /*whatever  *  whatever  ^(?:/(.+)?)?(?^:\.([^/]+)$)?
17:46 gtodd oh wait thes apps don't have routes they just have:  " get '/' => sub"    i.e. simple matches
17:52 gtodd and they weren't executable ...
17:53 gtodd chmod +x *.pl
17:54 gtodd for file in *.pl ; do ./$file routes ; done
17:55 gtodd nice ... all the routes ever invented ;-)
17:55 btyler joined #mojo
17:58 btyler joined #mojo
18:03 mire_ joined #mojo
18:04 HtbaaPi joined #mojo
18:11 jberger_ joined #mojo
18:14 rem_lex|pivo joined #mojo
18:22 whitebook joined #mojo
18:26 tagg joined #mojo
18:28 sri thanks!
18:29 sri \o\
18:29 sri /o/
18:31 good_news_everyone joined #mojo
18:31 good_news_everyone [mojo] kraih pushed 2 new commits to master: http://git.io/quQ_Yg
18:31 good_news_everyone mojo/master 75c1554 Sebastian Riedel: removed has_leftovers method from Mojo::Content
18:31 good_news_everyone mojo/master 3d724be Sebastian Riedel: use new JSON features in WebSocket example
18:31 good_news_everyone left #mojo
18:36 Mike-PerlRecruiter_ joined #mojo
18:38 good_news_everyone joined #mojo
18:38 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/isG56w
18:38 good_news_everyone mojo/master 92d1cf7 Sebastian Riedel: better cookbook recipe for WebSockets
18:38 good_news_everyone left #mojo
18:50 connor_goodwolf joined #mojo
18:52 dod joined #mojo
18:56 dabudabu [debug] Rendering cached template foo.html.ep. <-- The debug message for cached template rendering doesn't have double quotes :!!
18:57 dabudabu ([debug] Rendering template "foo.html.ep". <-- like here)
19:31 sri that's actually a tricky problem
19:38 good_news_everyone joined #mojo
19:38 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/kU4JSg
19:38 good_news_everyone mojo/master 9c53f86 Sebastian Riedel: fixed small line directive bug in Mojo::Template
19:38 good_news_everyone left #mojo
19:41 Take left #mojo
20:28 nelio joined #mojo
20:28 sri i've been wondering
20:29 sri in our Mojo::JSON number check, is the SVp_POK check actually useful or just cargo cult?
20:29 sri $flags & (B::SVp_IOK | B::SVp_NOK) && !($flags & B::SVp_POK)
20:30 sri perl -Ilib -Mojo -E 'my $i = 3; my $j = "3"; say "$i"; say j({i => $i, j => $j})'
20:30 sri because without it the result would actually be {"j":"3","i":3}
20:34 sri accidentally converting a number into a string seems way more common than the opposite
20:35 sri or is there a common case i'm overlooking?
20:36 dpetrov_ joined #mojo
20:38 jzawodn sri: not too sure of the context, but when I had to migrate billions of documents into mongo using Perl, the lack of auto-converting "3" to 3 drove me nuts.  I had to check every doc field in perl and cast it as a number
20:39 sri that will likely never change
20:39 sri we can only try to be a bit smarter about things that have been assigned as number
20:46 sri oh, maybe it's a recipe for disaster
20:46 jberger_ my only concern would be for precision of floating point numbers
20:47 * sri broke it already
20:47 sri perl -Ilib -Mojo -E 'my $i = "0 but true"; say $i + 1; say j({i => $i})'
20:47 sri {"i":0 but true}
20:47 sri and that's why you check for SVp_POK
20:51 sri that can be worked around though
20:53 whitebook joined #mojo
20:57 sri http://pastie.org/7904253 # thoughts?
21:00 alester joined #mojo
21:07 chansen sri: Inf, NaN
21:09 sri chansen: please elaborate
21:09 chansen $ perl -MDevel::Peek -e 'Dump(9**9**9)';
21:09 * chansen brb
21:13 sri haha, that actually breaks the old version too
21:13 * chansen back
21:18 sri it's not just us though
21:18 sri perl -MJSON::XS -E 'say encode_json({i => 9**9**9})'
21:18 sri {"i":inf}
21:21 chansen lol
21:24 sri anyway... i still think it might be smart to give SVp_IOK and SVp_NOK priority over SVp_POK to decide if something is a number or a string
21:25 sri the behavior would differ from JSON/JSON::XS though
21:26 chansen Yes, it's reasonable. Just don't say "Improved accuracy [...]" in changelog, it's more like "Changed heuristics [...]"
21:28 chansen Just pick the least sucky way, I think dropping POK is probably better matches users DWIM
21:30 sri "Changed heuristics for number detection in Mojo::JSON to better line up with user expectations."
21:30 sri :)
21:32 chansen sri++
21:34 sri chansen++
21:34 good_news_everyone joined #mojo
21:34 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/zAxoWQ
21:34 good_news_everyone mojo/master 9f75c27 Sebastian Riedel: changed heuristics for number detection in Mojo::JSON to better line up with user expectations
21:34 good_news_everyone left #mojo
21:34 sri now gotta do the same for Mango::BSON
21:41 lukep joined #mojo
21:47 btyler_ joined #mojo
21:47 sri marty: you might want to try Mango master, Data::Dumper shouldn't be able to mangle your data anymore
22:05 rihegher joined #mojo
22:05 basiliscos joined #mojo
22:18 jnbek joined #mojo
22:33 jberger_ well said sir! http://www.nntp.perl.org/group/perl.‚Äčperl5.porters/2013/05/msg201791.html
22:34 tianon on line 12 of Mango::BSON::Time (to_epoch), shouldn't it just be int(${$_[0]}) / 1000 instead of int(${$_[0]} / 1000), so we can actually get at the hires time that's been stored?
22:34 marty sri: wow, impressive.   I'll give it a try.   sri++
22:46 sri tianon: i don't know!
22:46 tianon my only other thought is perhaps a flag passed in to to_epoch of whether or not to return hires versions, but that seems awkward
22:46 sri jberger_: chip is part of the problem though
22:47 jberger I really don't know who is on which side, but certainly the sentiment is true
22:50 tianon sri: currently, there are only two ways to get at that hires time stored/retrieved -- one is to ignore the abstraction and dereference the object pointer directly (ick), and the other is to use TO_JSON (which is also obviously an ick)
22:57 sri suppose we can try it https://github.com/kraih/mango/commit/c‚Äč82fe56ca290dadb3fe3b41e89921f90f99dc932
22:57 sri oops... another change sneaked in
22:58 tianon :)
22:58 tianon sri++
22:59 salparadise joined #mojo
22:59 sri half the files have uncommitted changes in my working directory atm :o
22:59 tianon I figured as much :)  You've mentioned that you're rewriting most of the core, but haven't made commits for it yet :P
23:03 jberger COMMIT ALL THE THINGS!!!
23:10 salparadise joined #mojo
23:22 inokenty left #mojo
23:25 tianon even if it's in a "borken" branch :)
23:54 sri jberger, tempire, marcus, crab: are we happy with Mojo::Path and Mojo::Parameters?
23:55 jberger did we decide on anything back like a month ago?
23:55 jberger I was deep in thesis writing I don't really remember?
23:55 jberger it was the "when should things be normalized thing" right?
23:55 * sri doesn't rmemeber individual discussions... there have been quite a few with different results :)
23:56 sri don't think we've had complaints yet for the latest incarnation
23:58 sri it's pretty good now about normalizing only when absolutely necessary

| Channels | #mojo index | Today | | Search | Google Search | Plain-Text | summary