(09:00:37 AM) jsdelfino: hi all
(09:00:44 AM) ant_: hi
(09:00:52 AM) rfeng: hi
(09:00:54 AM) rfeng: hi, ant
(09:01:07 AM) rfeng: a question about the jar layout for tomcat
(09:01:23 AM) haleh [EMAIL PROTECTED]
entered the room.
(09:01:49 AM) jsdelfino: rfeng, go ahead, we'll wait a few mins for the
sdo and das guys
(09:01:51 AM) kgoodson [EMAIL PROTECTED] entered
the room.
(09:02:04 AM) rfeng: you require some changes? shouldn't we fix the
problem at the tomcat acceptance testing?
(09:02:35 AM) rfeng: the distribution follows the acceptance testing
(09:02:45 AM) ant_: is that a q for me? relating to the comments i added
to the distr jira?
(09:02:49 AM) rfeng: yes
(09:03:08 AM) ant_: i forget what i said, whats the jira no.?
(09:03:26 AM) rfeng: 282:
(09:03:27 AM) rfeng: A couple of changes to enable exposing
JavaScript/E4X components with web service entryPoints:
(09:03:27 AM) rfeng: - the container-rhino jar needs to be included in
the tomcat server/lib directory and the js and xbean jars need to be
included in the tomcat common/lib.
(09:03:27 AM) rfeng: - the StAX jars need to be moved from Tomcat
server/lib to common/lib, thats stax-api and wstx-asl.
(09:04:20 AM) ant_: ok, right yes i guess so. i haven't looked at how
the dist. stuff works, if its just using testing/tomcat then yes
(09:04:23 AM) jsdelfino: why do we need to move these jars to common/lib?
(09:04:28 AM) kevin [EMAIL PROTECTED] entered
the room.
(09:04:32 AM) jsdelfino: hi kevin
(09:04:45 AM) ant_: an application can't use stax with them in server/lib
(09:05:11 AM) jsdelfino: simple answer :) ok makes sense
(09:05:14 AM) rfeng: do we have such usages in the sample applications?
(09:05:29 AM) rfeng: BB packages its own version in the web app
(09:05:30 AM) cr22rc: why wouldn't an application put it in its webinf/lib ?
(09:05:45 AM) cr22rc: not anymore it was removed
(09:05:50 AM) rfeng: :-)
(09:05:55 AM) cr22rc: but don't understnad the reasoning
(09:05:58 AM) ant_: i found this trying to use JavaScript/E4X in a webapp
(09:06:15 AM) ant_: we don't have a e4x web app sample...yet
(09:06:44 AM) rfeng: so your sample has direct api call to Stax?
(09:06:49 AM) cr22rc: I needed it for bb and had it in webinf/lib .. saw
no obvious problem them
(09:07:08 AM) cr22rc: yes when you activiate the live stockquote
(09:07:15 AM) ant_: no, but javascript uses it internally, but it uses
the context classlaoder
(09:07:28 AM) rfeng: that's the problem I saw before
(09:07:48 AM) jsdelfino: just after this discussion - hoping we resolve
it quick :) I'd like to go over the remaining hot JIRAs for the release
(09:07:50 AM) rfeng: I personally like to have it in common/lib to avoid
classloading issues
(09:08:01 AM) jsdelfino: and some coordination to build the distribution
with all the pieces we need in it
(09:08:09 AM) rfeng: sure
(09:09:03 AM) jsdelfino: so what are u guys deciding for this stax lib,
do we have a jscript web app sample?
(09:09:40 AM) ant_: can we decide this later?
(09:10:43 AM) cr22rc: I'm not opposed... just worried about putting too
much in common/lib and exposing jars "tuscany" needs to all apps
(09:11:22 AM) jsdelfino: yes, let's decide later, can u create a JIRA
issue for this so we don
(09:11:35 AM) jsdelfino: 't forget about it and know when/if we get this on?
(09:11:36 AM) jsdelfino: thx
(09:12:10 AM) ant_: ok
(09:12:15 AM) jsdelfino: okay, I categorized the JIRA issues yesterday,
M1, M1-tentative, Mx
(09:12:36 AM) jsdelfino: thought we could first see if there's any
disagreements with how I categorized them
(09:12:51 AM) jsdelfino: anybody wants to move some issues up/down?
(09:13:09 AM) ant_: i already changed one back from M1-tentative to M1
(09:13:16 AM) jsdelfino: ok, which one?
(09:13:25 AM) ant_: e4x, ...
(09:13:45 AM) ant_: 82
(09:14:30 AM) kevin: looks fine from my point of view
(09:15:11 AM) jsdelfino: ok, 82 still shows as M1-tentative for me
because the affect version is still "M1-tentative" :)
(09:15:22 AM) isilval [EMAIL PROTECTED] entered
the room.
(09:15:35 AM) simonnash [EMAIL PROTECTED] entered the room.
(09:15:48 AM) jsdelfino: when do you think you'll get this EAX stuff
done? is it ready yet?
(09:15:57 AM) simonnash: hello
(09:16:03 AM) jsdelfino: hi simon
(09:16:19 AM) ant_: mostly yes, but with SVN down ...
(09:16:41 AM) jsdelfino: ok, is this going to run in tomcat? or j2se?
(09:17:02 AM) ant_: both if we move the stax jar, otherwise only j2se
(09:17:08 AM) rfeng: what's the impact on other things, such as
acceptance, distribution, doc?
(09:17:38 AM) jsdelfino: I'm asking because I'm worried about moving
stax jars around at this point, this can have big class loader impacts...
(09:18:16 AM) jsdelfino: ant, would you be ok with either doing only
j2se at this point? or packaging the necessary stax jars in the app for
this speicfic sample?
(09:18:16 AM) ant_: i understand
(09:18:31 AM) jsdelfino: I let you decide what you prefer
(09:18:51 AM) jsdelfino: between these 2 options :)
(09:19:01 AM) ant_: i prefer to move them to common/lib, all our
existing tests samples work like that
(09:19:28 AM) jsdelfino: do we have stax in common/lib in the acceptance
test right now?
(09:19:35 AM) ant_: no
(09:19:44 AM) rfeng: not yet, I tried once before, seems to be fine
(09:19:58 AM) jsdelfino: *seems*?
