| Time |
S |
Nick |
Message |
| 00:06 |
|
|
tholen42 left #mojo |
| 00:21 |
|
|
ruz left #mojo |
| 00:39 |
|
|
D4RK-PH0ENiX left #mojo |
| 00:59 |
|
|
SmokeMachine joined #mojo |
| 01:04 |
|
|
dku joined #mojo |
| 01:26 |
|
|
GitHub90 joined #mojo |
| 01:26 |
|
GitHub90 |
[mojo] kraih pushed 1 new commit to master: http://bit.ly/lO6pAH |
| 01:26 |
|
GitHub90 |
[mojo/master] fixed a few more application mount bugs and added more tests - Sebastian Riedel |
| 01:26 |
|
|
GitHub90 left #mojo |
| 01:27 |
|
sri |
phew... tough bugs |
| 01:38 |
|
sri |
looks like mount might grow into a full feature :) |
| 01:53 |
|
perlrocks |
Twitter: "Anyone check out Mojolicious' security? Comments?" --chrissullo http://twitter.com/chrissullo/[…]80815089317449728 |
| 02:03 |
|
|
SmokeMachine left #mojo |
| 02:22 |
|
perlrocks |
Twitter: "@henq none yet, but I think #mongodb is starting to become the "preferred" database for a lot of #mojolicious users" --perlninja http://twitter.com/perlninja/s[…]80822364711026689 |
| 02:23 |
|
|
dvinciguerra left #mojo |
| 02:38 |
|
|
GitHub113 joined #mojo |
| 02:38 |
|
GitHub113 |
[mojo] kraih pushed 1 new commit to master: http://bit.ly/iEGhip |
| 02:38 |
|
GitHub113 |
[mojo/master] added a few more application mount tests - Sebastian Riedel |
| 02:38 |
|
|
GitHub113 left #mojo |
| 03:39 |
|
|
tholen42 joined #mojo |
| 03:41 |
|
|
dku left #mojo |
| 03:43 |
|
crab |
hi |
| 04:09 |
|
sri |
ho |
| 04:27 |
|
|
Foxcool joined #mojo |
| 04:41 |
|
|
sugar joined #mojo |
| 05:02 |
|
|
dekimsey joined #mojo |
| 05:06 |
|
|
dekimsey1 left #mojo |
| 05:20 |
|
|
koban joined #mojo |
| 05:21 |
|
|
koban left #mojo |
| 05:30 |
|
crab |
it's really absurd that my production app (running on hypnotoad on a hosted server >100ms away) is significantly faster to use than my development version running with daemon --reload on localhost (a dual-core 1.6ghz atom laptop with 1gb of ram) |
| 05:30 |
|
|
metaperl joined #mojo |
| 05:30 |
|
crab |
someone must be punished for this. and i think it should be metaperl. |
| 05:31 |
|
metaperl |
who me???? |
| 05:31 |
|
purl |
you are confused but I never really fully idgested MOP and Moose, so what is Mojo and what is Stevan's voodoo is all greek to me or TBONE |
| 05:33 |
|
crab |
yes. don't act innocent. i know it's your fault. |
| 06:04 |
|
yko |
i think crab should be punished for running production server in --reload mode |
| 06:06 |
|
crab |
i think yko should be punished for not being able to read. |
| 06:07 |
|
marcus |
yko is actually ircing through google translate |
| 06:07 |
|
marcus |
:D |
| 06:07 |
|
yko |
hey! alligators are almost blind! you should know that! |
| 06:07 |
|
crab |
marcus: bahah |
| 06:08 |
|
|
d4rkie joined #mojo |
| 06:08 |
|
* marcus |
is trying to rent an apartment in sicilly for a 10 day trip in september. Just got the most interesting reply back: |
| 06:08 |
|
yko |
hmm. one day i should really try to irc through google translate |
| 06:08 |
|
marcus |
For the confirmation it has to send an account of €100.00 with banking credit transfer point out her the banking coordinates when I will receive the credit transfer we will send her an e-mail of confirmation: |
| 06:11 |
|
yko |
looks like google translated ja? |
| 06:11 |
|
marcus |
'it rubs the lotion on it's skin, or it gets the hose again'. |
| 06:15 |
|
|
sugar left #mojo |
| 06:17 |
|
yko |
hehe, you broke my brain. never heard that in english but it sounds so known. god save the google |
| 06:19 |
|
marcus |
yko: silence of the lambs. |
| 06:19 |
|
yko |
you should like Ukraine when Euro-2012 begins. taxi drivers will speak in similar manner and ask for at least €300 |
| 06:19 |
|
marcus |
yko: http://www.youtube.com/watch?v=tDgS6qLsVM4 |
| 06:20 |
|
yko |
yeah, already found it. god save google |
| 06:26 |
|
|
AmeliePoulain joined #mojo |
| 06:39 |
|
metaperl |
god save http://www.DuckDuckGo.com |
| 06:56 |
|
crab |
i wonder if i should add this onsubmit="<disable submit button>" bit of js on all my forms. |
| 07:12 |
|
perlrocks |
Twitter: "Mojolicious::Plugin::SslAuth 0.05: SSL client certificate auth helper http://twurl.nl/0535r4" --lordnaastik http://twitter.com/lordnaastik[…]80895312255922176 |
| 07:12 |
|
perlrocks |
Twitter: "Mojolicious::Plugin::SslAuth 0.05: SSL client certificate auth helper http://twurl.nl/oxcue4" --lordnaastik http://twitter.com/lordnaastik[…]80895327267340288 |
| 07:35 |
|
|
bosphorus joined #mojo |
| 07:35 |
|
|
Sugar joined #mojo |
| 07:35 |
|
|
koban` joined #mojo |
| 07:42 |
|
