Camelia, the Perl 6 bug

IRC log for #mojo, 2013-02-08

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

All times shown according to UTC.

Time Nick Message
00:02 TitanOfOld joined #mojo
00:15 Molaf_ joined #mojo
00:27 Averna joined #mojo
00:32 _xaka_ joined #mojo
00:33 good_news_everyone joined #mojo
00:33 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/ppeh1A
00:33 good_news_everyone mojo/master d4a0998 Sebastian Riedel: small optimizations
00:33 good_news_everyone left #mojo
00:42 KindTwo joined #mojo
00:53 jzawodn joined #mojo
00:59 ACE joined #mojo
01:08 jberger joined #mojo
01:08 mire joined #mojo
01:12 sri $cursor->explain works now too \o/
01:18 PanzerBjorn Well, I got my stuff jiggered up. I added some extra stuff to my controller methods to take a partiel render parameter and return a scalar if it's set.
01:18 ka2u joined #mojo
01:21 PanzerBjorn (And then updates to all the error/warning helpers to also take that parameter and return partial as necessary)
01:24 tempire mojocasts have returned
01:24 sri tempire++
01:24 Skylos sweet
01:25 tempire Skylos: calling a controller from within a controller is a bad direction to go in, so there will be never be an obvious way to do it in mojolicious. If you want to break the rules, you have to work for it :)
01:27 Skylos tempire - so I noticed, as I was scrutinizing the dispatch code.
01:27 Skylos tempire - how would you answer the situation I outlined though, within an application?
01:28 tempire abstract the json retrieval functionality into a helper
01:28 tempire than call it from both controllers as necessary
01:28 Skylos thanks.
01:29 tempire or make a bridge (upstream function), get the json there, and store it in the stash for the controller actions to use down the pipe.
01:30 Skylos upstream function?
01:30 Skylos Oh, like... a middleware?
01:33 tempire https://gist.github.com/tempire/4735882
01:34 tempire somecontroller#someaction will be called first, whenever /someroute or /someotherroute is requested.
01:35 tempire the helper is usually the better way to go, though, unless there's another reason for a bridge.
01:35 tempire but it's an option.
01:37 asarch joined #mojo
01:37 * Skylos puzzles at the semantics of bridge
01:41 tempire it's the same as under in lite
01:41 tempire http://mojocasts.com/e3#Under%20directive
01:41 Skylos yeah, but I don't know it well enough to understand that.
01:42 Skylos the bridge is...?
01:43 Skylos what do you do with the bridge, eventually?
01:43 Skylos your gist doesn't show the utilization just the setup
01:45 tempire usually a bridge is used for authentication, but it doesn't matter what you do with it.  it only matters that you return true if you want whatever is under the bridge to execute.
01:46 Skylos return true from what
01:47 tempire somecontroller#someaction
01:51 Skylos hm.
01:52 Skylos so bridge just connects two(ormore) routes?
01:52 Skylos or controllers?
01:53 marty sri++ mango is growing up.
01:53 sri actually, it was just shrinking by 50 lines or so :D
01:54 * sri is quite happy with how the basics are mostly covered now
01:54 marty I couldn't agree more
01:55 sri the rest is mostly sugar coating for commands
01:55 marty I can't way to play with ->explain
01:55 marty is it possible to ->drop a collection?
01:55 sri that's just a command :)
01:55 marty ahhh
01:55 marty then it is
01:55 marty and I will
01:55 marty but carefully
01:55 sri all the collection/db manipulation stuff is commands
01:56 azawawi joined #mojo
01:56 PanzerBjorn I use a bridge in my app for all paths - it handles auth and various duty that needs done for all routes.
01:56 sri i think we could generate wrappers for all the commands that return whole documents, and add like 30 at once :)
01:56 azawawi Hi
01:57 marty like a command wrapper orgy
01:57 sri o/
01:57 marty hi azawawi
01:57 sri what slows me down is just the testing of non-blocking stuff, only time when i actually have to think for a bit ;p
01:58 azawawi **** I finally migrated to mojo websockets in Farabi... https://metacpan.org/source/​AZAWAWI/Farabi-0.32/Changes   # sri++  for an excellent implementation
01:58 sri \o/
01:58 azawawi The speed is really good
01:58 azawawi i found some interesting things that i would like to share...