(09:20:12 AM) rfeng: worked well in my test :-)
(09:20:18 AM) cr22rc: can we open a jira that explains why this is
needed and decide it there?
(09:20:46 AM) jsdelfino: yes
(09:21:21 AM) jsdelfino: is it reasonable to say E4X is M1, E4X in
tomcat is M1-tentative if we find a way to not impact the packaging of
the JARs
(09:21:38 AM) jsdelfino: and just package stax in web-inf/lib for this app
(09:22:08 AM) ant_: the main useful part of JavaScript/E4X is for WS
support of we can't do that having JavaScript at all is a very low priority
(09:22:31 AM) Mike_Edwards [EMAIL PROTECTED] entered
the room.
(09:22:42 AM) adisakala [EMAIL PROTECTED] entered the room.
(09:22:52 AM) adisakala left the room.
(09:23:34 AM) simonnash: ant, why would we ont have WS support?
(09:23:38 AM) simonnash: ..not..
(09:23:46 AM) dkulp: However, it should work for E4X webservices with
celtix, right? That's all J2SE
(09:24:26 AM) jsdelfino: I agree that we need E4X, I'm just concerned
about moving jars around under the tomcat lib at this point, can we
create a JIRA issue for the packaging of the stax lib, and decide there?
(09:24:29 AM) ant_: can we do what cr22rc suggestes and discuss this in
a JIRA/later discussion
(09:24:34 AM) jsdelfino: yes :)
(09:24:51 AM) cr22rc: +1
(09:25:21 AM) jsdelfino: ant_, will u create the JIRA?
(09:25:26 AM) ant_: yup
(09:25:39 AM) jsdelfino: thx, let's take a look at the remaining
critical ones
(09:25:46 AM) rfeng: can we move 61 (?wsdl) to M1? I just fixed it and
the patch is in JIRA
(09:26:06 AM) jsdelfino: 343 is getting fixed by Frank, he's got the
fix, will check it in when SVN is back up
(09:26:50 AM) rfeng: svn is up now, right?
(09:27:08 AM) jsdelfino: now yes, he'll do it later this morning
(09:27:15 AM) jsdelfino: ok let's discuss 61
(09:27:50 AM) ant_: its still down for me, is it sposed to be up?
(09:28:01 AM) cr22rc: down for me
(09:28:25 AM) rfeng: I tried svn update and it worked 1 hour ago
(09:28:35 AM) kevin: down for me now
(09:28:38 AM) ant_: readonly works
(09:28:39 AM) jsdelfino: down too
(09:28:45 AM) rfeng: :-(
(09:28:58 AM) cr22rc: mabe more we need to work via patches :-)
(09:28:58 AM) jsdelfino: you can update
(09:29:01 AM) haleh: who do we need to follow up with for SVN?
(09:29:10 AM) jsdelfino: but not commit anything
(09:29:23 AM) jsdelfino: watch the infrastructure mailing list
(09:29:24 AM) gwinn [EMAIL PROTECTED] entered the room.
(09:29:31 AM) jsdelfino: this is an apache infra issue
(09:29:56 AM) jsdelfino: the read access is back up, committers got an
email on that this morning
(09:30:05 AM) jsdelfino: next: 61
(09:30:36 AM) rfeng: it's a simple one line change, ant, are you ok with it?
(09:30:36 AM) ant_: move it to M1 its harmless
(09:30:57 AM) ant_: several moe lines to put the testcase back in ;)
(09:31:11 AM) jsdelfino: ok, if you're comfortable with it, we move it to M1
(09:31:20 AM) jsdelfino: what about fixing the endpoint address?
(09:31:35 AM) ant_: will have to be another day
(09:31:56 AM) ant_: Mx
(09:32:03 AM) jsdelfino: ok, I'm good with that, move to M1, document
the incorrect endpoint address
(09:32:10 AM) jsdelfino: and, will u apply the patch?
(09:32:14 AM) ant_: yes
(09:32:24 AM) jsdelfino: when?
(09:32:25 AM) jsdelfino: :)
(09:32:41 AM) ant_: its already on my local code so whenever SVN is back
(09:32:41 AM) jsdelfino: just need to get a feel of when can close it
(09:32:44 AM) jsdelfino: ok
(09:32:48 AM) jsdelfino: next:
(09:32:59 AM) rfeng: the incorrect endpoint has already been reported by
an Axis2 jira
(09:33:25 AM) jsdelfino: great, can u add that JIRA number to ours, link
it? thanks
(09:33:33 AM) kevin: i would like to discuss 327
(09:33:40 AM) jsdelfino: next is 326 (I'm going in prio order)
(09:33:44 AM) kevin: ok
(09:33:48 AM) jsdelfino: then we can do 327 for kevin :)
(09:34:00 AM) rfeng: ok
(09:34:24 AM) jsdelfino: 326, I'm looking at it, if I find an easy
non-disruptive fix, I'll try to fix it, otherwise move to Mx
(09:34:37 AM) jsdelfino: it's about some cases of wiring between fragments
(09:34:56 AM) jsdelfino: on one hand it looks like fragments just do not
work
(09:35:03 AM) jsdelfino: cause u can't wire between them
(09:35:35 AM) jsdelfino: on the other jim had a good point in the JIRA
issue that it still made sense to have islands of isloated components in
a module
(09:35:48 AM) jsdelfino: so... update on this one later today
(09:36:06 AM) jsdelfino: next: 327
(09:36:24 AM) kevin: had a regression and fixed that
(09:36:29 AM) kevin: but ...
(09:36:36 AM) jsdelfino: kevin, so you want this fixed in M1 right?
(09:36:42 AM) kevin: that sample has now moved to das/samples
(09:37:00 AM) kevin: so i would like to provide the automation there
rather than with bigbank
(09:37:08 AM) kevin: it is not currently automated
(09:37:16 AM) jsdelfino: what do u mean by automated?
(09:37:16 AM) kevin: the readme describes set of manual stepse
(09:37:28 AM) jsdelfino: ah
(09:37:43 AM) kevin: so, while i fixed the npe ... you cannot today run
the testing build and then point the browser
(09:37:44 AM) gwinn left the room ("Leaving").
