| Time |
S |
Nick |
Message |
| 00:06 |
|
|
jieryn joined ##javaee |
| 00:06 |
|
|
jieryn joined ##javaee |
| 00:09 |
|
|
SLovenberg joined ##javaee |
| 00:16 |
|
sfisque |
anyone have issues running update (plugins) in 7.4 on macosx.7 |
| 00:58 |
|
|
hackyo joined ##javaee |
| 01:52 |
|
pdurbin |
whartung: cpanm is newish and easier than older clients |
| 03:18 |
|
jieryn |
how can i force cdi to @Inject on new Foo() ? |
| 03:18 |
|
jieryn |
i need to create many Foo() depending on rows in a db, e.g. some Persons to monitor() |
| 03:19 |
|
jieryn |
so i create for (Person person : personEjb.findAll()) { new PersonMonitor(person); } |
| 03:19 |
|
jieryn |
where PersonMonitor has @EJB NotifyCenter notify; |
| 03:52 |
|
sfisque |
constructor injection? public Foo( @Inject T param, …. ); |
| 03:52 |
|
sfisque |
oh, you mean in an entity |
| 03:52 |
|
sfisque |
don't go down that route |
| 03:52 |
|
sfisque |
entities are payload objects. treat them that way |
| 03:57 |
|
|
[[thufir]] joined ##javaee |
| 05:41 |
|
sfisque |
is .sar a jboss only construct? i'm trying to find general docs on the difference between a rar and a sar. cant find any |
| 05:43 |
|
sfisque |
ah so it looks like a sar is like a cross between a rar and an osgi module |
| 05:43 |
|
sfisque |
for jboss containers |
| 06:52 |
|
|
tangatools joined ##javaee |
| 06:52 |
|
|
tangatools joined ##javaee |
| 06:59 |
|
|
AlexCzar joined ##javaee |
| 07:21 |
|
|
neuro_sys joined ##javaee |
| 07:27 |
|
|
Superman_ joined ##javaee |
| 07:37 |
|
|
AlexCzar joined ##javaee |
| 08:18 |
|
sess |
Anyone got experience with JAX-WS combined with EJBs? Not sure how to handle the annotations on interface vs implementation |
| 08:28 |
|
neuro_sys |
Does anyone know of a library to convert html special codes (e.g. ç which stands for 'ç') to their utf-8 equivalents? |
| 08:30 |
|
neuro_sys |
jakarta commons lang. |
| 08:57 |
|
|
weyer joined ##javaee |
| 09:04 |
|
neuro_sys |
I've a curious small need. How do you think best I can split a long body of string text into fixed size smaller chunks but never splitting in the middle of a word? |
| 09:19 |
|
|
weyer joined ##javaee |
| 09:35 |
|
|
weyer joined ##javaee |
| 09:42 |
|
|
weyer joined ##javaee |
| 10:29 |
|
|
[[thufir]] joined ##javaee |
| 10:35 |
|
|
weyer joined ##javaee |
| 10:37 |
|
|
weyer joined ##javaee |
| 11:02 |
|
pdurbin |
neuro_sys: `fold -s` ;) |
| 11:04 |
|
neuro_sys |
org.apache.commons.lang.WordUtils |
| 11:04 |
|
neuro_sys |
I used this one :) |
| 11:04 |
|
pdurbin |
ok :) |
| 11:08 |
|
|
weyer joined ##javaee |
| 13:10 |
|
|
drspockbr joined ##javaee |
| 13:56 |
|
|
neuro_sys joined ##javaee |
| 14:01 |
|
|
weyer joined ##javaee |
| 14:01 |
|
|
neuro_sys joined ##javaee |
| 14:21 |
|
|
Naros joined ##javaee |
| 14:26 |
|
|
weyer joined ##javaee |
| 14:29 |
|
|
Naros joined ##javaee |
| 14:30 |
|
neuro_sys |
@OneToMany and @OrderColumn(name="ID") somehow requires the referenced table to have PARENT_ID, where Parent table. Does anyone know some other way that I can do it? |
| 14:31 |
|
neuro_sys |
where Parent is the master table*. |
| 14:31 |
|
neuro_sys |
Erm, wait, forget it |
| 14:36 |
|
pdurbin |
figured it out? :) |
| 14:38 |
|
|
sajjadg joined ##javaee |
| 15:04 |
|
Superman_ |
Can we trim to a certain length in DB2? |
| 15:14 |
|
|
sheenobu joined ##javaee |
| 15:15 |
|
acuzio |
Trim what to a certain length |
| 15:16 |
|
neuro_sys |
he has quit |
| 15:17 |
|
neuro_sys |
but he could use SUBSTRING I guess |
| 15:24 |
|
acuzio |
if he is using SQL then there is TRIM |
| 15:32 |
