| Time |
S |
Nick |
Message |
| 00:02 |
|
pdurbin |
sfisque: so you're pretty happy with JPA |
| 00:03 |
|
sfisque |
my only complaint is that it is tough to use when doing fairly "exotic" joins. but otherwise, it's pretty good |
| 00:03 |
|
sfisque |
it definitely solves the "dynamic query" problem in a fairly painless api |
| 00:04 |
|
sfisque |
makes building query wizards less problematic than in the past .. String.concat() and StringBuilder.append() are not nice to use for SQL |
| 00:05 |
|
whartung |
yen, I use ELs ExpressionBuilder for that. |
| 00:05 |
|
whartung |
I haven't converted over to Critera, as we're still not on JPA 2 |
| 00:05 |
|
whartung |
even then, not sure if it's worthy the effort |
| 00:05 |
|
sfisque |
i find pre-2 jpa unusable now taht i've drank the koolaid |
| 00:06 |
|
sfisque |
it's like hand rolling servlets compared to using a web framework |
| 00:07 |
|
whartung |
we use the latest EL, but not JPA 2 |
| 00:12 |
|
pdurbin |
we started looking a bit at mongodb but... I'm pretty sure we're gonna stick with JPA and postgresql |
| 00:12 |
|
sfisque |
pdurbin didnt you tell me that your org uses neo4j |
| 00:13 |
|
pdurbin |
yeah, more as a library though... for analysis of graphml files |
| 00:13 |
|
sfisque |
gotcha |
| 00:14 |
|
pdurbin |
we were looking at mongo because we want to be flexible in the types of data we accept and process |
| 00:14 |
|
pdurbin |
i.e support a few fields such as social science, astronomy, and biomed but be ready to support other fields in the future |
| 00:15 |
|
pdurbin |
every field has its own set of metadata it cares about when in comes to publishing data |
| 00:15 |
|
sfisque |
aye |
| 00:15 |
|
pdurbin |
everyone will give their data a title of some sort. author. etc. |
| 00:16 |
|
pdurbin |
but there's a lot of variation |
| 00:16 |
|
pdurbin |
and do we want to have to cut a release every time we want to support other field? neuroscience or what have you... not really |
| 00:17 |
|
pdurbin |
so maybe a plugin architecture of some sort |
| 00:17 |
|
pdurbin |
but on JPA and postgres, I guess |
| 00:24 |
|
whartung |
well, remember this tenet -- database schema are designed around the questions to be asked of them, not the data to be stored. |
| 00:24 |
|
sfisque |
yup |
| 00:25 |
|
sfisque |
if you're just interested in purely data store/retrieve, there are simpler mechs to use. rdbms are for asking questions |
| 00:25 |
|
whartung |
yup |
| 00:25 |
|
whartung |
you can always convert it later, depending on your volume |
| 00:29 |
|
semiosis |
imo, unless your dataset is extremely large, use an RDBMS -- they're the most well understood |
| 00:30 |
|
sfisque |
i guess the point though, is if you're not asking questions and just round tripping data, there are better mechs (flat files, nosql, object stream, etc.) for doing that |
| 00:30 |
|
whartung |
or clobs/blob in the db |
| 00:31 |
|
whartung |
those are at least transactional |
| 00:31 |
|
sfisque |
i guess, but rdbms' dont have a monopoly on TXs. any decent persistence mech will have some form of acid |
| 00:32 |
|
whartung |
but then why manage two servers, two installs, two mechanisms of backup, blah blah blah blah |
| 00:33 |
|
whartung |
reach for rdbms until it proves itself unworthy, rather than you think it's unworthy :) |
| 00:33 |
|
* semiosis |
mixes mysql with glusterfs :O |
| 00:34 |
|
semiosis |
...in java |
| 00:34 |
|
semiosis |
rdbms not the best thing to serve photos & videos to the web |
| 00:34 |
|
pdurbin |
well, sure |
| 00:35 |
|
sfisque |
thank you semiosis. my point precisely. why wait for production to fail when you can avoid that by thinking through the problem appropriately |
| 00:35 |
|
whartung |
but they're not absolutely dreadful. All the major DBs will stream blobs if you ask them nicely. |