(09:37:49 AM) jsdelfino: in our distribution, we want to pre-configure
tomcat with the relevant samples
(09:38:18 AM) jsdelfino: so it's up to you, either get with rfeng today
to get it installed as part of the distro build
(09:38:25 AM) kevin: this is a stand-alone das sample and so is probably
not relevant
(09:38:33 AM) jsdelfino: best is to provide the pom or build.xml to him :)
(09:38:37 AM) jsdelfino: or u just doc the steps
(09:38:44 AM) kevin: but, i would like to automate this in a similar way
but under das samples
(09:38:51 AM) kevin: yes
(09:38:56 AM) jsdelfino: is it running in tomcat? or j2se?
(09:39:02 AM) kevin: i will automate as much as i can for m1
(09:39:08 AM) kevin: tomcat
(09:39:36 AM) kevin: with goal of complete automation with a build.xml
(09:40:13 AM) jsdelfino: raymond, can u tell kevin what u need, and then
u guys decide if you're comfortable with putting the automation in M1
(09:40:41 AM) rfeng: sure, he can directly look into the tomcat
acceptance build.xml
(09:40:53 AM) kevin: if today is the cut-off then we will have to go
with documented steps
(09:41:03 AM) jsdelfino: I was thinking about requiring a stable distro
build by tomorrow morning
(09:41:22 AM) ant_: your time? so UK has most of tomorrow?
(09:41:23 AM) kevin: ok, then i wiill see how far i can get
(09:41:32 AM) jsdelfino: cut-off for the distro build is a bigger
question, how about tomorrow morning PDT?
(09:41:39 AM) jsdelfino: can we get a quick poll from the room?
(09:42:00 AM) ant_: seems harsh to make cut off today while SVN is still
down
(09:42:04 AM) jsdelfino: by cut-off I mean no jars moving around, no new
directories, no structure changes
(09:42:13 AM) kevin: ahh
(09:42:15 AM) jsdelfino: yeah :)
(09:42:23 AM) cr22rc: hard for that without svn
(09:42:44 AM) cr22rc: need to do a lot to get it in sync with what we
discussed yesterday
(09:42:45 AM) rfeng: cr22rc, can you commit the distribution patch as
the base when svn is up?
(09:43:03 AM) cr22rc: sure
(09:43:04 AM) dkulp: If he cannot, I can.
(09:43:13 AM) rfeng: so people can try out to see if it fits
(09:43:41 AM) jsdelfino: ok, good, can u guys please update the status
of the distro JIRA when u do that so we know where its stands
(09:44:07 AM) rfeng: it's 282
(09:44:18 AM) jsdelfino: looks like it's difficult to set a cut-off time
for the distro build without SVN :) so let me ask a diff. question then:
(09:44:47 AM) jsdelfino: can u all tell us now if you need any
significant change, addition to the distro build?
(09:44:51 AM) jsdelfino: so far I have:
(09:44:55 AM) jsdelfino: kevin's change
(09:45:02 AM) jsdelfino: ant's e4x sample
(09:45:05 AM) slaws left the room (quit: "Bye").
(09:45:13 AM) jsdelfino: dan, do u have anything?
(09:45:26 AM) dkulp: Yea, working on some samples for celtix stuff.
(09:45:32 AM) dkulp: Mostly a XML over JMS thing.
(09:45:44 AM) jsdelfino: ok, one more sample
(09:45:45 AM) dkulp: But that will just be an addition to the
samples/sca dir.
(09:45:47 AM) ant_: i dont really have an e4x sample yet, i have the
code changes to make it work from TUSCANY-82
(09:46:26 AM) dkulp: I also have a lot of bug fixes for the celtix
binding, but those don't affect the distribution at all.
(09:46:50 AM) jsdelfino: can we have JIRA issues for all these additions
or changes, and add a link from 282? so that the guys working on 282
know what to expect and pick up?
(09:47:22 AM) ant_: the ISL guys have the java2wsdl plugin patch that
makes a few dir changes
(09:47:28 AM) ant_: is that in?
(09:47:45 AM) cr22rc: besides bb in the "current" java/samples there is
"JavaScript" ... where does this go ?
(09:47:47 AM) jsdelfino: don't want to be a pain with JIRA, this will
help communication and coordination I think :)
(09:48:18 AM) jsdelfino: the java2wsdl update is assigned to me, I had a
chat with venkata earlier, will apply the change later today
(09:48:23 AM) jsdelfino: when SVN is back up :)
(09:49:42 AM) jsdelfino: ok, so if JIRA gets back up in 2 or 3 hours,
are u guys comfortable with a cut-off for significant changes to the
distro build tomorrow morning 10am PDT?
(09:50:12 AM) jsdelfino: just need to set up a goal, so we know when to
sync up all the pieces...
(09:50:25 AM) ant_: +1
(09:50:27 AM) jsdelfino: tell me if it doesn't make sense :)
(09:50:47 AM) jsdelfino: dan, rfeng, kevin, will that work for you?
(09:50:50 AM) kevin: yes
(09:50:52 AM) dkulp: Sure
(09:51:11 AM) jsdelfino: cr22rc do u have any big upcoming changes? did
u get a chance to refactor the samples dir before SVN went down?
(09:51:43 AM) cr22rc: no... did not have a "clean" version ... worried a
little about what i have
(09:51:43 AM) rfeng: yes
(09:52:32 AM) jsdelfino: ok so there's a risk with the samples, I guess
we can't really say more about this without knowing when SVN gets back up
(09:52:47 AM) jsdelfino: if there's nothing else on this subject, moving
on to 333
(09:52:48 AM) cr22rc: right
(09:52:57 AM) jsdelfino: this one is worrying me...
(09:53:15 AM) jsdelfino: random ClassCastExceptions in bigbank
(09:53:29 AM) cr22rc: I spent time on it ... still haven't narrowed it.
(09:53:48 AM) jsdelfino: a serious problem, you get this exception very
quickly
(09:54:01 AM) jsdelfino: first step I'd say is to gather all the info
(09:54:07 AM) jsdelfino: then get multiple eyes on it
(09:54:24 AM) cr22rc: I've automated bb ... it can't go through that
without encountering this
(09:54:31 AM) jsdelfino: a famous open source guy said "before 1000 eyes
all bugs are shallow" :)
(09:54:37 AM) jsdelfino: or something like that
(09:54:49 AM) jsdelfino: can we get the info in front of everybody
(09:54:53 AM) ant_: who else has seen this besides cr22rc and me?