|
|
weyer joined ##javaee |
| 15:34 |
|
pdurbin |
dunno if DB2 does SQL or its own thing |
| 15:46 |
|
|
semiosis joined ##javaee |
| 15:46 |
|
|
semiosis joined ##javaee |
| 15:48 |
|
sfisque |
db2 has sql as well as it's own native api |
| 15:52 |
|
sfisque |
in fact sql was pretty much invented at ibm. according to wikipedia it started life as SEQUEL and targeted System R. |
| 16:09 |
|
|
AlexCzar joined ##javaee |
| 16:16 |
|
pdurbin |
javaeebot: lucky system r |
| 16:16 |
|
javaeebot |
pdurbin: http://en.wikipedia.org/wiki/IBM_System_R |
| 16:29 |
|
|
sajjadg joined ##javaee |
| 16:34 |
|
|
trollolol joined ##javaee |
| 16:37 |
|
|
sross07 joined ##javaee |
| 16:48 |
|
|
neuro_sys joined ##javaee |
| 16:48 |
|
|
neuro_sys joined ##javaee |
| 17:10 |
|
neuro_sys |
man, 28 yo, and still have yet to graduate for a bachelors degree. :< |
| 17:10 |
|
neuro_sys |
I can't be arsed to write this final term paper |
| 17:10 |
|
neuro_sys |
so tired at work already |
| 17:15 |
|
|
Guest6933 joined ##javaee |
| 17:16 |
|
whartung |
I don't have a bachelors degree… :) |
| 17:17 |
|
neuro_sys |
I guess I'd be better off now if I didn't enter the uni in the first place. |
| 17:17 |
|
whartung |
Mind, I'm not suggesting my path is a good, best, or better path… times are a lot different. |
| 17:17 |
|
neuro_sys |
It's a literature major, bah |
| 17:18 |
|
whartung |
ah |
| 17:18 |
|
whartung |
well |
| 17:20 |
|
acuzio |
db2 is actually quite a good DB - large . very large retail stores as in where millions of rows are added every hour usually use DB2 |
| 17:20 |
|
acuzio |
on AS400 |
| 17:20 |
|
neuro_sys |
yeah, that's the case where I'm working for all our customers |
| 17:21 |
|
neuro_sys |
still do a lot of as400 programming |
| 17:21 |
|
whartung |
so tell me about the "object nature" of the as 400? I've always wanted to talk to a AS /400 coder |
| 17:22 |
|
whartung |
the AS/400 has some pretty neat stuff as I understand it |
| 17:22 |
|
acuzio |
Mainframes is where its at |
| 17:23 |
|
neuro_sys |
I don't do as400 programming, but at the least I guess DB2 lives in kernel level or something similar. |
| 17:24 |
|
acuzio |
I havent done AS400 programming i was very briefly involved ina project where they had AS/400 + DB2 running and the volumes were just off the charts . Some intrepid idiot suggested Oh convert the whole thing to EJB and have CMP ., (this was 2003) - i walked out the next day |
| 17:24 |
|
neuro_sys |
Where schemas are called libraries, and tables are files. |
| 17:24 |
|
whartung |
" volumes were just off the charts" ? |
| 17:25 |
|
acuzio |
The average transaction in an hour neared a million ; on a "normal" day - i.e. no Xmas, Thanskgiving etc day |
| 17:25 |
|
whartung |
!! |
| 17:25 |
|
whartung |
wow |
| 17:25 |
|
whartung |
nice |
| 17:25 |
|
acuzio |
This was a daily transactional DB |
| 17:25 |
|
whartung |
yea |
| 17:26 |
|
whartung |
what was the industry? |
| 17:26 |
|
acuzio |
Retail |
| 17:26 |
|
acuzio |
One of the largest Retail stores in UK |
| 17:27 |
|
whartung |
very nice |
| 17:27 |
|
whartung |
yup -- in the trenches, transactional DB work. |
| 17:28 |
|
acuzio |
They had a project to consolidate this at a "zonal" level and this was running on like i said - DB2 + AS/400 - they wanted to move to something else and one of the largest DB firms was contracted to for an initial case study (I was an intern) - even i knew that Java + <DB> was not going to work |
| 17:29 |
|
acuzio |
IN those days Java was not on the AS/400 , Mainframe platforms |
| 17:32 |
|
acuzio |