perlrocks |
Twitter: "New release: Mojolicious::Plugin::SslAuth 0.05 http://bit.ly/jMkrgJ" --softpedialinux http://twitter.com/softpediali[…]80903020216451072 |
| 07:42 |
|
perlrocks |
Twitter: "Mojolicious::Plugin::SslAuth 0.05: SSL client certificate auth helper http://bit.ly/jxK1b7" --J03_PY http://twitter.com/J03_PY/stat[…]80903021395062784 |
| 08:14 |
|
|
spleenjack joined #mojo |
| 08:43 |
|
|
marcus_ joined #mojo |
| 09:02 |
|
marcus_ |
good morning nurse |
| 09:03 |
|
|
marcus left #mojo |
| 09:03 |
|
|
marcus_ is now known as marcus |
| 09:43 |
|
crab |
nurse? |
| 09:43 |
|
purl |
hmmm... nurse is soo sweet =) |
| 09:44 |
|
|
fmerges joined #mojo |
| 09:54 |
|
marcus |
crab: http://www.youtube.com/watch?v=X6-UftBHk9A |
| 09:54 |
|
marcus |
hrm, irccloud is pretty darn cool. |
| 10:01 |
|
|
Foxcool left #mojo |
| 10:05 |
|
|
Foxcool joined #mojo |
| 10:29 |
|
|
psiman joined #mojo |
| 10:32 |
|
|
psimanx1 left #mojo |
| 10:32 |
|
|
psimanx1 joined #mojo |
| 10:37 |
|
|
psiman left #mojo |
| 10:46 |
|
|
Akron joined #mojo |
| 11:34 |
|
|
fhelmberger left #mojo |
| 11:35 |
|
|
fhelmberger joined #mojo |
| 11:37 |
|
|
fmerges left #mojo |
| 11:41 |
|
|
psimanx1 left #mojo |
| 11:58 |
|
chansen |
I just signed up for an invite, seems like a cool idea |
| 11:59 |
|
|
janus left #mojo |
| 12:03 |
|
koban` |
http://blog.zawodny.com/2011/0[…]d-learning-rails/ |
| 12:04 |
|
|
bosphoru_ joined #mojo |
| 12:09 |
|
|
bosphorus left #mojo |
| 12:17 |
|
|
dotan joined #mojo |
| 12:21 |
|
|
Sugar|2 joined #mojo |
| 12:25 |
|
|
Sugar left #mojo |
| 12:43 |
|
marcus |
koban: So very fucking cool! |
| 12:48 |
|
sri |
crab: wait... --reload is slow? SHOCKING! |
| 12:49 |
|
|
bosphoru_ left #mojo |
| 12:51 |
|
perlrocks |
Twitter: "Very cool to see that @jzawodn is a #Mojolicious user. # http://t.co/HoxLC3p" --marcusramberg http://twitter.com/marcusrambe[…]80980649909039104 |
| 12:51 |
|
|
bosphorus joined #mojo |
| 12:51 |
|
marcus |
Aslo, interesting to see that everybody is recommending $favorite_framework in the comments of that post. |
| 12:51 |
|
marcus |
Also even |
| 12:51 |
|
sri |
he's even here :o |
| 12:52 |
|
* sri |
waves at jzawodn |
| 12:53 |
|
sri |
btw. anyone tested the mount plugin? |
| 12:55 |
|
marcus |
sri: sorry, been too busy mounting other things |
| 12:55 |
|
marcus |
:p |
| 12:55 |
|
|
baton8 left #mojo |
| 12:55 |
|
|
sh9 joined #mojo |
| 12:56 |
|
sri |
broken filesystems? |
| 12:56 |
|
marcus |
http://nordaaker.com/docs/skit[…]110615-145546.jpg thing I will stick with irccloud for a while. |
| 12:56 |
|
marcus |
*think |
| 12:57 |
|
|
SmokeMachine joined #mojo |
| 12:58 |
|
sri |
hmm, irc in the cloud... that's crazy talk |
| 12:59 |
|
moritz |
it doesn't look more cloudy to me than other server that offers an IRC interface |
| 12:59 |
|
sri |
does it stay conected when you disconnect? |
| 13:00 |
|
elb0w |
man I wanted to do that |
| 13:00 |
|
elb0w |
and sell it to financial tech divisions |
| 13:00 |
|
elb0w |
as a support tool |
| 13:00 |
|
|
baton8 joined #mojo |
| 13:01 |
|
|
Foxcool left #mojo |
| 13:02 |
|
|
Foxcool joined #mojo |
| 13:06 |
|
MisterHatt |
elb0w: friend of mine sells a drac communication tool, all it does is install irssi on the proc and autoconnect, all system logs get sent to it is all, lol |
| 13:06 |
|
MisterHatt |
allows for "real time log analysis on drac" |
| 13:06 |
|
MisterHatt |
such a load of shit~ |
| 13:06 |
|
elb0w |
yeah, larger firms have a lot of compliance rules needed |
| 13:07 |
|
MisterHatt |
word |
| 13:10 |
|
|
metaperl left #mojo |
| 13:22 |
|
|
mattastrophe joined #mojo |
| 13:29 |
|
crab |
sri: i'd rather think that hypnotoad is fast, personally |
| 13:29 |
|
sri |
crab: production vs development mode also makes a huge difference |
| 13:29 |
|
crab |
anyway, i like having a devel installation that is slower than the real thing. keeps me honest when it comes to performance. |
| 13:29 |
|
|
psimanx1 joined #mojo |
| 13:29 |
|
crab |
sri: i know |
| 13:29 |
|
sri |
:) |
| 13:30 |
|
sri |
but yea, i'm quite happy with hypnotoad too |
| 13:30 |
|
sri |
ALL GLORY TO THE HYPNOTOAD! |
| 13:30 |
|
|
sh9 left #mojo |
| 13:31 |
|
* marty |
yawn, good morning every.... ALL GLORY TO THE HYPNOTOAD! |
| 13:31 |
|
marty |
wow, what a way to start the day. :) |
| 13:34 |
|
marcus |
sri: Yes |
| 13:35 |
|
sri |
so you're getting rid of all your coins in norway? |
| 13:35 |
|
moritz |
all hail the bitcoin? |
| 13:36 |
|
sri |
heh |
| 13:36 |
|
|
AmeliePoulain left #mojo |
| 13:37 |
|
|
bosphoru_ joined #mojo |