(09:55:03 AM) kevin: i have
(09:55:04 AM) jsdelfino: the stack trace
(09:55:07 AM) jsdelfino: me too
(09:55:15 AM) jsdelfino: do we have stack trace?
(09:55:24 AM) cr22rc: in the jira
(09:55:38 AM) kevin: if, i remember, the stack trace varied
(09:55:58 AM) kevin: but always CCE
(09:56:04 AM) jsdelfino: http://issues.apache.org/jira/browse/TUSCANY-333
(09:56:08 AM) ant_: i think i always get the same one, but its not the
same as the one in the JIRA
(09:56:33 AM) ant_: i'll add mine to the jira
(09:56:39 AM) jsdelfino: rick, could you do the following?
(09:57:02 AM) jsdelfino: in the component that returns the SDO
(09:57:16 AM) jsdelfino: print theSDO.getClass().getClassLoader()
(09:57:36 AM) cr22rc: I "think" but am not sure it because BB uses the
SAME classes in the account and webclient webapps
(09:57:37 AM) jsdelfino: in the client, print
CustomerProfile.class.getClassLoader()
(09:58:00 AM) jsdelfino: and post that to the JIRA so we can take a look?
(09:58:25 AM) cr22rc: if you run this on two machines or I think two
seperate jvms it works solid
(09:58:28 AM) cr22rc: sure
(09:58:52 AM) jsdelfino: when you print the classloaders, try to print
as much info on them, like their parent classloader as well
(09:58:57 AM) jsdelfino: so we know what they are
(09:59:19 AM) cr22rc: I did that alread.. but don't lost results ... I
saw nothing obvious
(09:59:32 AM) cr22rc: the classes were loadded with the webapp loaader
(09:59:36 AM) jsdelfino: where the classloaders the same?
(09:59:42 AM) jsdelfino: both?
(09:59:45 AM) cr22rc: yup
(10:00:19 AM) cr22rc: I put the code in a finally statement .. I saw no
diff between when it worked and didnt
(10:00:23 AM) jsdelfino: if the classes had the same name, loaded by the
same classloader, we wouldn't get a CC exception
(10:00:35 AM) cr22rc: I'll do it again
(10:00:58 AM) jsdelfino: post the code to the JIRA as well, so we can
see both the output, and what produces it, will help understand
(10:01:00 AM) rfeng: can you set a "exception" breakpoint against CCE to
capture it?
(10:01:02 AM) jsdelfino: thanks
(10:01:24 AM) cr22rc: yes .. the expection in the proxy class
(10:01:29 AM) rfeng: then you can inspect the classloaders
(10:01:29 AM) jsdelfino: so the bottom line is, all of us are running
into this one, we can't have an M1 without this fixed IMO
(10:01:55 AM) cr22rc: spinned several hours on it yesterday
(10:02:03 AM) rfeng: I can also help if you let me know how to reproduce it
(10:02:14 AM) cr22rc: ah ... run bb
(10:02:20 AM) rfeng: ok
(10:02:23 AM) jsdelfino: debugger as suggested by rfeng, but then if you
want others to be able to help as well, put all the info you have from
the debugging session in the JIRA
(10:02:36 AM) ant_: the way i reproduce it is tsop/start TC, then run
BB. maybe 1 in 10 times it fails
(10:02:37 AM) jsdelfino: ok, so rick you're working on 333?
(10:03:03 AM) cr22rc: that and the big sample move
(10:03:18 AM) jsdelfino: okay
(10:03:44 AM) jsdelfino: let us know on IRC or dev list if u need help
with either one
(10:03:47 AM) cr22rc: no one answered what that javascript was in
samples today
(10:03:56 AM) jsdelfino: depending on how things go today
(10:03:58 AM) cr22rc: need svn :-)
(10:04:15 AM) ant_: they're the 'old' javascript sample sthat no one
seemed to like
(10:04:23 AM) jsdelfino: what that javascript was in samples? what do u
mean?
(10:04:31 AM) jsdelfino: oh yeah
(10:04:42 AM) simonnash: need to leave now. i will check back later to
see if you guys are still chatting.
(10:04:57 AM) jsdelfino: ant_, it's not that we didn't like them :)
(10:05:00 AM) cr22rc: afaik there should only be bb in there.. right?
(10:05:10 AM) jsdelfino: yes only bb should be there
(10:05:24 AM) jsdelfino: and ant_ will take one or two good js samples
to sca/samples
(10:05:32 AM) jsdelfino: I already moved the jsonrpc one there
(10:06:00 AM) jsdelfino: ant_, would you be ok to delete the whole
JavaScript dir for now?
(10:06:13 AM) jsdelfino: or move it to somewhere else?
(10:06:27 AM) jsdelfino: and just have the one or two js samples under
sca/samples?
(10:07:22 AM) jsdelfino: or move the whole javascript dir under sca/samples
(10:07:29 AM) ant_: yes its fine to move it. will have to wait and see
if i've the time to make new ones
(10:07:34 AM) jsdelfino: ok good
(10:07:55 AM) cr22rc: goes in samples with sca sdo etc
(10:07:59 AM) ant_: or delete it. whatever
(10:08:14 AM) jsdelfino: if u don't have the time to make new ones, I
suggest you just rename one of the 6 or 7 to a more obvious name
(10:08:19 AM) jsdelfino: like helloworldjs
(10:08:26 AM) jsdelfino: and have that under sca/samples
(10:08:30 AM) ant_: yep
(10:08:54 AM) jsdelfino: ok, are u on track to do that before the distro
build cut-off? again when SVN gets back up...
(10:09:40 AM) jsdelfino: are u happy with that? again if you want to
keep two of the samples instead of one it's OK with me, just give them
obvious names
(10:10:16 AM) ant_: yes
(10:10:22 AM) jsdelfino: ok, BTW is it samples/sca instead of sca/samples?
(10:10:29 AM) simonnash left the room (quit: Read error: 104 (Connection
reset by peer)).
(10:10:33 AM) ant_: what?
(10:10:55 AM) jsdelfino: I said sca/samples earlier, but I guess it's
samples/sca, rick is gonna make that change
(10:10:58 AM) ant_: i'll leave it till all the moving has happened
(10:10:59 AM) cr22rc: it should be sample/sca/<etc> right ?
