| Time |
S |
Nick |
Message |
| 00:55 |
|
|
philbot joined ##javaee |
| 00:57 |
|
|
VB1 joined ##javaee |
| 00:57 |
|
|
ibaca joined ##javaee |
| 00:57 |
|
|
Naros joined ##javaee |
| 00:57 |
|
|
acuzio joined ##javaee |
| 00:57 |
|
|
semiosis joined ##javaee |
| 00:57 |
|
|
Lappro joined ##javaee |
| 00:57 |
|
|
book` joined ##javaee |
| 00:57 |
|
|
TinkerTyper joined ##javaee |
| 00:57 |
|
|
pdurbin joined ##javaee |
| 00:57 |
|
|
whartung joined ##javaee |
| 00:57 |
|
|
tjsnell joined ##javaee |
| 00:57 |
|
|
liecno_ joined ##javaee |
| 00:57 |
|
|
SoniEx2 joined ##javaee |
| 00:57 |
|
|
VB2 joined ##javaee |
| 00:57 |
|
|
Sircle_ joined ##javaee |
| 00:57 |
|
|
javaeebot joined ##javaee |
| 00:57 |
|
|
Bombe joined ##javaee |
| 00:57 |
|
|
mikee joined ##javaee |
| 00:57 |
|
|
wicketn01b joined ##javaee |
| 00:57 |
|
|
huhlig joined ##javaee |
| 00:57 |
|
|
sfisque joined ##javaee |
| 00:57 |
|
|
pdave joined ##javaee |
| 00:57 |
|
|
rektide_ joined ##javaee |
| 00:57 |
|
|
CJ_ joined ##javaee |
| 00:57 |
|
|
sess joined ##javaee |
| 00:57 |
|
|
ChanServ joined ##javaee |
| 00:57 |
|
|
Fubar^ joined ##javaee |
| 00:57 |
|
|
Topic for ##javaee is now Core Java (Java SE) AND Java Enterprise Edition (Java EE) discussion | ##javaee-offtopic for all kinds of non-tech chat| logs at http://irclog.greptilian.com/javaee/today |
| 01:01 |
|
|
TxRx127 joined ##javaee |
| 01:01 |
|
|
VB1 joined ##javaee |
| 01:01 |
|
|
ibaca joined ##javaee |
| 01:01 |
|
|
Naros joined ##javaee |
| 01:01 |
|
|
acuzio joined ##javaee |
| 01:01 |
|
|
semiosis joined ##javaee |
| 01:01 |
|
|
Lappro joined ##javaee |
| 01:01 |
|
|
book` joined ##javaee |
| 01:01 |
|
|
TinkerTyper joined ##javaee |
| 01:01 |
|
|
pdurbin joined ##javaee |
| 01:01 |
|
|
whartung joined ##javaee |
| 01:01 |
|
|
tjsnell joined ##javaee |
| 01:01 |
|
|
liecno_ joined ##javaee |
| 01:01 |
|
|
SoniEx2 joined ##javaee |
| 01:01 |
|
|
Sircle_ joined ##javaee |
| 01:01 |
|
|
javaeebot joined ##javaee |
| 01:01 |
|
|
Bombe joined ##javaee |
| 01:01 |
|
|
mikee joined ##javaee |
| 01:01 |
|
|
wicketn01b joined ##javaee |
| 01:01 |
|
|
huhlig joined ##javaee |
| 01:01 |
|
|
sfisque joined ##javaee |
| 01:01 |
|
|
pdave joined ##javaee |
| 01:01 |
|
|
rektide_ joined ##javaee |
| 01:01 |
|
|
CJ_ joined ##javaee |
| 01:01 |
|
|
sess joined ##javaee |
| 01:01 |
|
|
ChanServ joined ##javaee |
| 01:01 |
|
|
Fubar^ joined ##javaee |
| 01:09 |
|
|
sajjadg joined ##javaee |
| 01:13 |
|
|
firebird1 joined ##javaee |
| 01:51 |
|
|
sajjadg joined ##javaee |
| 03:01 |
|
|
VB2 joined ##javaee |
| 05:30 |
|
|
cem_ joined ##javaee |
| 05:30 |
|
|
ramsrib joined ##javaee |
| 05:48 |
|
|
sajjadg joined ##javaee |
| 06:05 |
|
cem_ |
integrating testng with spring and JSF |
| 06:05 |
|
cem_ |
:( |
| 06:05 |
|
cem_ |
it wont detect managed proeperty |
| 06:13 |
|
|
sajjadg joined ##javaee |
| 06:34 |
|
|
sajjadg joined ##javaee |
| 06:46 |
|
|
neuro_sys joined ##javaee |
| 07:00 |
|
|
sajjadg joined ##javaee |
| 07:23 |
|
|
sajjadg joined ##javaee |
| 07:41 |
|
|
sajjadg joined ##javaee |
| 07:59 |
|
cem_ |
sajjadg help |
| 08:00 |
|
cem_ |
sajjadg get a good proxy |
| 08:04 |
|
|
firebird1 joined ##javaee |
| 08:09 |
|
|
TxRx127 left ##javaee |
| 08:17 |
|
sajjadg |
cem_: what? |
| 09:00 |
|
|
ramsrib joined ##javaee |
| 09:01 |
|
|
cem_ joined ##javaee |
| 11:13 |
|
|
neuro_sys joined ##javaee |
| 12:25 |
|
CJ_ |
In your opinion, is it better to create a file on the local file system and then transfer it to the download stream for a user or just do everything on the fly? |
| 12:26 |
|
neuro_sys |
disk i/o will be slow, but memory is limited. choose your destiny. |
| 12:31 |
|
CJ_ |
I think I'm going to stream it. I have to bring everything into memory to create it on the file system anyways. |
| 13:35 |
|
|
ramsrib joined ##javaee |
| 14:19 |
|
|
firebird1 joined ##javaee |
| 14:31 |
|
|
Naros joined ##javaee |
| 15:40 |
|
|
sfisque joined ##javaee |
| 15:52 |
|
|
jieryn joined ##javaee |