| 13:37 |
|
crab |
misterhatt: what is "drac"? |
| 13:40 |
|
|
bosphorus left #mojo |
| 13:45 |
|
marcus |
sri: Yes, irccloud stays connected when I disconnect |
| 13:48 |
|
|
metaperl joined #mojo |
| 13:51 |
|
sri |
neat |
| 13:52 |
|
MisterHatt |
crab: dell's remote access console |
| 13:52 |
|
MisterHatt |
shunts vga output and power control and keyboard/mouse into an ipmi interface that you can access over a java or activex controller |
| 13:53 |
|
|
Sugar|2 is now known as Sugar |
| 13:53 |
|
MisterHatt |
handy way of remote control of hardware/software on servers |
| 13:53 |
|
MisterHatt |
horrendous to use though, and if you're running xen on it, it gets a bit tricky, hence being able to stream logs in realtime based on where in drac they're coming from |
| 13:54 |
|
MisterHatt |
not all that diff to regular vm's I guess aside from needing a way to access it all via the java app, an d most corporate envs are run by people too dumb to know what irssi is |
| 13:54 |
|
MisterHatt |
amazing that he gets away with it really |
| 13:54 |
|
* MisterHatt |
sleeps |
| 14:01 |
|
|
mattastrophe left #mojo |
| 14:34 |
|
crab |
random access console |
| 14:34 |
|
crab |
ssh to it and you never know which system you'll end up controlling! |
| 14:36 |
|
|
bangbang joined #mojo |
| 14:37 |
|
|
Foxcool left #mojo |
| 14:38 |
|
|
bosphoru_ left #mojo |
| 14:46 |
|
|
koban` left #mojo |
| 14:56 |
|
|
Foxcool joined #mojo |
| 14:57 |
|
|
fmerges joined #mojo |
| 15:01 |
|
perlrocks |
Twitter: "Blogs written by # perl # mojolicious / Mojolicious to deploy a web design - jamadam weblog2 http://htn.to/cPneFE" (ja) --jamadam http://twitter.com/jamadam/sta[…]81013516454207488 |
| 15:14 |
|
|
D4RK-PH0ENiX joined #mojo |
| 15:14 |
|
|
kaare joined #mojo |
| 15:18 |
|
|
d4rkie left #mojo |
| 15:20 |
|
|
metaperl_ joined #mojo |
| 15:25 |
|
|
metaperl left #mojo |
| 15:29 |
|
|
sh9 joined #mojo |
| 15:44 |
|
|
fmerges left #mojo |
| 15:45 |
|
|
metaperl_ left #mojo |
| 15:46 |
|
|
metaperl_ joined #mojo |
| 15:49 |
|
|
Sugar left #mojo |
| 15:51 |
|
|
D4RK-PH0ENiX left #mojo |
| 15:52 |
|
|
d4rkie joined #mojo |
| 15:53 |
|
|
Akron left #mojo |
| 15:56 |
|
perlrocks |
Twitter: "Deploying a web design Mojolicious - jamadam weblog2 http://bit.ly/lGJt8p" (ja) --wPerlnews http://twitter.com/wPerlnews/s[…]81027200689774592 |
| 15:57 |
|
sri |
funny, even deep nested mounting works |
| 15:58 |
|
sri |
just tried 4 nested apps |
| 15:59 |
|
sri |
interesting |
| 15:59 |
|
sri |
the cost doesn't go up much after the first layer |
| 16:01 |
|
sri |
and there are a lot of low hanging fruits... lots of potential should anyone need more performance |
| 16:07 |
|
|
Foxcool left #mojo |
| 16:09 |
|
|
d4rkie left #mojo |
| 16:23 |
|
|
dotan left #mojo |
| 16:25 |
|
|
Foxcool joined #mojo |
| 16:31 |
|
|
GitHub138 joined #mojo |
| 16:31 |
|
GitHub138 |
[mojo] kraih pushed 1 new commit to master: http://bit.ly/kipvfS |
| 16:31 |
|
GitHub138 |
[mojo/master] updated prettify.js to version 1-Jun-2011 - Sebastian Riedel |
| 16:31 |
|
|
GitHub138 left #mojo |
| 16:38 |
|
metaperl_ |
so I have this line: $tx = $self->ua->post_form($url => $self->req->body_params->to_hash); # and I want to get the json that was returned from the post ... $tx->res is the start of the object navigation... but how would I get the json value for a key named 'patientname' for instance |
| 16:40 |
|
sri |
->res->json->{patientname}? |
| 16:41 |
|
sri |
http://mojolicio.us/perldoc/Mojo/Message#json |
| 16:47 |
|
metaperl_ |
sri was it illogical of me to look here for a json method - http://search.cpan.org/~kraih/[…]ssage/Response.pm |
| 16:47 |
|
sri |
yes it was |
| 16:48 |
|
|
sh9 left #mojo |
| 16:48 |
|
sri |
json deserialization of course affects both sides |
| 16:49 |
|
sri |
very little functionality is unique to one side |
| 16:51 |
|
|
SmokeMachine left #mojo |
| 16:54 |
|
perlrocks |
Twitter: "@freekey Which software did you use for the #Mojolicious slides? I'm preparing some slides for my final degree project, and PowerPoint sucks" --miquelruiz http://twitter.com/miquelruiz/[…]81041978061631488 |
| 16:57 |
|
|
SmokeMachine joined #mojo |
| 17:01 |
|
perlrocks |
Twitter: "@miquelruiz I've used showoff with a custom CSS mostly copied from mojolicious socumentation :-) http://bit.ly/iK6MG6 http://bit.ly/irNxH7" --freekey http://twitter.com/freekey/sta[…]81043548924944384 |
| 17:02 |
|
|
dekimsey1 joined #mojo |
| 17:02 |
|
|
SmokeMachine left #mojo |
| 17:02 |
|
|
metaperl_ left #mojo |
| 17:03 |