| 00:35 |
|
sfisque |
of course they will.. but will it scale accordingly |
| 00:36 |
|
* whartung |
has streamed videos from a DB |
| 00:36 |
|
semiosis |
like i said, not the best |
| 00:36 |
|
whartung |
but that doesn't mean it's unusable. |
| 00:36 |
|
whartung |
is my point |
| 00:36 |
|
pdurbin |
speaking of acid, has anyone worked with http://en.wikipedia.org/wiki/H-Store or http://en.wikipedia.org/wiki/VoltDB ? (or http://en.wikipedia.org/wiki/NuoDB I guess.) I heard about them at http://www.se-radio.net/episode-199-michael-stonebraker/ |
| 00:37 |
|
sfisque |
i mean, you could run an ecommerce site using berkelydb in the back end… if you really wanted… but would you make that decision "because it can" or woudl you consider the ramifications and use a more "appropriate" tool |
| 00:37 |
|
pdurbin |
"The significance of the H-Store is that it is the first implementation of a new class of parallel database management systems, called NewSQL, that provide the high-throughput and high-availability of NoSQL systems, but without giving up the transactional guarantees of a traditional DBMS." |
| 00:37 |
|
whartung |
alleyways give Stonebrakers stuff 10 years to settle down :) |
| 00:38 |
|
pdurbin |
whartung: that interview with him is really good |
| 00:38 |
|
semiosis |
pdurbin: sounds a little like http://en.wikipedia.org/wiki/MySQL_Cluster |
| 00:38 |
|
whartung |
the primary thing that make bdb an issue, imho, is lack of a query language. But, other than that, BDB is fine. It has locks, it has btrees, it caches pages, blah blah blah. |
| 00:39 |
|
pdurbin |
semiosis: hmm, well both seem to be "shared nothing" |
| 00:39 |
|
semiosis |
and in-memory |
| 00:39 |
|
pdurbin |
yeah |
| 00:39 |
|
semiosis |
row-oriented |
| 00:39 |
|
semiosis |
auto-sharding |
| 00:40 |
|
whartung |
I have a catalog app that stores all of its images in a DB, but I cache them to the disk for serving them to the web…I think. |
| 00:40 |
|
semiosis |
many similarities |
| 00:40 |
|
whartung |
isn't all of his stuff currently memory based? |
| 00:40 |
|
whartung |
Postgres started as Stonebrakers work. |
| 00:40 |
|
pdurbin |
whartung: good point about the questions to be asked. and some of those questions we'll ask of solr, not necessarily the database itself |
| 00:41 |
|
whartung |
the primary benefit of solr is the index features of lucene underneath it. |
| 00:41 |
|
* whartung |
has written inverted indexes in SQL before... |
| 00:41 |
|
pdurbin |
yep. sure |
| 00:42 |
|
whartung |
how much data are you storing pdurbin ? |
| 00:42 |
|
pdurbin |
I find solr much more hackable than lucene... easy to hit a web api to figure our what it can do |
| 00:42 |
|
whartung |
yea, it's easier to integrate with in a SOA sort of way |
| 00:42 |
|
* whartung |
has integrated lucene with a RDBMS also :D |
| 00:43 |
|
whartung |
it wasn't transactional, but thems the breaks. |
| 00:43 |
|
pdurbin |
whartung: in the database? buh. 5GB? I forget what prod has |
| 00:43 |
|
whartung |
more? |
| 00:43 |
|
whartung |
6GB? 600 GB? |
| 00:44 |
|
whartung |
if the database isn't bigger than the ram on a netbook, then look, I mean, really, who cares what you use lol |
| 00:44 |
|
pdurbin |
let's say between 5 and 20 GB. not sure |
| 00:44 |
|
sfisque |
because you can lock up a db with 100k TXs per minute, even if the |
| 00:44 |
|
sfisque |
the TX is only a few bytes each |
| 00:44 |
|
whartung |
what's your load pdurbin ? |
| 00:44 |
|
whartung |
how many users? |
| 00:45 |
|
pdurbin |
let's say thousands of users. not sure how to quantify load for ya |
| 00:45 |
|
whartung |
how many of them on line at any one time? TPS |
| 00:46 |
|
pdurbin |
you know... we have new relic... maybe that would tell me |
| 00:46 |
|
whartung |