| 15:52 |
|
|
jieryn joined ##javaee |
| 16:45 |
|
|
neuro_sys joined ##javaee |
| 16:45 |
|
|
neuro_sys joined ##javaee |
| 16:59 |
|
pdurbin |
starting soon in ##virtualjug: "WebSocket Applications using Java EE 7" by Arun Gupta - virtual Java User Group (London, England) - Meetup - http://www.meetup.com/virtualJUG/events/163666492/ |
| 16:59 |
|
|
Naros left ##javaee |
| 16:59 |
|
pdurbin |
live stream: ▶ "WebSocket Applications using Java EE 7" by Arun Gupta - YouTube - http://www.youtube.com/watch?v=lbANLOUFe58 |
| 17:05 |
|
neuro_sys |
I've been using websocket for a SOA for the past month |
| 17:05 |
|
neuro_sys |
it's great |
| 17:05 |
|
pdurbin |
cool |
| 17:05 |
|
neuro_sys |
all traffic going through websocket with fallback support |
| 17:05 |
|
neuro_sys |
ah it's live |
| 17:06 |
|
pdurbin |
yeah |
| 17:07 |
|
|
VB1 joined ##javaee |
| 17:13 |
|
whartung |
via a web page neuro_sys ? |
| 17:26 |
|
neuro_sys |
yes whartung |
| 17:29 |
|
neuro_sys |
switched to atmosphere framework for it |
| 17:40 |
|
whartung |
yea, heard good things about it. Francois is a stand up guy, does good work |
| 17:43 |
|
pdurbin |
semiosis: you're paying more attention than I am :) http://irclog.perlgeek.de/crimsonfu/2014-02-21#i_8322898 |
| 17:43 |
|
semiosis |
i started late then had to pause it :/ |
| 17:44 |
|
pdurbin |
semiosis: it'll get saved on youtube. you can watch it later |
| 17:46 |
|
neuro_sys |
http sucked mainly because it was stateful |
| 17:47 |
|
neuro_sys |
now ws is a stateful protocol, gotta love it. |
| 17:47 |
|
neuro_sys |
oops |
| 17:47 |
|
neuro_sys |
I meant stateless for http |
| 17:47 |
|
neuro_sys |
whenever you feel like you push data to all the clients, or to the ones you want. |
| 17:48 |
|
whartung |
lol |
| 17:48 |
|
whartung |
HTTP is where it is today because it's stateless |
| 17:48 |
|
neuro_sys |
that's kindda philosophical |
| 17:48 |
|
whartung |
we'd be mired in a sea of awful stateful protocols if HTTP wasn't stageful. |
| 17:48 |
|
whartung |
no, it's not. |
| 17:48 |
|
neuro_sys |
back in the day the bw was limited, so yeah |
| 17:48 |
|
whartung |
making stateless even MORE expensive, and it STILL exploded |
| 17:48 |
|
neuro_sys |
yes that is also true |
| 17:49 |
|
whartung |
whereas every stateful protocol has failed to take root, particularly in the ubiquitous way HTTP has. |
| 17:49 |
|
whartung |
RPC, CORBA, NFS, RMI, etc. etc. etc. |
| 17:49 |
|
neuro_sys |
regardless, I find http having been a stateless protocol caused a shit ton of workaround solutions to provide more interactive and responsive stateful apps that feels like desktop ones. |
| 17:50 |
|
whartung |
the only stageful protocol of any prevalence is SMTP |
| 17:50 |
|
whartung |
yea, that wouldn't have anything to do with the whole concept of HTTP not being GEARED for those applications whatsoever. |
| 17:50 |
|
whartung |
"I keep mashing this square peg and it won't. go. in. the round hole!!11!!!" |
| 17:51 |
|
neuro_sys |
hehe |
| 17:52 |
|
neuro_sys |
do you think using tomcat on production with about 1k simultanous users poses any risk? |
| 17:53 |
|
semiosis |
no |
| 17:54 |
|
tjsnell |
1k is nothing |
| 17:54 |
|
semiosis |
around 10k is where things get interesting |
| 17:58 |
|
neuro_sys |
also, there's a silly bottleneck that I have which is like, the json serialization occurs at the time of user request, while the data is already sitting there in the memory (not fetched from DB). |
| 17:58 |
|
neuro_sys |
so I should serialize them into json and keep them as such |
| 17:58 |
|
neuro_sys |
and only hand them over to the clients when they need it |
| 17:58 |
|
neuro_sys |
is that where such solutions like nosql caching dbs are useful, like couchebase or mongodb? |
| 17:59 |
|
neuro_sys |
(no experience in using them myself) |
| 17:59 |
|
whartung |
choose your database on how you want to query and manipulate your data. Whatever you think you may save in marshaling is being paid somewhere, unless you DB is the file system. |