|
|
metaperl_ joined #mojo |
| 17:04 |
|
perlrocks |
Twitter: "Indeed…installed NextGen web framework for #Perl http://ub0.cc/6m/Yq ☝ #GNU/ #Linux #Debian: sh -c "curl -L cpanmin.us | perl - Mojolicious"" --Metztli_IT http://twitter.com/Metztli_IT/[…]81044425668689920 |
| 17:06 |
|
|
dekimsey left #mojo |
| 17:10 |
|
crab |
what mojolicious slides, i wonder? |
| 17:11 |
|
sri |
yea, they are not in the wiki! |
| 17:11 |
|
* sri |
pokes diegok |
| 17:11 |
|
diegok |
sri: hi |
| 17:12 |
|
* sri |
waves |
| 17:15 |
|
diegok |
sri: why!? |
| 17:15 |
|
sri |
you forgot to link your slides in the wiki! |
| 17:16 |
|
diegok |
sri: oh!, are in spanish, but, ok, I will :) |
| 17:16 |
|
sri |
\o/ |
| 17:18 |
|
|
bangbang left #mojo |
| 17:20 |
|
|
spleenjack left #mojo |
| 17:21 |
|
|
kaare left #mojo |
| 17:22 |
|
|
metaperl joined #mojo |
| 17:23 |
|
|
sugar joined #mojo |
| 17:27 |
|
|
metaperl_ left #mojo |
| 17:30 |
|
|
zipkiddy joined #mojo |
| 17:30 |
|
|
zipkiddy left #mojo |
| 17:30 |
|
|
zipkiddy joined #mojo |
| 17:31 |
|
crab |
is diegok == freekey? |
| 17:31 |
|
sri |
ye |
| 17:32 |
|
diegok |
crab: :) |
| 17:37 |
|
|
Foxcool left #mojo |
| 17:38 |
|
|
zipkiddy left #mojo |
| 17:50 |
|
perlrocks |
Twitter: "Deploying a web design Mojolicious - jamadam weblog2 "web development" and "web design" I wonder if a different genre. Uebuapurikeshonfu trendy to have a living trust primarily to the production of ... http://bit.ly/irxXDs Koporetouebu" (ja) --webdesignStatio http://twitter.com/webdesignSt[…]81055942078246912 |
| 17:51 |
|
|
sh9 joined #mojo |
| 17:57 |
|
|
Foxcool joined #mojo |
| 18:02 |
|
|
mattastrophe joined #mojo |
| 18:20 |
|
|
mattastrophe left #mojo |
| 18:21 |
|
|
fhelmberger left #mojo |
| 18:22 |
|
|
mattastrophe joined #mojo |
| 18:23 |
|
|
trone joined #mojo |
| 18:25 |
|
|
GabrielVieira2 left #mojo |
| 18:28 |
|
|
trone_ left #mojo |
| 18:31 |
|
|
Sjors joined #mojo |
| 18:31 |
|
Sjors |
Hey guys |
| 18:32 |
|
Sjors |
I've written my first Mojo application but I can't get something to work |
| 18:32 |
|
Sjors |
A prerequisite for the application I'm developing is that it can "stream" HTML |
| 18:32 |
|
Sjors |
i.e. you write the first part of the HTML, then do something else which takes a few seconds, print some more HTML, do something else, print some HTML, do something, then print the last bit of HTML |
| 18:33 |
|
Sjors |
so I've written a template which simply prints "<p>foo</p>"; % sleep 1; "<p>bar</p>"; %sleep 1 and so on |
| 18:33 |
|
Sjors |
it turns out the page load simply takes a few seconds, then all HTML is printed as-is |
| 18:33 |
|
sri |
Sjors: http://mojolicio.us/perldoc/Mo[…]/Controller#write |
| 18:33 |
|
Sjors |
is there some way I can tell Mojolicious to return the HTML as it is being generated? |
| 18:34 |
|
sri |
the template system does not support streaming itself, you have to break down everything in smaller ones and write them yourself |
| 18:35 |
|
Sjors |
is there a way I can still use "most" of the template system, so I can let the pages look exactly alike? |
| 18:35 |
|
Sjors |
or does the template system have the problem where the page needs to be completed in order to send it back? |
| 18:35 |
|
sri |
yes, use small partial templates |
| 18:35 |
|
sri |
you can't use features like layouts |
| 18:35 |
|
Sjors |
at all? :( |
| 18:36 |
|
Sjors |
I liked those |
| 18:36 |
|
sri |
doesn't work for streaming |
| 18:37 |
|
sri |
we would need coroutines/fibers to make it "just work", but perl doesn't have those |
| 18:38 |
|
sri |
you'll need something like "$self->write($self->render('header', partial => 1), sub {...})" |
| 18:38 |
|
Sjors |
hmm, what do you mean? |
| 18:38 |
|
Sjors |
yeah |
| 18:39 |
|
sri |
no other way |
| 18:39 |
|
Sjors |
let the template system start writing its output before it's completed |
| 18:39 |
|
Sjors |
i.e. in this case: |
| 18:40 |
|
Sjors |
http://pastebin.com/2ETEdaSx |
| 18:40 |
|
Sjors |
let it print as much of the default layout as it can |
| 18:40 |
|
sri |
that's impossible |
| 18:41 |
|
Sjors |
mm yeah, I think I see what you mean |
| 18:41 |
|
Sjors |
it has to parse index.html.ep before it knows <%= title =%> |
| 18:41 |
|
Sjors |
and it has to parse it before it even knows what layout to use |
| 18:41 |
|
sri |
you can't assume that you have a direct blocking socket connection to the client |
| 18:42 |
|
Sjors |
that's not a problem, there's a HTTP standard for this |
| 18:42 |
|
sri |
because unless you're programming PHP you don't |
| 18:42 |
|
sri |