(10:11:00 AM) jsdelfino: right?
(10:11:03 AM) jsdelfino: yes
(10:11:14 AM) jsdelfino: I mispoke before: samples/sca
(10:11:26 AM) jsdelfino: it is samples/sca
(10:11:27 AM) cr22rc: k had me freaked :-)
(10:11:33 AM) jsdelfino: :)
(10:12:15 AM) jsdelfino: if there's no other critical issue, let's move
on to the major ones
(10:12:54 AM) jsdelfino: looking at JIRA, ant you have 4 assigned to you
(10:13:04 AM) jsdelfino: 221 307 295 287
(10:13:40 AM) jsdelfino: which ones do u want in M1? what is their status?
(10:14:18 AM) jsdelfino: 287 we covered
(10:14:19 AM) ant_: 295 not done anything , was waiting to see what
others did to copy the format
(10:14:44 AM) ant_: 221, 307 mostly done waiting to commit
(10:14:46 AM) jsdelfino: 295 is not critical now, we can doc this on the
web site later after the distro
(10:14:52 AM) jsdelfino: k, good!
(10:15:10 AM) cr22rc: we closing out on this chat soon?
(10:15:30 AM) jsdelfino: I'd like to close in like 15 mns, does that
work for you guys?
(10:15:34 AM) kevin: yes
(10:15:42 AM) jsdelfino: k, dan, u have two:
(10:15:51 AM) jsdelfino: 357 354
(10:16:01 AM) jsdelfino: I guess u want 354 in M1
(10:16:03 AM) jsdelfino: ?
(10:16:13 AM) jsdelfino: what about 357? any update on that one?
(10:17:25 AM) jsdelfino: dkulp, not sure you're there, so moving on to
the next, ping when u get back
(10:17:43 AM) dkulp: Sorry. On the SCA/Java phone call.
(10:17:52 AM) jsdelfino: haleh has 160, 183, 345, all on the web site,
not critical for the distro
(10:18:22 AM) jsdelfino: dan, no pb, take your time, that's what I
guessed, a number of people on the spec call now
(10:18:47 AM) jsdelfino: isilval, you have one: 270
(10:19:15 AM) jsdelfino: what's your feeling about this one? do you want
any change to the BB sample to show async in M1?
(10:19:32 AM) jsdelfino: do u have anything - small enough change -
ready to contribute?
(10:20:38 AM) jsdelfino: waiting for ignacio's answer, moving on in the
meantime
(10:20:51 AM) isilval: I'm working on 270
(10:21:07 AM) jsdelfino: jeremy has 289 and 292 but he's out this week,
both are web site updates, so no critical for the distro
(10:21:11 AM) isilval: but there are a few issues with the data service
that I am trying to get around
(10:21:55 AM) jsdelfino: ok, just give us an update in the JIRA whenever
you have one, and depending on the extent of the patch, the cutoff for
the distro, and the state of SVN we'll include it or not
(10:22:59 AM) isilval: ok, I hope to have something, but I am being
hampered by unrelated exceptions, like the one in ant's latest comment
to jira-333
(10:23:02 AM) jsdelfino: personally I'd like this async stuff in BB, I
had discussed this with Jim before and he had also submitted his ideas
for this, so please describe as much as u can in the JIRA it'll help us
make the call on whether/when we apply the patch
(10:23:15 AM) jsdelfino: ok thx
(10:23:41 AM) jsdelfino: if anything is blocking you (in addition to
333) ask on the dev list
(10:23:51 AM) jsdelfino: thx a lot, next:
(10:24:00 AM) isilval: I already have a logger service that logs to
derby (tho not via DAS, but plain JDBC)
(10:24:17 AM) isilval: the web interface is becoming the tricky part
(10:24:22 AM) jsdelfino: ah I see
(10:24:27 AM) isilval: I can add a comment to that effect
(10:24:33 AM) isilval: to the jira
(10:25:10 AM) kevin: we can look at using the DAS when it is appropriate
(10:25:44 AM) jsdelfino: ok next, we have a few issues assigned to Jim,
he's not on the chat now
(10:26:00 AM) jsdelfino: I think we need 217 fixed for M1
(10:26:58 AM) jsdelfino: and maybe 301 but I'm not sure if it really
really important, will need to discuss with him and find the balance
between complexity of the fix and inconvenience for users
(10:27:14 AM) jsdelfino: in terms of usability, lets take a quick poll here:
(10:27:34 AM) jsdelfino: a POJO component can expose only ONE Java interface
(10:28:07 AM) jsdelfino: I tested this with Java5 annotations, wondering
if the same limitation applies if u use a .componentType file instead
(10:28:15 AM) cr22rc: dont see that as a huge limitation for this m1 IMO
(10:28:20 AM) jsdelfino: what do u guys think about this?
(10:28:41 AM) cr22rc: I prefer stablity and that limitation at this point
(10:28:46 AM) jsdelfino: we can always use Java interface inheritance to
have a component implement multiple interfaces
(10:28:51 AM) ant_: +1 to defer
(10:28:52 AM) jsdelfino: ?
(10:28:57 AM) jsdelfino: ok
(10:29:01 AM) cr22rc: +1 defer
(10:29:04 AM) jsdelfino: anybody else? dan what do you think?
(10:29:20 AM) dkulp: Defer is fine.
(10:29:49 AM) ant_: 217 seems a bit more serrious
(10:30:19 AM) jsdelfino: k, I'd like jim's opinion too, so I'm going to
mark M1-tentative and if he's got a safe fix then we can put it back in
M1, or we defer
(10:30:36 AM) jsdelfino: ant_, agreed
(10:30:50 AM) ant_: for 217 or 310?
(10:30:56 AM) ant_: 301
(10:31:06 AM) jsdelfino: 301 tentative/defer
(10:31:07 AM) ant_: +1 to defer 301
(10:31:17 AM) jsdelfino: 217 in M1
(10:31:28 AM) jsdelfino: if no objection, let's move on:
(10:31:36 AM) ant_: what does that mean, no M1 till its fixed?