At that volume , throughput , the simplest solutions were the only ones that mattered - Tables were essentially Files ., CSV's with all kinds of special characters to short-cut processing., everything was single-threaded with explicit locking - start;read;write;commit/rollback. I thought this was really rudimentary , years later i understood the beauty of it |
| 17:33 |
|
whartung |
CSVs? Really? Were they fixed length? and append only? |
| 17:34 |
|
acuzio |
Thats it - |
| 17:34 |
|
acuzio |
And for a young whippersnapper it all sounded horribly outdated , now i sort of know better |
| 17:36 |
|
acuzio |
No updates, deletes, joins. |
| 17:37 |
|
acuzio |
BAtch Processes would "lock" part of the table (no row locks) - when doing calculations ., |
| 17:38 |
|
acuzio |
The CSV files themselves had special characters that would allow for skipping of column processing |
| 17:38 |
|
acuzio |
Right - laters |
| 17:52 |
|
|
sfisque joined ##javaee |
| 17:57 |
|
Naros |
I have a method I where I want to handle OptimisticLockException and retry the operation. If I have already got the entity via em.find(id,lockmodetype), should I simply call refresh on the entity and then reapply the operation? |
| 17:58 |
|
Naros |
I dont think refresh vs evict & find are different |
| 17:58 |
|
sfisque |
be careful with refresh. if the entity is dirty, you'll whack any changes. |
| 17:59 |
|
sfisque |
find will dig from the L1 and L2 caches before going to the db, so be careful with find(pk) |
| 18:00 |
|
Naros |
Hm, maybe this case im concerned about isn't an issue. |
| 18:01 |
|
Naros |
we use pessimistic_write operation on the row so it should be locked, so in this case, when users download the attachment, the download counter should (in theory) be handled atomicly. |
| 18:01 |
|
Naros |
and the version updated accordingly |
| 18:01 |
|
Naros |
just if users are changing other fields of the row, they'll get mid-air collisions |
| 18:01 |
|
Naros |
when they merge |
| 18:02 |
|
Naros |
thats the part i dont like. |
| 18:02 |
|
sfisque |
if you need "subtle" control over the Tx boundary, i recommend BMT. |
| 18:02 |
|
Naros |
I'm sorry, BMT? |
| 18:02 |
|
sfisque |
bean managed tx |
| 18:02 |
|
|
weyer joined ##javaee |
| 18:02 |
|
sfisque |
rather than the default CMT where the container implicitly controls the boundaries |
| 18:03 |
|
sfisque |
that way you can get flush, commit, and other explicit calls to be useful rather than no-oped |
| 18:03 |
|
Naros |
Oh I have access to flush/commit, etc already |
| 18:04 |
|
Naros |
When a user opens a record, the database row's bean will be stored in a conversation context. when they hit update, the context will be persisted back to the datastore. If the versions mismatch, i'll naturally get an optimistic lock. |
| 18:04 |
|
Naros |
the issue is tho, when users download an attachment, the download counter increments this version too |
| 18:06 |
|
Naros |
pondering is there anyway to have this download counter but not have it affect the version value. |
| 18:06 |
|
Naros |
guess it would be possible having it in another joined entity. |
| 18:07 |
|
Naros |
but that feels just ugh |
| 18:07 |
|
semiosis |
a hits table |
| 18:07 |
|
Naros |
pretty much |
| 18:08 |
|
semiosis |
avoid update wherever possible, imho |
| 18:08 |
|
sfisque |
yeah you'd have to externalize it or have a custom EM that does not roll the version on that column update |
| 18:08 |
|
Naros |
Wonder if that is even possible with any hibernate trickery. |
| 18:09 |
|
Naros |
doubtful |
| 18:14 |
|
pdurbin |