HTTP has nothing to do with this |
| 18:42 |
|
Sjors |
http://weblog.rubyonrails.org/[…]hy-http-streaming |
| 18:42 |
|
Sjors |
not that I've done any Ruby on Rails, but that page describes it nicely |
| 18:42 |
|
sri |
ok, do you actually know how rails does it? |
| 18:42 |
|
sri |
they have Fibers |
| 18:43 |
|
sri |
they have cheap concurrency in Ruby |
| 18:43 |
|
sri |
if we had those we could do it too, but we don't |
| 18:43 |
|
Sjors |
hmm, is this a fundamental Perl problem? |
| 18:43 |
|
sri |
i guess you could say that |
| 18:43 |
|
|
stephan48 left #mojo |
| 18:44 |
|
sri |
but we do support streaming, just not as dwim looking |
| 18:45 |
|
Sjors |
I implemented something in Perl a while ago |
| 18:45 |
|
Sjors |
sec |
| 18:46 |
|
Sjors |
I experimented with implementing something like fifo pipes using tied handles in Perl |
| 18:46 |
|
|
bangbang joined #mojo |
| 18:46 |
|
sri |
we actually have better streaming support than rails, we just can't do the high level abstraction in Perl |
| 18:46 |
|
Sjors |
I wrote a class which returns a tied handle, then everytime you call readline on it, it returns the next result as an object |
| 18:47 |
|
Sjors |
it works quite beautifully |
| 18:47 |
|
sri |
how is that related? |
| 18:47 |
|
Sjors |
hold on, I'm going there ;-) |
| 18:48 |
|
|
stephan48 joined #mojo |
| 18:48 |
|
Sjors |
I have no clue what's currently missing in the template engine to make it work, but it sounds to _me_ (considering I've never even looked at the Mojo sources) that there's something like a chicken-and-egg problem |
| 18:48 |
|
Sjors |
without completely generating the template, you can't substitute the placeholders in the layout with the results of the template, etc |
| 18:48 |
|
Sjors |
and you need all the results until the template can be generated, so the page is only done when all the results are in |
| 18:49 |
|
sri |
that's not the problem |
| 18:49 |
|
sri |
problem is context switching for io |
| 18:49 |
|
Sjors |
if this is the problem, it could help (I'm just suggesting) to not /store/ the results of the template, but initialise the layout and simply feed them as they go, printing the final results as they are finalized |
| 18:49 |
|
Sjors |
hmm |
| 18:49 |
|
Sjors |
could you elaborate? |
| 18:49 |
|
sri |
we need to preserve state in one thread while the event loop writes everything we have buffered |
| 18:49 |
|
Sjors |
oh it's working with threads? |
| 18:49 |
|
Sjors |
ouch |
| 18:50 |
|
sri |
threads/coroutines/fibers... all works |
| 18:50 |
|
sri |
that's how rails did it |
| 18:50 |
|
Sjors |
it doesn't sound impossible to me in Perl, and it doesn't sound impossible to me to do it in a nice (as you said dwim) way too |
| 18:51 |
|
sri |
patches welcome :D |
| 18:51 |
|
bangbang |
famous last words |
| 18:51 |
|
purl |
please hold while I connect you. |
| 18:51 |
|
Sjors |
but I'd have to look at the Mojo sources to see if I can figure out where the problem is now |
| 18:51 |
|
sri |
you would be my hero if you make it work |
| 18:51 |
|
Sjors |
yeah, it sounds like a valuable addition which would be worth patching |
| 18:52 |
|
sri |
all the primitives are there with write and write_chunk |
| 18:52 |
|
sri |
if it can be done without threads we have the tools in place |
| 18:52 |
|
Sjors |
for now I'll try to continue my application without streaming, and when it becomes a problem I'm going to look at Mojo and where I could make improvements |
| 18:52 |
|
Sjors |
I don't think threads are at all necessary here :) |
| 18:52 |
|
Sjors |
though they could make stuff faster once it's done without threads |
| 18:53 |
|
Sjors |
that's a job for someone else, then. ;-) |
| 18:53 |
|
sri |
i think you're missing the context switching part |
| 18:53 |
|
Sjors |
yeah me too |
| 18:53 |
|
sri |
that's what makes it tricky |
| 18:53 |
|
Sjors |
didn't understand what you meant there |
| 18:54 |
|
sri |
a write in mojolicious is not blocking |
| 18:55 |
|
sri |
it will be passed to the event loop, which will then take over control and call the drain callback once everything has reached the kernel send buffer |
| 18:55 |
|
sri |
you get control back via drain callback |
| 18:57 |
|
sri |
a template is a perl sub, you can't pass control back to the event loop and stay in it, you have to return first and abort everything |
| 18:58 |
|
sri |
if you did you would just get deep recursion issues at some point |
| 18:58 |
|
|
sh9 left #mojo |
| 18:58 |
|
Sjors |
however |
| 18:58 |
|
Sjors |
you do have tied IO handles |