(10:31:55 AM) jsdelfino: yes
(10:32:28 AM) jsdelfino: I think it's in the same category as 333
(10:32:39 AM) jsdelfino: confusing, in your face issues
(10:32:47 AM) jsdelfino: that break us in a server environment
(10:33:18 AM) jsdelfino: next: I have 180 assigned to me
(10:33:25 AM) jsdelfino: looks like a simple fix
(10:33:46 AM) jsdelfino: and will avoid that we have to explain to
everybody that our syntax for portTypes is completely different from the
spec
(10:34:06 AM) jsdelfino: ns#name instead of the spec'd notation
(10:34:20 AM) jsdelfino: could create lots of confusion
(10:34:29 AM) jsdelfino: I'm planning on fixing it today
(10:34:36 AM) cr22rc: we don't any plain sdo samples -- right ?
(10:34:42 AM) jsdelfino: no
(10:34:56 AM) jsdelfino: for 180 I'll keep the exsiting notation in
place to not break anybody
(10:35:06 AM) jsdelfino: but I'll adjust the samples ready at the time
to the new notation
(10:35:09 AM) jsdelfino: ok?
(10:35:16 AM) ant_: doe sthe fix to 180 mean all the existing scdl needs
to change?
(10:35:18 AM) ant_: ok ok
(10:36:17 AM) jsdelfino: rick you have 349 and 347, these are important
for the distro but can be done after tomorrow morning's cut-off
(10:36:23 AM) jsdelfino: small fixes to the readmes
(10:36:41 AM) jsdelfino: speaking of readmes...
(10:37:01 AM) jsdelfino: I'm gonna create 3 JIRA issues for the 3 sca,
sdo, das readmes
(10:37:18 AM) cr22rc: ?
(10:37:24 AM) jsdelfino: which we need to have ready for our release build
(10:37:54 AM) cr22rc: little more explanation
(10:38:25 AM) jsdelfino: thinking... we need ONE README instead sorry
(10:38:44 AM) jsdelfino: with the list of limitations, issues etc for
SCA, SDO, and DAS
(10:38:57 AM) jsdelfino: so one REAME.txt file
(10:39:02 AM) jsdelfino: README
(10:39:13 AM) cr22rc: at the top .. a readme.htm
(10:39:18 AM) jsdelfino: 3 JIRA issues for the 3 projects
(10:39:25 AM) kevin: sounds good
(10:39:28 AM) jsdelfino: and we close the JIRA issues when the work is done
(10:39:59 AM) jsdelfino: that'll be a GO signal to know when it's ready
to go into the distro build
(10:40:07 AM) jsdelfino: question:
(10:40:13 AM) jsdelfino: TXT or HTM?
(10:40:25 AM) cr22rc: htm.. links to the other htms
(10:40:37 AM) jsdelfino: which could be a nightmare to set up :)
(10:40:44 AM) rfeng: I caught the ClassCastException problem
(10:40:50 AM) jsdelfino: what other Apache projects do?
(10:40:52 AM) rfeng: it could be a tomcat issue
(10:40:55 AM) ant_: txt
(10:41:05 AM) jsdelfino: I vote for txt then
(10:41:10 AM) jsdelfino: easier to write
(10:41:15 AM) jsdelfino: no linking problems
(10:41:16 AM) jsdelfino: :)
(10:41:27 AM) rfeng: I'm seeing two classloaders for the same BB web app
(10:41:30 AM) ant_: axis2 does txt
(10:41:36 AM) jsdelfino: rfeng, can u create the empty file? README.txt
(10:41:47 AM) jsdelfino: include it in the distro build
(10:41:52 AM) rfeng: ok
(10:42:05 AM) jsdelfino: people will fill in the blanks
(10:42:09 AM) jsdelfino: thanks
(10:42:34 AM) jsdelfino: we have 3 unassigned major issues
(10:42:37 AM) cr22rc: all of the rest or our readme s or htm ... seems
odd IMO
(10:42:43 AM) jsdelfino: 276 310 328
(10:43:09 AM) jsdelfino: IMO 276 can be defered to after we cut the distro
(10:43:37 AM) jsdelfino: 310 needs to be in M1
(10:43:51 AM) jsdelfino: 328, I'm not sure, ant_ can u give us more info
about this one?
(10:43:54 AM) ant_: 276 we have some in interop - they'r 'real'
(10:44:03 AM) jsdelfino: yes you're right
(10:44:36 AM) jsdelfino: I can close 310 as a dup of 276
(10:44:43 AM) jsdelfino: if nobody objects that's what I'll do
(10:44:49 AM) ant_: but they should be more, but i'm fine with defering
that now
(10:45:01 AM) jsdelfino: you're right sorry I was too quick
(10:45:10 AM) jsdelfino: let's keep this one open, defer to Mx
(10:45:11 AM) jsdelfino: :)
(10:46:01 AM) ant_: 328 i think is a problem with the testcase not the
code but i couldn't figure out whta was going on
(10:46:16 AM) jsdelfino: 310, I wanted to help here but I'm not a Maven
expert, rfeng could you find out with Dan how to best gen the javadoc,
and then I volunteer to review it and adjust as necessary
(10:46:50 AM) rfeng: sure
(10:46:56 AM) dkulp: Not easy.
(10:47:00 AM) jsdelfino: ok, let's move 328 to M1-tentative, if jboynes
is available before we cut the release then we can maybe fix it
(10:47:23 AM) jsdelfino: ok, we went through all the JIRA issues I think
(10:47:23 AM) dkulp: Actually, is all the stuff that needs javadocing in
one sub-project?
(10:47:44 AM) jsdelfino: I'm not sure
(10:47:48 AM) jsdelfino: we have the spec project
(10:47:54 AM) jsdelfino: then we have core
(10:47:56 AM) jsdelfino: and model
(10:48:28 AM) dkulp: If we want three separate "javadoc" dirs, the
javadoc stuff is easy.
(10:48:29 AM) jsdelfino: not eveyrthing in core and model absolutely
needs javadocing but I'd be OK with javadocing the whole project for now
(10:48:40 AM) dkulp: If you want a single cross linked javadoc, it's not.
(10:48:52 AM) jsdelfino: i'm fine with the easy option as a starting
point :)
(10:48:56 AM) haleh left the room (quit: Read error: 110 (Connection
timed out)).
(10:49:04 AM) ant_: why do we need javadoc for all of core and model?
(10:49:09 AM) jsdelfino: when do u guys think that javadoc should be built?
(10:49:15 AM) jsdelfino: the regular build?