1000's of idle users don't count for much :) |
| 00:46 |
|
pdurbin |
:) |
| 00:46 |
|
whartung |
the biggest traffic hit of most systems today is basically the access log. |
| 00:47 |
|
pdurbin |
is it useful to think of the shape of your data? or is the answer always "square.... put it in a relational database" ;) |
| 00:47 |
|
whartung |
relations databases data is shape as trees and graphs |
| 00:47 |
|
whartung |
…with square nodes |
| 00:49 |
|
* whartung |
ran a web site late 90's a http://en.wikipedia.org/wiki/SPARCstation_2 -- we had 10's of thousands of users registered on that site. |
| 00:49 |
|
whartung |
*on a |
| 00:49 |
|
whartung |
JSPs, servlets, and Oracle…pre-WARs |
| 00:49 |
|
whartung |
running on JRun(!!) |
| 00:52 |
|
whartung |
how much of your data is active on any day? |
| 00:53 |
|
pdurbin |
whartung: I should just give you the new relic login ;) |
| 00:53 |
|
whartung |
I don't know what that is |
| 00:53 |
|
pdurbin |
javaeebot`: lucky new relic |
| 00:53 |
|
javaeebot` |
pdurbin: http://newrelic.com/ |
| 00:53 |
|
whartung |
ah k |
| 00:54 |
|
pdurbin |
it's pretty cool. it can tell you how many joins happen for a given http request |
| 00:54 |
|
whartung |
wow -- adobe finally killed JRun only last year! |
| 00:56 |
|
pdurbin |
a terrible loss? |
| 00:57 |
|
whartung |
nah |
| 00:57 |
|
whartung |
tomcat was just getting out the gate back then |
| 00:57 |
|
whartung |
we actually started with Netscapes server, but it would randomly lock up |
| 00:57 |
|
whartung |
so we got JRun instead. It was cheap enough, and the software moved right over \o/ standards! |
| 01:07 |
|
pdurbin |
back when I ran netscape servers (mail and ldap) I didn't quite get what an "app" server was :) |
| 01:07 |
|
sfisque |
TC didnt appear until after 2000. back in 99 the only option was either the sun server or jserv plugin for apache (which was a P I T A to configure) |
| 01:11 |
|
whartung |
3.0 was in 1999, accordiing to Mr. Pedia |
| 01:13 |
|
sfisque |
i could very well have been, but i do not recall the actual day that it was made available to the outside world. it was basically sun turning over code to the ASF. eventually it came out as tomcat to the rest of the world. but before it was available, there wasnt much available for servlets other than jserv, sun server, and i think weblogic (back when it was called tenga server) |
| 01:13 |
|
whartung |
yea |
| 01:13 |
|
whartung |
well, we had netscape and jrun. |
| 01:15 |
|
whartung |
and I know we were toying with tomcat of somekind |
| 01:46 |
|
sfisque |
aye. i always forget that netscape had their own server product. |
| 01:49 |
|
pdurbin |
they were doing js on the server before it was cool: http://en.wikipedia.org/wiki/JavaScript#Server-side_JavaScript |
| 01:51 |
|
sfisque |
yah, i vaguely remember SSI javascripting |
| 02:12 |
|
|
tommmied joined ##javaee |
| 04:46 |
|
kotten |
In JSF is there anyway to set value on one of the clients bean from within another bean? |
| 04:55 |
|
sfisque |
inject the bean and call a setter |
| 04:55 |
|
sfisque |
@Inject BeanType bean; bean.setXXX( xxx ); |
| 05:05 |
|
|
cutepro joined ##javaee |
| 05:05 |
|
cutepro |
hi |
| 05:10 |
|
cutepro |
http://pastie.org/8638713#34,49,138 lines 35,50 and 144. http://i43.tinypic.com/4tld3o.png . I am getting 404 for response.sendRedirect("jsp/emailconfirmation.jsp"); now . any clue why? |
| 05:13 |
|
sfisque |
maybe change the path to "/jsp/emailconfirmation.jsp" ? |
| 05:29 |
|
kotten |
sfisque , I constructed a reference to the calling bean dont know if this will work.. |
| 05:29 |
|
kotten |
@ManagedProperty(value="#{MessageBean}") |
| 05:29 |
|
kotten |
private MessageBean messagebean; |
| 05:29 |
|
kotten |
then i call this ones function to get what i want.. I suspect it wont work as it is viewscoped |