| 18:59 |
|
sri |
i think the important information here is that even if $self->write('foobar'); returns doesn't mean anything has been written |
| 18:59 |
|
Sjors |
you can set them as the default out handle (replacing stdout), then a function inside the tied class will be called when something is printed |
| 18:59 |
|
Sjors |
if you do this, the handler could easily call $self->write_chunk() if it has access to $self |
| 18:59 |
|
sri |
i don't see at all how tied handles help in any way here |
| 19:00 |
|
sri |
Sjors: that doesn't change anything |
| 19:00 |
|
Sjors |
I think I'm not seeing the complete problem |
| 19:00 |
|
sri |
$self->write_chunk('foobar') will return even if the whole chunk is still buffered |
| 19:01 |
|
Sjors |
ah |
| 19:01 |
|
Sjors |
so everything is already directly written in chunks? |
| 19:01 |
|
sri |
non-blocking io, we don't know for sure until the drain callback has been called |
| 19:01 |
|
Sjors |
and if you allow an option to set it to blocking i/o? |
| 19:02 |
|
sri |
impossible |
| 19:02 |
|
purl |
You must mean impassable - *nothing's* impossible! |
| 19:02 |
|
Sjors |
'cause in this case, the kernel will drop it on the network socket, which is in-memory at this point, so I don't really care about the time my write_chunk takes |
| 19:03 |
|
sri |
no, it might still be in our event loop |
| 19:03 |
|
sri |
not the kernel send buffer |
| 19:03 |
|
Sjors |
especially since my processing takes long enough for it to matter less than if I have to wait until the full processing is done |
| 19:03 |
|
Sjors |
hmm |
| 19:03 |
|
Sjors |
and if there's a call for "flush the event loop right now, no matter how long it takes"? |
| 19:03 |
|
marcus |
yo |
| 19:03 |
|
Sjors |
I understand the event loop also contains calls like "print $foo to the networking socket"? |
| 19:04 |
|
sri |
the event loop will be accepting new connections in the meantime |
| 19:04 |
|
sri |
or handling the other few thousand it already accepted |
| 19:04 |
|
|
sh9 joined #mojo |
| 19:04 |
|
sri |
you can't block |
| 19:04 |
|
Sjors |
oh, hmm, right |
| 19:05 |
|
Sjors |
so the event loop both contains all intermediate results my templates have come up with (i.e. while a template which contains some sleeps is halfway through processing), and all new connections, etc? |
| 19:05 |
|
Sjors |
an option that any print / print_chunks are not even entered into the event loop but handled immediately? |
| 19:06 |
|
Sjors |
would that help? |
| 19:06 |
|
sri |
on platforms that support it the event loop with attempt a direct write |
| 19:07 |
|
sri |
but there are no guarantees, you never know if it works |
| 19:07 |
|
Sjors |
so for some platforms, a template with a couple of sleeps would write in chunked encoding, and for others, it wouldn't? |
| 19:07 |
|
Sjors |
or is there still some work to be done for that? |
| 19:08 |
|
sri |
sleep does absolutely nothing except hurt your app |
| 19:08 |
|
Sjors |
I know - it was only a test |
| 19:08 |
|
sri |
it's the most useless thing you could possibly do ;p |
| 19:08 |
|
sri |
btw. no it would not write in chunked encoding by default |
| 19:08 |
|
Sjors |
it's a good test at that: if the application supports chunked sending, sleep is the best way to freeze the whole application, so if your output appears between the calls to sleep, you're sure your lines are printed constantly ;-) |
| 19:09 |
|
Sjors |
it's not my application will eventually use sleep too :P |
| 19:09 |
|
Sjors |
but it's an easy way to simulate "I'm too busy for a few seconds to give control back to the framework" |
| 19:12 |
|
sri |
anyway, i don't think it can be done properly in perl atm, but please prove me wrong :) |
| 19:13 |
|
Sjors |
I would certainly like to do so ;-) |
| 19:13 |
|
Sjors |
I'm going to continue my application without chunked encoding and once it's a problem, I'll start looking at what actually causes it |
| 19:13 |
|
Sjors |
maybe in the meantime it will already be fixed by someone else, but if not, I'll definitely give it a try |
| 19:15 |
|
* marcus |
roots for Sjors |
| 19:16 |
|
marcus |
http://seldo.com/weblog/2011/0[…]is_an_antipattern # think I agree with this |
| 19:16 |
|
|
preflex left #mojo |
| 19:19 |
|
|
preflex joined #mojo |
| 19:23 |
|
* sri |
grabs a root beer |
| 19:24 |
|
* marcus |
grabs a real beer |
| 19:26 |
|
* metaperl |
continues with day 16 of his orange juice mono-diet |
| 19:28 |
|
Sjors |
will <%= "\n" %> also write a HTML <br/> tag> |
| 19:28 |
|
Sjors |
?* |
| 19:31 |
|
