Chat log attached. In summary:
* we talked about getting checkstyle/pmd working
there's an inability to build individual modules in the sandbox that
is not an issue in trunk
some concern over compatibility with Eclipse reformatting (IDEA
seems to be OK)
* want to have the sandbox code build on a clean machine
need to fix top-level pom version
need to include new spec module
need to fix version of sdo dependency in databinding module
need to set up to use the snapshot repo (for the sdo dependency)
* we will leave the new spec stuff in the sandbox for now
--
Jeremy
On 6/24/06, ant elder <[EMAIL PROTECTED]> wrote:
Thought about canceling this due to ApacheCon starting at the same
time...but there are a lot of people not going who'll still be able to make
the chat, and for those who'll already be at ApacheCon remember theres
little geekier than participating in your own projects IRC chat while at at
an ApacheCon! Ideally sit at the same table as others on the chat and don't
speak to each other :-)
Hi folks!
This is a reminder that the weekly Tuscany developer chat will be
occurring on Monday, June 19, at: 15:30 GMT, 16:30 BST, 08:30am PDT,
11:30am EDT, 21:00 Bangalore
The chat takes place on the freenode IRC network, (use server
irc.freenode.net), on channel #tuscany, and is scheduled to last one
hour, though it may run longer. Please join us!
If you need an IRC client for windows, check out http://www.mirc.com,
and http://www.mirc.com/links.html has some links to clients for other
OS's.
Thanks,
...ant
[08:31] <ant_> hi
[08:31] <jboynes> hi
[08:31] <kgoodson> hi
[08:31] <gwinn> hi
[08:31] <cr22rc> hi
[08:32] <jliu> hi guys
[08:32] *** kevin ([EMAIL PROTECTED]) joined
[08:34] <jboynes> I'd like to remind any committers who haven't done so yet to
sign up on the private mailing list
[08:35] <rfeng> what are the topics for today?
[08:35] *** dkulp ([EMAIL PROTECTED]) joined
[08:38] <ant_> does anyone have any? what about that launcher thread on the
mailing list? the sandbox...a few of us are at apachecon this week so will be
talking a lot aboutthat i expect
[08:39] <jboynes> happy to chat about the launcher if anyone's interested
[08:39] <jboynes> beats quiet :-)
[08:40] <jboynes> I'd also ask if anyone is interested in setting up the
profile for checkstyle/pmd
[08:41] *** jliu quit ("Download Gaim: http://gaim.sourceforge.net/";)
[08:41] <jboynes> the thing Jim, Dan and I were chatting about
[08:42] *** jliu ([EMAIL PROTECTED]) joined
[08:42] <cr22rc> seems like some key people missing jim, sebastien
[08:43] <ant_> what does 'setting up the profile' involve?
[08:44] <dkulp> Two parts:
[08:44] <cr22rc> (turning everyting off ) JUST KIDDING !
[08:44] <dkulp> 1) Setting up a "pluginManagement" section to configure the
plugins (not a big deal, almost done)
[08:44] <dkulp> 2) Setting up a "profile" that turns on the wanted plugins when
that profile is invoked
[08:45] <dkulp> Actually: 3) Updating the "default' (non-profile) case to
remove the unwanted plugins
[08:45] <dkulp> FYI: it's ALL already done in the trunk. It's just the
sandbox code that isn't done.
[08:45] <jboynes> I'd suggest doing it like celtix i.e. on by default but with
a profile to turn it off
[08:46] <ant_> looking back on the ML, there was a comment that you with
checkstyle on you can't build individual modules. has that been fixed?
[08:46] <dkulp> On trunk, there is a "properties" flag in each module that says
whether to fail the build or just "warn" for that module
[08:47] <dkulp> For modules that aren't fixed yet, the flag is set to just
"warn". For Celtix and others that are fixed, it's set to fail.
[08:48] <jboynes> ant_, not that I know of, I know there were some suggestions
[08:48] <jboynes> dkulp, is that also a problem in trunk?