| 05:44 |
|
sfisque |
i can't see why not. the only issue i would see is if the bean is request scoped, setting the var and watching it disappear when the page transition happens |
| 05:44 |
|
sfisque |
you "could" move the var into a session scoped bean and then inject the session scoped bean into both beans |
| 05:44 |
|
sfisque |
that way it wont be effermeral |
| 05:45 |
|
sfisque |
***ephemeral |
| 07:34 |
|
|
neuro_sys joined ##javaee |
| 08:08 |
|
|
fabioportieri joined ##javaee |
| 08:09 |
|
fabioportieri |
o/ |
| 08:16 |
|
neuro_sys |
\o |
| 08:18 |
|
fabioportieri |
\o/ <- friday super happy mode |
| 08:48 |
|
|
weyer joined ##javaee |
| 08:48 |
|
|
firebird1 joined ##javaee |
| 08:48 |
|
firebird1 |
hi |
| 08:48 |
|
firebird1 |
Exception in thread "http-bio-8080-exec-9" java.lang.OutOfMemoryError: PermGen space |
| 08:48 |
|
firebird1 |
? |
| 08:50 |
|
fabioportieri |
firebird1: it happens to me too sometimes with jsf mojarra and tomcat6 |
| 08:50 |
|
fabioportieri |
not sure the thread is the same, the error is tho |
| 08:52 |
|
|
weyer joined ##javaee |
| 08:53 |
|
firebird1 |
how to fix and what is this error ? |
| 09:10 |
|
fabioportieri |
i fix by restarting the server, if it doesn't work you may want to augment memory space in the jvm |
| 09:21 |
|
|
sajjadg joined ##javaee |
| 09:52 |
|
|
weyer joined ##javaee |
| 09:59 |
|
|
fabioportieri joined ##javaee |
| 10:06 |
|
|
weyer joined ##javaee |
| 10:09 |
|
firebird1 |
how to support japanese langiag |
| 10:49 |
|
|
SEx2 joined ##javaee |
| 11:34 |
|
firebird1 |
java is far touch then C :/ |
| 11:35 |
|
firebird1 |
magic<---ask |
| 11:35 |
|
firebird1 |
magic--->give |
| 11:36 |
|
fabioportieri |
uh? |
| 11:37 |
|
firebird1 |
? |
| 11:46 |
|
neuro_sys |
wut? |
| 12:15 |
|
cutepro |
hi |
| 12:15 |
|
cutepro |
Whats the advantage of spring core? the core only |
| 12:16 |
|
fabioportieri |
dependency managmentDI |
| 12:17 |
|
cutepro |
ok, we have one. any other? |
| 12:18 |
|
fabioportieri |
EL, context |
| 12:18 |
|
cutepro |
spring expression language? |
| 12:18 |
|
cutepro |
whats context here? |
| 12:19 |
|
fabioportieri |
yes |
| 12:19 |
|
fabioportieri |
the way spring access objects |
| 12:20 |
|
fabioportieri |
aniway i'm not spring expert |
| 12:20 |
|
firebird1 |
how to display status in primefaces datatable , status means row no |
| 12:20 |
|
fabioportieri |
firebird1: look at varStatus attr |
| 12:20 |
|
|
weyer joined ##javaee |
| 12:21 |
|
firebird1 |
fabioportieri, i looked at the doc cant find varstatus |
| 12:23 |
|
fabioportieri |
mm maybe in primefaces is not called that way |
| 12:23 |
|
fabioportieri |
firebird1: should be called "var" if i'm not mistaken |
| 12:24 |
|
firebird1 |
got it |
| 12:24 |
|
firebird1 |
its not var |
| 12:24 |
|
fabioportieri |
what was it? |
| 12:24 |
|
firebird1 |
its not like loop it has its rowindex |
| 12:25 |
|
fabioportieri |
woo primefaces exploding popup is nice |
| 12:26 |
|
fabioportieri |
rowIndexVar should do it firebird1 |
| 12:26 |
|
firebird1 |
s |
| 12:45 |
|
cutepro |
fabioportieri what is the advantage of DI in the first place |
| 12:47 |
|
fabioportieri |
it follows single responsability principle |
| 12:47 |
|
fabioportieri |
and even the open closed principle |
| 12:48 |
|
fabioportieri |
permits a cleaner implementation of oop in general |
| 12:48 |
|
cutepro |
what advantage does single responsibilty principle gives us? + this can be atained without spring |
| 12:48 |
|
fabioportieri |
yeah but you'll have to code your own stuff |
| 12:48 |
|
cutepro |
ok |
| 12:49 |
|
fabioportieri |
with spring you already have a framework |
| 12:49 |
|
fabioportieri |
google SOLID |
| 12:49 |
|
cutepro |
ok what advantage does single responsibilty principle gives us? |
| 12:49 |
|
fabioportieri |