moritz |
no |
| 19:32 |
|
moritz |
btw only xhtml needs to close br tags |
| 19:33 |
|
Sjors |
my eyes will start bleeding if I see HTML that doesn't look like XHTML, even if it's HTML5, let alone if I write it myself :( |
| 19:34 |
|
Sjors |
what would be the easiest way to not only escape HTML content, but also replace all \n by <br/> ? |
| 19:35 |
|
Sjors |
I can simply s/\n/<br\/>/g in <%== but that will not escape the content, of course |
| 19:35 |
|
sri |
make a helper |
| 19:36 |
|
sri |
then it can look like <%=x "lalala\nlalala" %> |
| 19:37 |
|
Sjors |
ah |
| 19:37 |
|
Sjors |
cool |
| 19:47 |
|
|
GabrielVieira joined #mojo |
| 19:55 |
|
perlrocks |
Twitter: "Deploying a web design Mojolicious - jamadam weblog2 http://htn.to/LYaNNa" (ja) --yukaikousai http://twitter.com/yukaikousai[…]81087341908856832 |
| 20:02 |
|
|
sh9 left #mojo |
| 20:08 |
|
elb0w |
sri: you around? |
| 20:11 |
|
|
amoore joined #mojo |
| 20:11 |
|
|
sugar left #mojo |
| 20:27 |
|
elb0w |
who is the resident web socket expert here? |
| 20:32 |
|
moritz |
sri is! |
| 20:33 |
|
elb0w |
Im just trying to understand, so I would create a websocket at a location say '/' then from javascript connect to ws://site.com/ |
| 20:33 |
|
elb0w |
and then what haha |
| 20:33 |
|
elb0w |
just keep it in a loop that sends data? |
| 20:34 |
|
bangbang |
thats gonna leave a mark |
| 20:34 |
|
bangbang |
ouch |
| 20:40 |
|
|
amoore left #mojo |
| 20:44 |
|
sri |
elb0w: send and receive messages, websockets are bi-directional |
| 20:45 |
|
Sjors |
I just saw the Mojo daemon continues to generate a page even if the HTTP socket which did the actual request is closed |
| 20:46 |
|
Sjors |
that sounds DDOS-able... |
| 20:46 |
|
Sjors |
do a million intensive requests, and just close the connection before you can pick them up; the daemon will continue processing |
| 20:46 |
|
Sjors |
I imagine this is hard to fix when the request is "somewhere" in the event loop |
| 20:51 |
|
sri |
Sjors: everything facing the internet is DDOS-able |
| 20:56 |
|
sri |
for streaming that's not true btw |
| 20:56 |
|
sri |
only for monolithic responses |
| 20:57 |
|
sri |
and i doubt you expect us to kill the pocess :) |
| 20:58 |
|
elb0w |
how do I generate a lite app again? |
| 20:58 |
|
sri |
closing the connection also doesn't make a DDOS attack that much more effective, you can DDOS just as well with a million open ones |
| 20:58 |
|
* sri |
pushes elb0w into the alligator pit |
| 20:58 |
|
elb0w |
nvm |
| 20:58 |
|
elb0w |
:D |
| 21:02 |
|
* stephan48 |
throws elb0w out... no more room! |
| 21:09 |
|
Sjors |
sri: oh, true, you're right |
| 21:13 |
|
elb0w |
I am seeing the connection but my onopen does not fire |
| 21:14 |
|
sri |
elb0w: which browser? |
| 21:14 |
|
purl |
which browser is being stupid? ie? |
| 21:14 |
|
elb0w |
chrome |
| 21:15 |
|
sri |
mojo version? |
| 21:15 |
|
elb0w |
updating now |
| 21:15 |
|
sri |
won't work |
| 21:15 |
|
sri |
different protocol versions |
| 21:15 |
|
elb0w |
gotcha |
| 21:15 |
|
sri |
you need firefox aurora |
| 21:15 |
|
elb0w |
Really? |
| 21:15 |
|
sri |
mojo is on ietf-09 now |
| 21:16 |
|
sri |
chrome is draft-76 |
| 21:16 |
|
sri |
very very old |
| 21:17 |
|
elb0w |
is auroa a beta? |
| 21:17 |
|
elb0w |
or 4.0.1? |
| 21:17 |
|
purl |
i guess 4.0.1 is a bit scared |
| 21:17 |
|
sri |
6.0 beta |
| 21:17 |
|
sri |
or you use mojolicious 1.16, which was the last draft-76 release for us |
| 21:18 |
|
elb0w |
There a easy way to downgrade? |
| 21:20 |
|
sri |
downgrading completely is a bad idea |
| 21:21 |
|
sri |
for experimenting you should just bundle 1.16 |
| 21:24 |
|
elb0w |
Yeah so I dont even see it connecting using aurora |
| 21:24 |
|
elb0w |
but I did with chrome |
| 21:27 |
|
sri |
ok time to step back |
| 21:28 |
|
sri |
run "mojo daemon" and connect with aurora to "http://127.0.0.1:3000/diag/websocket" |
| 21:28 |
|
yko |
sri stop pushing people here! there's not so much space as you can imagine! |
| 21:29 |
|
sri |
yko: maybe you should stop slacking then and eat more people? |
| 21:29 |
|
yko |
yum-yum |
| 21:29 |
|
yko |
anyway, i should do something with bones |
| 21:29 |
|
sri |
elb0w: then you'll know if websockets actually work |
| 21:32 |
|
tempire |
behold, my 512 ssd lives |
| 21:32 |
|
tempire |
someone behold it |
| 21:32 |
|
tempire |
go |
| 21:32 |
|
* marty |
beholds |
| 21:32 |
|
* marty |
bows to tempire's 512 ssd |
| 21:34 |
|
* sri |
beholds |
| 21:35 |
|
* tempire |
bestows its greatness to the mojicans. |
| 21:35 |
|
sri |