"Java is one of the languages I know fairly well and I thought that implementing an NIO.2 filesystem provider would be a fun challenge" --semiosis at http://www.gluster.org/2013/12/how-picture-marketing-is-using-and-extending-glusterfs/ |
| 18:16 |
|
semiosis |
Is it ever! |
| 18:18 |
|
pdurbin |
:) |
| 19:15 |
|
|
weyer joined ##javaee |
| 19:37 |
|
|
javier joined ##javaee |
| 19:40 |
|
|
weyer joined ##javaee |
| 19:45 |
|
Naros |
sfisque: using @OptimisticLock(excluded=true) on the property prevents it from participating in the version tracking |
| 19:45 |
|
Naros |
so I can set that on the download counter and there wont be any contention. |
| 19:50 |
|
sfisque |
nifty. i did not know that |
| 19:56 |
|
Naros |
yah neither did i. decided to do some snooping around and found it. |
| 20:05 |
|
sfisque |
\o/ nice find! |
| 21:32 |
|
|
1JTABH5NG joined ##javaee |
| 21:40 |
|
|
pustyn joined ##javaee |
| 22:09 |
|
sfisque |
so i have a conceptual question. i'm working on a JCA adapter to consume/publish to IRC/JMS (aka remote IRC gets published to local JMS and local JMS pushes get pushed out to a configured IRC channel). should i really care about TXs or just punt on them? thoughts? |
| 22:10 |
|
pdurbin |
transactions? |
| 22:11 |
|
sfisque |
when you design a JCA you can support "no Tx", "local Tx", XA Tx, or some set of those three |
| 22:48 |
|
pdurbin |
ok |
| 22:49 |
|
sfisque |
i'm thinking because it should be operating outside of any "user situation" that no Tx should be fine |
| 22:49 |
|
sfisque |
user's would interact with the JMS queue so it should be purely asynch |
| 22:51 |
|
pdurbin |
sfisque: write your facebook killer quick in ruby first, endure some fail whales, and if it takes off... re-write on top of the JVM with JCA/JMS :) |
| 22:56 |
|
sfisque |
my ruby skill is "neglible". i'd be stumbling over learning more of the language whereas i can spend my time working out the subtle details and briskly work through the main stuff |
| 22:57 |
|
sfisque |
the UX is already starting to solidify. i'm working out the external stuff, auth/auth, and credential management |
| 22:58 |
|
sfisque |
as a first cut, i'm going to persist stuff in rdbms, but eventually leverage nosql and ldap |
| 22:59 |
|
sfisque |
i figure do hierarchy in ldap, blobs in nosql, and referential stuff in sql |
| 23:00 |
|
pdurbin |
sfisque: neo4j as your nosql? |
| 23:00 |
|
sfisque |
possibly |
| 23:00 |
|
sfisque |
still evaluating options |
| 23:01 |
|
sfisque |
neo4j looks nice. not sure how well it will play with JBoss datasources |
| 23:02 |
|
sfisque |
jboss has their own product (jboss data grid) but i'm not sure of its licensing model |
| 23:13 |
|
sfisque |
on a side note, 2 phone screen interviews this week. intel and adp |
| 23:13 |
|
sfisque |
intell == arch, adp == sr dev |
| 23:17 |
|
|
javier joined ##javaee |
| 23:25 |
|
pdurbin |
I'm just thinking a social network is a graph |
| 23:27 |
|
sfisque |
of course it is :-D |
| 23:27 |
|
sfisque |
but a person and their "things" is hiearchical (hence ldap) |
| 23:27 |
|
sfisque |
but their connection to outside things is an obvious graph |
| 23:28 |
|
sfisque |
i'm just more comfy with rdbms and ldap atm, so nosql will be a migration strategy |
| 23:29 |
|
sfisque |
first cut will just be single user (me). second iteration will add multi-tenancy. third iteration will integrate with IRC/XMPP/RSS and maybe do FB integration, though that might be discarded because FB has removed their S2S communication protocol |
| 23:45 |
|
|
trollolol joined ##javaee |