01:59 azawawi firefox ... press ESC ... kills websocket connection... had to workaround it by disabling ESC on $(window)
01:59 azawawi actually kills both AJAX and websockets...
01:59 azawawi so lame
01:59 marty congratulations, azawawi
01:59 sri the $cursor->hint/snapshot stuff is really trivial to add, but i'm not sure how to test it yet
01:59 azawawi sri: i have a question about "Protocol change". What's that?
02:00 sri tailable cursors will be a little more tricky
02:00 sri azawawi: please elaborate
02:01 marty but I'll bet they come in very handy if you can get it working
02:01 sri tailable cursors will be great, for all kinds of queue, pubsub stuff
02:01 * marty nods
02:02 azawawi sri: sometime i get "101 Switching Protocols" and afterwards websocket seems to die in firefox
02:02 d4rkie joined #mojo
02:02 sri azawawi: that's the websocket handshake
02:03 sri websockets are a binary protocol, with an http based handshake
02:04 azawawi my bad... it is still alive... i used farabi on my windows home dir... and tried the open file feature which reads the whole dir to match your filename
02:05 sri if you want to know more about websockets, the rfc is actually pretty readable http://tools.ietf.org/html/rfc6455
02:05 azawawi sri: another question regarding memory-based session. I looked around and found a dancer plugin that implements that... Let me get it...
02:06 azawawi sri: basically that's Java's model of storing things in the memory session... and holding the key in the cookie...
02:06 azawawi sri: do we have something similar to this https://metacpan.org/modul​e/Dancer::Session::Simple
02:06 sri no clue
02:07 * sri doesn't use that kind of session
02:07 azawawi sri: when session store X in mojo, it is stored in the cookie and hence the 4K limitation, right?
02:08 azawawi sri:  using that memory model, we dont  have to store things in the cookie... Just the id to a memory-based store
02:08 azawawi sri: session id
02:10 marty sri:  I been using Mango::BSON::Binary to load 1000's of images and it has not barfed once.  \o/
02:10 sri \o/
02:10 sri that reminds me, GridFS is also on the todo list
02:11 marty \o/
02:11 marty That might challenge you a bit
02:11 sri it's also really simple
02:11 marty well, maybe not
02:11 sri basically just two collections with fancy names and specific document layout
02:12 sri one for meta data, one for chunks of data
02:12 marty ahh, makes sense.
02:13 azawawi joined #mojo
02:13 marty I suppose loading a file larger than 16mb in a db is not all that uncommon.
02:14 * azawawi sleeping time... &
02:17 sri marty: i'm also unsure about bson_bin()->type('generic') btw
02:17 sri (the type being a string)
02:17 marty That's what I have been using to load images
02:18 marty I did notice that there was a very slight file size difference between the mango and the other driver.
02:18 sri on the other hand ->type(5) where you have to look up the numbers first sucks
02:18 sri file size difference?
02:19 marty the byte count of the loaded image was different
02:19 sri that's very odd
02:19 marty lemme see if I can work up a more specific test
02:19 sri generated bson should be the same
02:20 marty also, just an fyi.  there seem to be some interaction between the drivers.  when I have the MongoDB driver loaded along side mango I get some intermittent warnings
02:21 sri also very odd
02:21 Skylos left #mojo
02:21 marty I'll see if I can recreate
02:22 sri Mango is very self contained interactions are pretty much impossible
02:24 sri now that i think about it, tailable cursors might be trivial to add
02:25 marty ahh.  Could it be that mango handles the objectId created by the mongo driver differently?
02:25 sri like, set one flag and be done
02:25 sri marty: wait, you mean you pass MongoDB::OID objects into Mango?
02:26 marty I'm not sure yet, but I think it was when I tried to read an object id created by the MongoDB driver and then insert it with mango.
02:26 sri yea, that's not gonna work
02:26 marty I'm guessing they are not a universal thing.
02:26 marty ok, thx, my bad
02:26 sri those special wrapper objects are matched by namespace
02:27 marty ahhhh
02:27 marty the light comes on
02:27 sri anything MongoDB::* will not work in Mango
02:27 sri you should get very funny results actually
02:27 marty whew, good to know.
02:28 sri MongoDB::OID has a TO_JSON method that returns {'$oid' => '1234567890'}
02:28 marty yep, I am conveting all things over now but that explains why I saw some strange behavior.