i was a bit scared earlier that my ssd might have died... macbook was hanging at they greyscreen for like 15 minutes :S |
| 21:35 |
|
tempire |
what was the problem? |
| 21:35 |
|
sri |
*-y |
| 21:36 |
|
marty |
he forgot to plug it in |
| 21:36 |
|
sri |
no idea, now it just works again |
| 21:36 |
|
marty |
:P |
| 21:36 |
|
sri |
timemachine backups make it a lot less scary though |
| 21:38 |
|
|
janus joined #mojo |
| 21:45 |
|
|
bangbang left #mojo |
| 21:52 |
|
marcus |
https://github.com/mberends/http-server-simple/ |
| 21:53 |
|
tempire |
ooh |
| 21:54 |
|
tempire |
is there a website for rakudo modules? |
| 21:54 |
|
tempire |
hmm, the proto tool |
| 21:55 |
|
tempire |
and rakudo is in homebrew |
| 21:55 |
|
tempire |
sweet |
| 21:59 |
|
sri |
didn't compile last i tried installing it |
| 22:04 |
|
|
jwang left #mojo |
| 22:25 |
|
|
metaperl_ joined #mojo |
| 22:29 |
|
|
metaperl left #mojo |
| 22:34 |
|
tempire |
rakudo-star compiles for me |
| 22:36 |
|
sri |
that server is horrible |
| 22:37 |
|
sri |
even for a proof of concept i would expect a little more |
| 22:39 |
|
* sri |
really should stop looking at perl6... so depressing |
| 22:41 |
|
|
mmcleric left #mojo |
| 22:42 |
|
|
metaperl_ left #mojo |
| 22:59 |
|
marty |
is perl 6 really that bad? |
| 23:20 |
|
|
jwang joined #mojo |
| 23:28 |
|
|
GitHub85 joined #mojo |
| 23:28 |
|
GitHub85 |
[mojo] kraih pushed 1 new commit to master: https://github.com/kraih/mojo/[…]cce62410c3ece4a9b |
| 23:28 |
|
GitHub85 |
[mojo/master] improved Mojo::Message documentation - Sebastian Riedel |
| 23:28 |
|
|
GitHub85 left #mojo |
| 23:29 |
|
sri |
marty: bad might not be the right term to describe it |
| 23:29 |
|
sri |
the language itself is great |
| 23:30 |
|
sri |
the implementation just doesn't make any visible progress |
| 23:31 |
|
sri |
from the outside it just looks like they are starting over and over again |
| 23:31 |
|
sri |
last blog post i saw was about yet another new implementation |
| 23:32 |
|
sri |
last time it was about using more parrot, this time about using less parrot |
| 23:33 |
|
sri |
my impression is that perl6 itself is a lost cause, but can be a good source for new perl5 feature ideas |
| 23:42 |
|
|
mattastrophe left #mojo |
| 23:46 |
|
|
purl left #mojo |
| 23:46 |
|
|
baton8 left #mojo |
| 23:46 |
|
|
marcus left #mojo |
| 23:46 |
|
|
memowe left #mojo |
| 23:46 |
|
|
k-man left #mojo |
| 23:46 |
|
|
alnewkirk|com left #mojo |
| 23:46 |
|
|
yakudzo left #mojo |
| 23:46 |
|
|
marty left #mojo |
| 23:46 |
|
|
trcjr left #mojo |
| 23:46 |
|
|
yko left #mojo |
| 23:46 |
|
|
gbacon left #mojo |
| 23:46 |
|
|
gabb left #mojo |
| 23:46 |
|
|
espent_ left #mojo |
| 23:46 |
|
|
memememomo left #mojo |
| 23:46 |
|
|
jwang left #mojo |
| 23:46 |
|
|
preflex left #mojo |
| 23:46 |
|
|
stephan48 left #mojo |
| 23:46 |
|
|
dekimsey1 left #mojo |
| 23:46 |
|
|
fish_ left #mojo |
| 23:46 |
|
|
gshank_ left #mojo |
| 23:46 |
|
|
robinsmidsrod left #mojo |
| 23:46 |
|
|
edenc left #mojo |
| 23:46 |
|
|
hidek left #mojo |
| 23:46 |
|
|
nuba left #mojo |
| 23:46 |
|
|
s0va left #mojo |
| 23:46 |
|
|
jzawodn left #mojo |
| 23:46 |
|
|
charsbar_ left #mojo |
| 23:46 |
|
|
mattp left #mojo |
| 23:53 |
|
|
GitHub113 joined #mojo |
| 23:53 |
|
GitHub113 |
[mojo] kraih pushed 1 new commit to master: https://github.com/kraih/mojo/[…]df2a985c9694e6955 |
| 23:53 |
|
GitHub113 |
[mojo/master] many small documentation improvements - Sebastian Riedel |
| 23:53 |
|
|
GitHub113 left #mojo |
| 23:58 |
|
|
yakudzo joined #mojo |
| 23:58 |
|
|
jwang joined #mojo |
| 23:58 |
|
|
preflex joined #mojo |
| 23:58 |
|
|
stephan48 joined #mojo |
| 23:58 |
|
|
dekimsey1 joined #mojo |
| 23:58 |
|
|
baton8 joined #mojo |
| 23:58 |
|
|
marcus joined #mojo |
| 23:58 |
|
|
fish_ joined #mojo |
| 23:58 |
|
|
memowe joined #mojo |
| 23:58 |
|
|
k-man joined #mojo |
| 23:58 |
|
|
alnewkirk|com joined #mojo |
| 23:58 |
|
|
gshank_ joined #mojo |
| 23:58 |
|
|
marty joined #mojo |
| 23:58 |
|
|
robinsmidsrod joined #mojo |
| 23:58 |
|
|
trcjr joined #mojo |
| 23:58 |
|
|
yko joined #mojo |
| 23:58 |
|
|
edenc joined #mojo |
| 23:58 |
|
|
gbacon joined #mojo |
| 23:58 |
|
|
gabb joined #mojo |
| 23:58 |
|
|
espent_ joined #mojo |
| 23:58 |
|
|
hidek joined #mojo |
| 23:58 |
|
|
nuba joined #mojo |
| 23:58 |
|
|
memememomo joined #mojo |
| 23:58 |
|
|
mattp joined #mojo |
| 23:58 |
|
|
s0va joined #mojo |
| 23:58 |
|
|
jzawodn joined #mojo |
| 23:58 |
|
|
charsbar_ joined #mojo |
| 23:58 |
|
|
purl joined #mojo |