(10:49:16 AM) kevin: i think 3 separate docs is best
(10:49:18 AM) jsdelfino: the distro build?
(10:49:21 AM) jsdelfino: the site build?
(10:49:41 AM) jsdelfino: ah, kevin, do u need any of the das project
javadoc'ed as well
(10:49:43 AM) jsdelfino: ?
(10:49:47 AM) kevin: i was thinking the dist
(10:49:48 AM) kevin: yes
(10:49:49 AM) dkulp: I'll profile it.... (mvn -Pjavadoc)
(10:50:06 AM) kevin: but, would like it published
(10:50:35 AM) jsdelfino: ok so we need spec, sca/model, sca/core, das,
I'm not sure if we need the sdo/impl
(10:50:44 AM) jsdelfino: profiled as dan said
(10:50:45 AM) kevin: could the dist build update the site with the jdoc?
(10:50:59 AM) jsdelfino: wow :)
(10:50:59 AM) kgoodson is now known as kg_away
(10:51:06 AM) jsdelfino: that would be nice
(10:51:31 AM) jmarino [EMAIL PROTECTED] entered the room.
(10:51:33 AM) jsdelfino: but I have no idea how you'd do it
(10:51:36 AM) jsdelfino: hey jim
(10:51:36 AM) pombreda left the room (quit: No route to host).
(10:51:39 AM) jmarino: hey
(10:51:47 AM) ant_: hi
(10:51:48 AM) jmarino: on the SCA call too but can watch
(10:51:50 AM) jsdelfino: we're discussing javadoc now, how we build it
(10:51:50 AM) jmarino: hey
(10:52:48 AM) jsdelfino: looks like dan is volunteering to help produce
the javadoc build, I'd say we let dan decide what's simplest for him now
(10:52:54 AM) jsdelfino: that will be a starting point
(10:52:56 AM) jmarino: yea
(10:53:01 AM) jsdelfino: multiple javadoc projects
(10:53:03 AM) jsdelfino: a profile
(10:53:17 AM) jsdelfino: and dan decides the level of automation
(10:53:32 AM) dkulp: The easiest is to add a javadoc profile to the top
level pom
(10:53:40 AM) dkulp: When run, EVERYTHING get's javadoced
(10:53:41 AM) jsdelfino: dan, are u ok with that, hope it's not too much
in addition to your celtix stuff?
(10:53:50 AM) jsdelfino: well, that works for me too :)
(10:53:57 AM) dkulp: Then it's just an assmebly thing to select the
javadoc we want to include in the dist.
(10:53:59 AM) jsdelfino: the more javadoc the better after all
(10:54:27 AM) dkulp: Each subproject is javadoced (and installed into
your ~/.m2/repository dir).
(10:54:32 AM) jsdelfino: I'd be ok to javadoc everything if you guys
think it still makes sense, jim what do you think? others?
(10:55:00 AM) kevin: sounds good. i have to run now
(10:55:25 AM) jsdelfino: dkulp, are u ok with that? you know this stuff
much better than most of us anyway :)
(10:55:35 AM) dkulp: Sure. Assign it to me.
(10:55:42 AM) jsdelfino: great, thx a lot
(10:56:17 AM) jsdelfino: jim since you're back
(10:56:34 AM) jsdelfino: we discussed earlier two issues that were
assigned to you
(10:56:54 AM) jsdelfino: 217 - we think it's important to have this
fixed in M1
(10:57:02 AM) kevin left the room (quit: ).
(10:57:03 AM) jsdelfino: 310 - consensus was to defer
(10:57:17 AM) jsdelfino: but I wanted to get your input, see if you'd
like this fixed for M1 or not
(10:58:16 AM) jsdelfino: and for 217 - this is the issue we talked about
before this week, that system components are not scoped per app, the
user experience can be real confusing with WSDLs from diff apps stepping
on each other, so I'd be ok with any workaround / hack to get over this
one for M1
(10:58:34 AM) jsdelfino: so what do you think?
(10:58:49 AM) jsdelfino: will give you a little bit of time to type
since I know you're on a call :)
(10:59:10 AM) jsdelfino: in the meantime, I'd like to have the next IRC
tomorrow
(10:59:17 AM) jsdelfino: same time?
(10:59:47 AM) jsdelfino: and also ask people if there is anything else
we forgot to talk about which is really important for the release
(11:00:29 AM) jsdelfino: and anything they would like me to do (or do
better) to facilitate things in the next few days as we are trying to
wrap up M1
(11:00:43 AM) ant_: we're going to need to vote on the tuscany-dev list
for a release at some point
(11:00:56 AM) dkulp: We need SVN to work first
(11:01:44 AM) dkulp: If they don't get SVN working till monday, we
obviously cannot vote before then.
(11:02:51 AM) jsdelfino: yes, I was thinking about this order: get svn
working first, build the distro, adjust the readmes etc, I'll upload the
distro to my people.apache.org page, give everybody a day to try it out,
anything big that comes up we respin a distro, and then we start a vote
on our dev list
(11:03:06 AM) jsdelfino: while the vote is going, everybody can polish
docs, the web site etc.
(11:03:16 AM) ant_: a week day or maybe sat/sun?
(11:03:47 AM) jsdelfino: isn't a week just 7 days of work?
(11:03:48 AM) jsdelfino: :)
(11:04:00 AM) jsdelfino: I don't think we'll vote on the weekend
(11:04:14 AM) ant_: ok thats what i was asking
(11:04:24 AM) jsdelfino: we'd start to vote as soon as we are
comfortable enough with the contents next week
(11:04:36 AM) jsdelfino: tentatively on monday
(11:04:52 AM) ant_: so not much cjance of a release before thursday
then. is that ok?
(11:05:08 AM) jsdelfino: there is a chance :)
(11:05:27 AM) ant_: IPMC need 3 days after our vote
(11:05:31 AM) jsdelfino: I could post the distro on sunday night, we
start voting on monday morning
(11:05:56 AM) pombreda
[EMAIL PROTECTED] entered the room.
(11:06:04 AM) jsdelfino: again it's difficult to set a cut-off with SVN
being down right now...
(11:06:17 AM) jmarino: sorry back
(11:06:18 AM) ant_: i know, just pointing out logistics
(11:06:25 AM) rfeng: hi, ant.