http://en.wikipedia.org/wiki/SOLID_%28object-oriented_design%29 |
| 12:50 |
|
fabioportieri |
cutepro: you have classes that have only one responsability -> easier to maintain |
| 12:50 |
|
cutepro |
one responsibility? can you elaborate? |
| 12:51 |
|
fabioportieri |
sure |
| 12:51 |
|
fabioportieri |
http://en.wikipedia.org/wiki/Single_responsibility_principle |
| 12:51 |
|
fabioportieri |
:) |
| 12:52 |
|
cutepro |
k |
| 12:53 |
|
cutepro |
ok. can you tell one place . by code or other, by which it facilitates? |
| 12:55 |
|
fabioportieri |
sorry? |
| 12:56 |
|
|
weyer joined ##javaee |
| 12:57 |
|
cutepro |
np/ |
| 12:57 |
|
cutepro |
thanks ! |
| 12:58 |
|
fabioportieri |
DI enforces SRP cause your class will not have to act as a factory for its dependencies (other objects the class uses and instantiate) |
| 13:00 |
|
cutepro |
ok |
| 13:00 |
|
cutepro |
Parent o = new Child(); o.childMethod(); valid ? childMehtod() == a method in child class |
| 13:03 |
|
fabioportieri |
uhh.. i think you need a cast if you do that, not sure now |
| 13:03 |
|
cutepro |
ok |
| 13:03 |
|
fabioportieri |
try it |
| 13:17 |
|
firebird1 |
why we use datainputstream ? |
| 13:52 |
|
|
tommmied joined ##javaee |
| 14:24 |
|
|
weyer joined ##javaee |
| 14:35 |
|
|
fabioportieri joined ##javaee |
| 14:39 |
|
|
sheenobu joined ##javaee |
| 14:42 |
|
|
Naros joined ##javaee |
| 14:50 |
|
|
sajjadg joined ##javaee |
| 15:32 |
|
|
weyer joined ##javaee |
| 16:01 |
|
|
weyer joined ##javaee |
| 16:19 |
|
|
Naros left ##javaee |
| 16:33 |
|
|
syncsys joined ##javaee |
| 17:14 |
|
|
kobain joined ##javaee |
| 17:17 |
|
fabioportieri |
byee |
| 17:19 |
|
|
neuro_sys joined ##javaee |
| 17:41 |
|
|
sajjadg joined ##javaee |
| 17:56 |
|
|
janpenuel joined ##javaee |
| 18:26 |
|
|
sess joined ##javaee |
| 20:57 |
|
syncsys |
in jpa , I think catch (PersistenceContextType cant be catched? |
| 20:59 |
|
sfisque |
any exception can be caught |
| 21:02 |
|
|
kobain joined ##javaee |
| 21:02 |
|
syncsys |
how to catch it? IDE says: incompatible types. Required: java.lang.Throwable Found: javax.persistence.PersistenceContextTyp |
| 21:03 |
|
syncsys |
and do I need @PersistenceContext(type = PersistenceContextType.TRANSACTIONAL ON private EntityManager entityManager; |
| 21:05 |
|
sfisque |
does PersistenceContextType extend Exception? i'm thinking it does not |
| 21:06 |
|
sfisque |
if it's not some implementation of Throwable (aka exception) then you can't catch it |
| 21:07 |
|
syncsys |
sfisque, last line. http://docs.oracle.com/javaee/6/api/javax/persistence/EntityManager.html#merge%28T%29 |
| 21:08 |
|
syncsys |
http://docs.oracle.com/javaee/6/api/javax/persistence/PersistenceContextType.html |
| 21:08 |
|
sfisque |
you need to catch TransactionRequiredException |
| 21:08 |
|
sfisque |
not the type of the EM |
| 21:08 |
|
|
neuro_sys joined ##javaee |
| 21:08 |
|
|
neuro_sys joined ##javaee |
| 21:08 |
|
sfisque |
as well as IllegalArgumentException |
| 21:08 |
|
syncsys |
http://pastie.org/8643525 |
| 21:09 |
|
sfisque |
yah that third catch is garbage, delete it |
| 21:09 |
|
syncsys |
why so? |
| 21:09 |
|
syncsys |
why its mentioned in last line. http://docs.oracle.com/javaee/6/api/javax/persistence/EntityManager.html#merge%28T%29 |
| 21:11 |
|
syncsys |
sfisque, now it says avax.transaction.TransactionRequiredException' is never thrown in the corresponding try block |
| 21:12 |
|
syncsys |
http://pastie.org/8643538 |
| 21:44 |
|
syncsys |
sfisque, sorry, disconnected |
| 23:07 |
|
|
syncsys joined ##javaee |
| 23:14 |
|
|
neuro_sy1 joined ##javaee |
| 23:22 |
|
sfisque |
syncsys basically waht i said was the third catch is nonsense. you only need the first two and if you're doing java7 you can do 1 combined catch |
| 23:43 |
|
|
Naros joined ##javaee |