| 18:00 |
|
tjsnell |
what json lib? |
| 18:00 |
|
neuro_sys |
using jackson |
| 18:00 |
|
tjsnell |
odd |
| 18:01 |
|
whartung |
some libs are faster than others, none of the top ones dramatically "suck worse" than the others |
| 18:01 |
|
neuro_sys |
why, haven't benchmarked them |
| 18:02 |
|
whartung |
no, others have. THe point being picking one json lib over another solely for performance is likely not productive. |
| 18:03 |
|
neuro_sys |
their use is pretty much straightforward, so wouldn't mind changing it, although I should definetely benchmark things. |
| 18:03 |
|
whartung |
marshaling is always expensive |
| 18:03 |
|
whartung |
always has been |
| 18:03 |
|
neuro_sys |
I'm still curious about if, say, the system specs of a server with 2ghz cpu and 2 gigs of ram would be neough |
| 18:04 |
|
whartung |
how much work space does each connection need? |
| 18:04 |
|
neuro_sys |
so I believe I should do marhsalling in advance wherever it's possible |
| 18:04 |
|
whartung |
yes, it's worth caching, certainly |
| 18:04 |
|
whartung |
and that's the point |
| 18:04 |
|
whartung |
I'd look at caching before I bothered worrying about the json lib |
| 18:06 |
|
semiosis |
neuro_sys: beware of premature optimization. make it work, then make it right, then make it fast. |
| 18:06 |
|
neuro_sys |
having not used any of the nosql stuff, I wonder if they're up to this kind of task |
| 18:06 |
|
neuro_sys |
like caching |
| 18:06 |
|
neuro_sys |
of json data |
| 18:06 |
|
neuro_sys |
semiosis: good call, it's the root of all evil after all :P |
| 18:06 |
|
whartung |
I thought that was money? |
| 18:06 |
|
neuro_sys |
lol yes |
| 18:08 |
|
whartung |
Consider neuro_sys, that you have a Customer record that's made of not just the customer demographics, but order history, etc. So, it actually involves quite a few tables to get "the customer record". If you have a Read Heavy process that typically deals with the Whole Customer Record (vs a small piece, like just the demographics), then there can be some value in rendering the entire record, and lazily keep it up to date, and store it |
| 18:08 |
|
whartung |
document form for ready retrieval. |
| 18:08 |
|
whartung |
you still have to pay a marshaling cost to convert the JSON, say, in to Java Classes |
| 18:09 |
|
whartung |
but you don't have to select and join to the dozen plus tables involved in the normal transaction. |
| 18:09 |
|
whartung |
you pay the prices of extra complexity, stale data, and synchronization to try and get that to work. |
| 18:09 |
|
whartung |
but it can be viable in some scenarios. |
| 18:10 |
|
whartung |
Now, where Mongo is "better" at storing that document than, say, Postgres or Oracle storing it in a CLOB -- meh, that's a different problem. |
| 18:10 |
|
whartung |
If you don't need Mongos scaling features, then no reason to introduce a new piece of kit just to store a chunk of text in a table. |
| 18:11 |
|
whartung |
the only way to actually "save" marshaling costs in to cache the binary form in memory. |
| 18:26 |
|
|
VB1 joined ##javaee |
| 18:29 |
|
pdurbin |
do all the JSON libs support JSR 353? Java API for JSON Processing (JSON-P) https://jcp.org/en/jsr/detail?id=353 |
| 18:38 |
|
|
AndroidLoverInSF joined ##javaee |
| 18:48 |
|
neuro_sys |
https://www.youtube.com/watch?v=LgF2aSpyQDs |
| 18:48 |
|
neuro_sys |
lol |
| 18:58 |
|
|
VB1 joined ##javaee |
| 19:02 |
|
neuro_sys |
umm, what's a good js minimizer? |
| 19:03 |
|
neuro_sys |
btw, whartung, noted everything you said, thanks |
| 19:03 |
|
whartung |
yw |
| 19:35 |
|
|
VB1 joined ##javaee |
| 19:58 |
|
|
drspockbr joined ##javaee |
| 21:56 |
|
|
VB1 joined ##javaee |
| 22:10 |
|
|
AndroidLoverInSF joined ##javaee |
| 22:13 |
|
|
AndroidLoverInSF joined ##javaee |
| 22:27 |
|
|
VB1 joined ##javaee |
| 22:33 |
|
|
AndroidLoverInSF joined ##javaee |
| 22:40 |
|
|
AndroidLoverInSF joined ##javaee |
| 23:36 |
|
|
Naros joined ##javaee |