02:28 sri if you go through the database it's all the same again, just don't mix them up on the perl side
02:28 marty I did not realize OID were tied to a namespace.  that's actually really cool
02:29 sri you should take a look into Mango::BSON
02:29 * marty peeks
02:29 sri it's pretty readable and you'll grok how it all works
02:29 sri really simple all
02:30 sri that's where the sausage is made https://github.com/kraih/mango/bl​ob/master/lib/Mango/BSON.pm#L235
02:34 sri for true and false i actually reuse Mojo::JSON->true/false, maybe you've noticed that those "just work" in mojolicious :)
02:34 marty That's a good read.  I think I'm getting it more.  yes, I have noticed how they work and it's allowed me to clean up my code a lot.  I love the tight integration with mojolicous
02:35 sri making GridFS work with Mojo::Asset might be fun
02:37 sri $mango->db->gridfs->put($self->param('upload'));
02:37 sri or even Mojo::Upload for put
02:38 marty the possibilities become huuuge
03:09 sivoais joined #mojo
03:24 sivoais joined #mojo
04:06 zivester joined #mojo
04:08 bizam hi all, please suggest me about bridge route: $r->bridge('/foo')->to(...)->route('/bar')->... What is '/foo' and what is '/bar'?
04:10 tempire bizam: a request from a client/browser: /foo/bar
04:11 marty is that syntax correct?
04:12 marty my $foo = $r->bridge('/foo')->to(controller => 'foo', action => 'baz');
04:14 bizam matry: where is chain?
04:15 bizam $r->bridge('/foo')->to(controller => 'foo', action => 'baz')->get(...)->to(...)
04:16 marty interesting.  I've never used bridge like that.
04:16 zivester joined #mojo
04:18 bizam tempire: thanks, but is this correct with many to: $r->get('/foo')->to(cb=>su​b{1;})->to(cb=>sub{...}); ?
04:19 marty You can also do... my $foo = $r->bridge('/foo')...  my $bar = $foo->bridge('/bar')...  my $baz = $bar->bridge('/foo')...
04:20 bizam i want bridging only '/foo' but not '/foo/bar' path
04:21 bizam marty: i know
04:26 bizam i found that $r->get('/foo')->to(cb=> sub {die 'die 1';})->to(cb=>sub{die 'die 2';}); always die 2
04:29 * bizam feature requesting $r->route(...)->to(...)->to(...)
04:30 jberger bizam, its policy not to request features via github fyi
04:31 jberger also I would recommend you have a really good use-case laid out before you do it. I don't see sri going for it either way
04:31 jberger of course that is my opinion
04:31 jberger I don't want to put words in his mouth
04:31 jberger a good use-case will help either way :--)
04:32 * jberger just noticed my nose growing!
04:35 xaka joined #mojo
04:38 bizam two things: one path and more
04:38 bizam apparent
04:38 asarch joined #mojo
04:39 bizam more apparent syntax of the route
04:40 bizam more apparent syntax of the bridge route
04:51 jberger sri: I herd you like benchmarking, so xdg put benchmarking in your benchmarking so you can benchmark while you benchmark: http://www.dagolden.com/index.php/184​9/adventures-in-benchmarking-part-1/
04:57 yakudza joined #mojo
05:07 ka2u joined #mojo
05:14 rem_lex|pivo joined #mojo
05:23 ka2u joined #mojo
05:52 spleenjack joined #mojo
06:04 Miked joined #mojo
06:09 ver joined #mojo
06:10 vishwanathms joined #mojo
06:12 vishwanathms_ joined #mojo
06:18 vishwanathms joined #mojo
06:28 vishwanathms_ joined #mojo
06:28 vishwanathms joined #mojo
06:42 dhg joined #mojo
06:46 Vandal joined #mojo
07:09 Mike-PerlRecruiter_ joined #mojo
07:12 ObseLeTe joined #mojo
07:19 Phk joined #mojo
07:26 yakudza joined #mojo
07:36 vishwanathms joined #mojo
07:36 vishwanathms_ joined #mojo
07:50 ObseLeTe joined #mojo
07:52 ver Hello! Can I run the Mojo application with Mojo::Server::Morbo on a non-3000 port?
07:58 egopro joined #mojo
08:01 bizam $ morbo -h
08:03 bizam has listen => sub { [split /,/, $ENV{MOJO_LISTEN} || 'http://*:3000'] };
08:09 dod joined #mojo
08:11 pierrick joined #mojo
08:13 suy joined #mojo
08:15 ComLock joined #mojo
08:17 ver bizam, syntax error at launcher line 5, near "has listen"
08:17 miniCruzer joined #mojo
08:18 miniCruzer :O 50% more refracted light?!