(11:06:35 AM) rfeng: I found the CCE problem in the ClassLoaderHelper
(11:06:56 AM) rfeng: it doesn't take the app switch into consideration
(11:07:25 AM) rfeng: even it's on the same thread, the TCCL can be
changed when it spans multiple apps
(11:07:25 AM) jsdelfino: we can cut the distro as soon as the critical
and major JIRAs are fixed and assuming that no other big issue gets
raised, and when we have all the readmes ready
(11:07:52 AM) ant_: do you have a patch? ( i wont even mention a testcase)
(11:07:54 AM) jmarino: to fix 217, when would that have to be in?
(11:08:07 AM) jsdelfino: rfeng, that's great news, one more step toward
our release!
(11:08:07 AM) rfeng: and it incorrectly switch to account CL when it
tries to deserialize the data for web-client
(11:08:25 AM) rfeng: we need to think about how to fix it
(11:08:32 AM) jsdelfino: jim, some time tomorrow before end of the day
would still be OK
(11:08:56 AM) jsdelfino: jim, what do you think? is that feasible?
(11:08:57 AM) rfeng: having jars in server/lib gives us a lot of trouble :-)
(11:08:58 AM) cr22rc is now known as cr22rc_gone
(11:09:03 AM) jmarino: k I can't get into svn now and am completely out
of sync which means I won;t be able to start until tomorrow
(11:09:18 AM) jmarino: the only way I can fix it is if I do a hack
similar to @SDOHelper
(11:09:26 AM) ant_: rfeng, little chance of changing that, look what it
was like suggesting moving just stax
(11:09:27 AM) jmarino: do you think that would work?
(11:09:37 AM) jsdelfino: iirc I added some comments to the JIRA issue
with some hints on how we may be able to implement workarounds for the
underlying issue
(11:09:46 AM) jsdelfino: jmarino, yes!
(11:09:47 AM) jmarino: k
(11:09:54 AM) rfeng: ant, it seems that each app needs to have a
ClassLoaderHelper
(11:10:01 AM) jsdelfino: when are you traveling back?
(11:10:27 AM) ant_: could maybe do that during build
(11:10:30 AM) ant_: phase
(11:10:38 AM) jsdelfino: jim, if you know you're gonna be disconnected
for some time, and u have a good idea on how to fix this, but no svn
access, maybe attach the patch to the JIRA
(11:10:43 AM) kg_away left the room (quit: Read error: 110 (Connection
timed out)).
(11:10:46 AM) jsdelfino: and then as soon as svn gets back up we can
apply iy
(11:10:48 AM) jsdelfino: it
(11:10:49 AM) cr22rc_gone: rfeng.. but wouldn't what you are suggesting
mean a solid failure ?
(11:10:53 AM) rfeng: yes, keep the ClassLoaderHelper in the context scope
(11:10:55 AM) ant_: but tricky to get at everywhere its needed
(11:11:11 AM) rfeng: right
(11:11:23 AM) cr22rc_gone: and thats not what we see .. right?
(11:11:24 AM) jsdelfino: jmarino, wrt tuscany-310, what do you think?
(11:11:37 AM) ant_: also this code is where i've been refactoring to get
e4x working so you don't have the same code as i have
(11:11:39 AM) jsdelfino: do you want to defer (evyerbody would be ok
with that)
(11:11:42 AM) jsdelfino: or do u want to fix it?
(11:11:54 AM) rfeng: ok
(11:12:05 AM) jmarino: hmm is it just javadoc for osoa?
(11:12:14 AM) jsdelfino: sorry wrong number
(11:12:19 AM) ant_: why is it intermitant?
(11:12:32 AM) rfeng: the ordering of app interaction matters
(11:12:40 AM) jmarino: jsdelfino which #?
(11:12:42 AM) jsdelfino: 301
(11:12:43 AM) rfeng: it can go out of sync
(11:12:58 AM) rfeng: I mean the CL switch bracketing
(11:13:06 AM) jmarino: I'd like to defer it for now if people are
alright with that
(11:13:19 AM) jmarino: I'd rather look at the wsdl thing even though I
may not be able to fix that
(11:13:23 AM) jsdelfino: yes people are alright with that
(11:13:25 AM) jsdelfino: good
(11:13:38 AM) jsdelfino: jim, yes, makes sense
(11:13:50 AM) rfeng: do we have accurate classloader for a given module?
(11:13:50 AM) jsdelfino: ok, so jim to give you an update
(11:14:08 AM) jsdelfino: I think we're looking pretty good now
(11:14:15 AM) ant_: the resourceloader has a classloader
(11:14:20 AM) jsdelfino: most of the remaining JIRA issues are web site
related
(11:14:20 AM) rfeng: and we can get the current module from the app context?
(11:14:31 AM) jsdelfino: and for the others we have a plan
(11:14:41 AM) jsdelfino: assuming that the apache svn gets back up soon :)
(11:15:10 AM) rfeng: then we can get the classloader from the current
module instead of playing threadlocal tricks?
(11:15:26 AM) jsdelfino: for all, if you run into anything that you
think we absolutely need to fix for M1, raise a JIRA issue, mark it
blocker to get attention
(11:15:49 AM) jsdelfino: last question:
(11:16:28 AM) jsdelfino: I'm thinking that it would be a good idea to
document the limitations, gotchas, useful info etc about the level of
code we're going to release in our Wiki
(11:16:39 AM) jsdelfino: instead of having too much in a frozen README
in the download
(11:17:02 AM) jsdelfino: this way we can evolve this as we go and play
around with the release, get users involved etc.
(11:17:12 AM) jsdelfino: does that make sense to you guys?
(11:17:45 AM) jsdelfino: ?
(11:19:19 AM) jsdelfino: I'm going to interpret silence as an agreement :)
(11:19:46 AM) ant_: a smaller readme pointing at the wiki sounds ok
(11:19:59 AM) jsdelfino: ok good
(11:20:10 AM) jmarino: yea +1
(11:20:45 AM) jsdelfino: k, I propose we close the chat for today, thank
you all for staying on this long chat
(11:21:01 AM) jsdelfino: I'll send the log to the dev list
(11:21:21 AM) jsdelfino: and will schedule another for tomorrow, we'll
make it shorter tomorrow
--
Jean-Sebastien