08:18 ver bizam, where I can put this string? http://hash.su/7552
08:18 * miniCruzer sunglasses
08:19 dod joined #mojo
08:31 bizam глянь http://hash.su/7553
08:34 daxim joined #mojo
08:35 xaka глянул, ничего интересного :-p
08:36 bizam лучше tak $ morbo --listen 'http://*:8080' ./fw
08:44 ver bizam, 10x
08:44 bizam what?
08:44 bizam кирпичей?
08:46 dhg joined #mojo
08:55 KindOne joined #mojo
09:23 mire joined #mojo
09:29 navi joined #mojo
09:34 fhelmber_ joined #mojo
09:35 D4RK-PH0ENiX joined #mojo
09:37 drift_ joined #mojo
09:37 mugenken joined #mojo
09:39 cosimo joined #mojo
09:42 nelio joined #mojo
09:49 maxhq joined #mojo
09:55 vervain joined #mojo
10:15 connor_goodwolf joined #mojo
10:46 suy joined #mojo
10:54 SmokeMachine joined #mojo
11:38 mattastrophe joined #mojo
11:45 ObseLeTe joined #mojo
11:45 ObseLeTe joined #mojo
12:01 spleenjack1 joined #mojo
12:13 biz joined #mojo
12:14 mire joined #mojo
12:36 Vandal joined #mojo
12:37 modnar joined #mojo
12:40 mugenken joined #mojo
13:16 Skylos joined #mojo
13:28 spleenjack joined #mojo
13:28 moltar joined #mojo
13:33 SmokeMachine joined #mojo
13:39 TitanOfOld joined #mojo
13:42 sri jberger: you're absolutely right, without a very good use case i'm not interested
13:45 asarch joined #mojo
13:50 hmepas2 joined #mojo
13:50 Britzel_ joined #mojo
13:54 marty sri:  I may have found something.   https://gist.github.com/MadMartigan/4739145
13:54 marty When I insert an arrayref the first element is undefined
13:54 Skylos_ my $arrayref = [];
13:55 Skylos_ I'd write it push @{$arrayref}, $_
13:55 Skylos_ that is an interesting effect though.
13:55 Skylos_ its not like that before you insert it is it?
13:57 marty It also happens when I push @{$arrayref}.   I'll test @array also...
13:59 Skylos_ I doubt it has to do with the pushes
13:59 marty yep, same behavior when instering an array    ->insert({array => \@array});
13:59 Skylos_ probably has to do with some kind of saving logic.
13:59 Skylos_ off-by-one
13:59 Skylos_ like how its encoded into bson or something?
14:00 * marty nods
14:00 Skylos_ you could submit a test for that.
14:07 gryphon joined #mojo
14:22 SmokeMachine joined #mojo
14:33 sri marty: should be fixed in master
14:33 Skylos_ left #mojo
14:33 marty sri++  Thank you!
14:33 sri arrays are encoded as documents, like {0 => 1, 1 => 2, 2 => 3...}
14:33 sri i started with index 1 instead of 0
14:34 sri so i guess mongodb assumed index 0 was undefined
14:38 sri marcus: parrot threads suck, i would answer on twitter, but i really don't care for a discussion with reini ;p
14:38 bluescreen joined #mojo
14:39 marcus yeah, he seems a bit dogmatic.
14:39 sri the fact that he mentions erlang should be quite telling, erlang has no threads at all, just cheap processes
14:40 marcus http://perl6advent.wordpress.com/​2012/12/11/day-11-parrot-threads/ is what he is referring on Parrot I guess.
14:40 sri under the hood parrot threads are ithreads, very expensive and isolated, they just map green threads on top of it
14:41 marcus :/
14:41 marcus "Editors note: I, rurban, does know almost nothing about threads." :D
14:41 sri heh
14:42 sri http://niner.name/Hybrid_Th​reads_for_the_Parrot_VM.pdf # isolated + green threads = hybrid threads
14:42 sri appratently they plan to do automatic non-blocking i/o to make it work, which is kinda erlang-ish
14:43 sri but doing that at the vm level seems fishy
14:43 sri on the jvm you just get cheap os threads, and can build whatever you like on top
14:44 sri with parrot you get safety by default, with the jvm performance
14:46 * sri thinks running one interpreter per thread is silly
14:50 marcus yeah, that sounds very complicated.
14:51 marcus and also close to a cheap fork, like erlang.
14:51 mire joined #mojo
14:54 wircus .
15:04 sri not to mention that erlang is actually a rather slow language
15:04 sri i don't think erlang really forks
15:04 sri the vm just has a process pool and good ipc
15:05 sri erlang can send erlang bytecode via ipc, since everything is immutable, it doesn't really matter where the bytecode is executed
15:09 sri imagine you could serialize each and every function in your perl code, without having to worry about state, and execute it in another process, even on another physical machine, then just send the results back :)
15:10 Molaf__ joined #mojo
15:18 sri while i hate erlang as a language, the BEAM vm is actually pretty interesting, maybe new languages like Elixir will make it more user friendly
15:18 sri http://vimeo.com/53221562 # really good talk, if anyone's interested, especially the demonstrations towards the end
15:36 SmokeMachine joined #mojo
15:43 SmokeMac_ joined #mojo
15:44 SmokeMachine joined #mojo
16:06 zivester joined #mojo
16:07 cfedde joined #mojo
16:09 cfedde is there a way to take some xml text and get back a Mojo::DOM?
16:09 cfedde Mojo::DOM->new derp
16:10 labrown joined #mojo
16:18 crab seriously, is it forgetting to convert to the local timezone from utc in the query that extracts the buckets?
16:18 crab oops
16:22 dod joined #mojo
16:27 sri seriously!
16:55 SmokeMachine joined #mojo
17:02 xaka joined #mojo
17:09 dod joined #mojo
17:42 good_news_everyone joined #mojo
17:42 good_news_everyone [mojo] kraih pushed 2 new commits to master: http://git.io/hllFSQ
17:42 good_news_everyone mojo/master 3928f81 chromatic: Update lib/Mojo/Message/Response.pm...
17:42 good_news_everyone mojo/master d175764 Sebastian Riedel: Merge pull request #452 from chromatic/master...
17:42 good_news_everyone left #mojo
18:02 SmokeMachine joined #mojo
18:05 jberger_ joined #mojo
18:07 ka2u joined #mojo
18:13 rihegher joined #mojo
18:19 bc547 joined #mojo
18:55 Miked joined #mojo
19:08 vishwanathms joined #mojo
19:08 sri oooh
19:08 * sri just noticed the new tab bar above document option in TM2 o/
19:09 Mike-PerlRecruiter_ joined #mojo
19:19 yakudza joined #mojo
19:20 basic6 joined #mojo
19:23 rihegher left #mojo
19:28 BeDa joined #mojo
19:33 ka2u joined #mojo
19:37 d4rkie joined #mojo
19:39 bc547_ joined #mojo
20:04 perlite_ joined #mojo
20:16 sri hmm, think i'll expose $cursor->build_query, so it gets easier to debug the final query a cursor generates from all the attributes
20:40 njlg joined #mojo
20:45 sri yea, that makes testing a lot easier https://github.com/kraih/mango/commit/0​5e8613f0ba6ad685081322ef900eeff97e08b59
20:47 marty sri++
20:59 Britzel joined #mojo
21:22 Britzel joined #mojo
21:26 zaltekk joined #mojo
21:27 xaka joined #mojo
22:10 bluescreen joined #mojo
22:14 jberger_ does anyone have any experience with javascript for file upload via websocket?
22:14 jberger_ I have found this, it seems active
22:14 jberger_ https://github.com/binaryjs/binaryjs
22:26 marcus sri: debugging UA requests with binary data is really annoying, isn't there something we can do?
22:27 marcus at least like just output [BINARY] or something. my terminal goes crazy
22:37 sri marcus: think i proposed something a few weeks ago but nobody said anything
22:37 marcus sri: I must have missed that. It's a constant annoyance to me.
22:39 jberger_ not only does that sound nice, but I could take it a step further ...
22:39 jberger_ it would be nice if Mojo::UA could detect a mojolicious development mode 404 page and get the relevant info from the (rather large) html dump
22:40 jberger_ not sure if thats possible or feasible, but man would it be nice
22:40 marcus jberger_: you probably mean a 500
22:40 jberger_ yes
22:40 jberger_ thats what I mean
22:40 marcus then I agree, that would rock too.
22:40 sri pretty sure you're talking about different things
22:40 marcus but I'm not sure if it's too tight coupling
22:41 marcus sri: I'm talking about MOJO_USERAGENT_DEBUG output
22:41 sri yea, that shows the raw data stream
22:42 jberger_ I'm talking about ./myapp.pl get /
22:42 sri you don't even know what is header and what is body, or which belongs to which message
22:42 jberger_ or Test::Mojo->get_ok('/')
22:42 sri anyway, i'm too busy with mango
22:42 marcus sri: yeah, I don't really think it'd be feasible to do jberger_'s suggestion on that level.
22:43 jberger_ but again, I can't see any good way to do it (without lots of unnecessary special-casing)
22:43 jberger_ but a guy can dream :-)
22:43 marcus jberger_: content-negotitation and get / requesting text ? :)
22:43 marcus and a exception.text.epl
22:44 jberger_ sri: stick with mango, much better use of time
22:44 sri or just replace exception.html.ep with a text version
22:45 jberger_ hmmmmm, I like that
22:45 sri @@ exception.html.ep %= $exception->message
22:45 sri done :)
22:45 jberger_ I like the pretty one when I'm starting, but it does very little good when I do text_ok() expecting "blah" and getting *mountain of HTML*
22:45 marcus sri: as for my request $output= [ BINARY DATA REDACTED] if(-B $output);?
22:46 sri marcus: i have no opinion about that
22:46 * jberger_ will probably stick with silly things like lvalue attributes
22:47 jberger_ much more fun
22:47 * sri doesn't like lvalues much
22:48 jberger_ sri probably likes autoloaded tied lvalues worse
22:48 sri i use them with vec() and the like, but they always feel alien
22:48 jberger_ PDL uses them a lot
22:48 jberger_ assign a vector to a slice of a matrix
22:49 sri marcus: i don't really remember what i proposed
22:52 marcus sri: searching for binary in the irc logs yields nothing
22:55 * sri can't find the pastie either
22:55 marty sri:  I'm playing with bson_bin and I'm wondering if there is anything special I need to do to read them after retrieving them from the db.
22:55 sri might have been a Mojo::Util::censor or so that redacted non printable characters, and that was then used by the DEBUG warnings
22:56 sri marty: nope
22:56 marty hmmm ok.   Thank you
22:57 sri binary data is basically just a raw binary blob with a length prefix in bson
22:57 sri nothing gets changed
22:58 marty that's what I thougtht.  I have not tracked it down yet and I dont think it's a mango thing.  but I used to be able to load images and just read them back out in Imager but for some reason when I load/retrive them with mango I get an error.  wierd
22:59 marty and i just load the data with my $data = bson_bin($foo);
23:05 sri marty: i just ran a test and both MongoDB::BSON and Mango::BSON generate the exact same BSON for binaries
23:05 sri perl -MMongoDB -MMongoDB::BSON -E 'say unpack "H*", MongoDB::BSON::encode_bson({binary => \(my $dummy = "hello!")})'
23:05 sri perl -MMango::BSON=:bson -E 'say unpack "H*", bson_encode {binary => bson_bin("hello!")}'
23:06 jberger_ joined #mojo
23:08 marty Thanks, sri.  I have no idea what I am doing wrong but I'm sure it a face plant event.  :)
23:08 sri it's also roundtripping fine perl -Ilib -MMango::BSON=:bson -Mojo -E 'say r(bson_decode(pack "H*", "180000000562696e6172790006​0000000068656c6c6f2100"))'
23:10 marty I can confirm those test work perfectly.
23:28 ObseLeTe joined #mojo
23:34 sri hmm, ensure_index is harder than it looks
23:35 sri there are no rules for generating the index name, and all drivers have different semantics
23:40 marty sounds like an opportunity to be creative.  :)
23:51 sri marty: i think the main problem is that i'm not sure how index specs can look like in extreme cases
23:53 marty You mean like terrabytes of data sort of thing?
23:53 sri no, i mean the ensure_index arguments
23:54 sri http://api.metacpan.org/source/FRIEDO/Mo​ngoDB-0.503.3/lib/MongoDB/Collection.pm
23:54 sri look at to_index_string
23:58 ObseLeTe joined #mojo
23:58 sri tbh. ensure_index looks totally broken, when compared to the implementations in other drivers

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