[08:49] <dkulp> No, trunk works fine. You should be able to run "mvn
-Psourcecheck" and it will run spitting out warnings for all the checkstyle
issues.
[08:49] <jboynes> even in individual modules?
[08:50] <dkulp> Yep. (may need to have built from root once) (aka; parent of
sca)
[08:51] <ant_> do we need to update our eclipse style template stuff to match
this?
[08:51] <jboynes> ok, so there's a template that shows what need to be done to
the sandbox modules
[08:51] <jboynes> yes?
[08:52] <dkulp> We can make the sourcecheck profile to be "activeByDefault" if
we want that, then create a new "noSourceCheck" or similar profile for
developers to use.
[08:52] <rfeng> Just want to make sure: Is PMD from http://pmd.sourceforge.net/?
[08:52] <dkulp> The eclipse style templates should be OK except for the line
length. Not really sure though.
[08:52] <dkulp> rfeng: yes.
[08:52] <ant_> what is the line length now?
[08:53] <jboynes> 120
[08:53] <dkulp> I think Jim changed the one in the sandbox to 120.
[08:53] <dkulp> Or Jeremy. :-)
[08:53] <jboynes> I thought that is what we were using
[08:54] <jboynes> I don't mind if we go wider :-)
[08:54] <dkulp> Honestly, I'd like to go less (110), but 120 is OK.
[08:54] <cr22rc> Are we suggesting ... 80 .. please say no.
[08:54] <ant_> have any of you been using eclipse and our current style
template with this? its goingto suck a bit if everytime i do a format in
eclipse it breaks somthing in checkstyle
[08:55] <cr22rc> I've used it.
[08:55] <jboynes> I've been using idea and needed to change one thing (the
indent on a wrapped "throws" clause")
[08:55] <jboynes> that was it
[08:56] <dkulp> I don't know. I never use the Eclipse reformat. If the code
starts off compliant, and I write compliant code, I don't need the reformat.
:-)
[08:57] <dkulp> There may need some changes to things like the eclipse import
order and such. I don't really know as I use celtix to setup the eclipse
workspace which handles all of that.
[08:57] <jboynes> I use reformat quite a bit and it seems ok
[08:58] <rfeng> reformat helps me as well
[08:58] <ant_> i use it all the time so i hope it can be compatable
[08:58] <rfeng> I trust machine more than human beings :-)
[08:59] <rfeng> we just have to make the check-style and the templates for IDEs
consistent
[08:59] <jboynes> what I don't like about eclipse is that it seems too literal
[09:00] <jboynes> but this isn't meant to be an ide feature fest :-)
[09:01] <ant_> do you have much idea of how much more work there is to get
everything passing? how many modules in trunk are set to fail or warn?
[09:02] <dkulp> Only celtix is set to fail I think. Axis and jsonrpc would
be easy as they are mostly there.
[09:03] <dkulp> I don't know about the rest. A simple reformat in
Eclipse/IDEA would probably fix 90% or more of the rest.
[09:03] <dkulp> However, that will KILL all the merges to/from the sandbox
[09:03] <dkulp> Thus, if we're planning on integrating the sandbox code, I
suggest just doing it there.
[09:04] <dkulp> Honestly, I'd like to see a COMPLETE branch (including spec,
sdo, das, etc...) instead of the sandbox if we are planning to use the sandbox
code. We talked about that a couple weeks ago.
[09:05] <ant_> i agree with you, but its the "...if we are plannig to use..."
that remains unanswered
[09:06] <cr22rc> If there is going to be a lot of merging ... maybe this isn't
the time to activate this?
[09:07] <jboynes> dkulp, you raise an interesting question - do we need to keep
everything in lockstep?
[09:08] <jboynes> I mean spec, sdo, sca, das, ...
[09:08] <jboynes> when we started out we set things up so that they could be
branched/released independently
[09:11] <ant_> i have to leave early sorry...ttyal
[09:11] *** ant_ changed nick to ant_away
[09:12] <jboynes> so for example I can see us doing several tuscany releases
against the same spec jars
[09:12] <jboynes> or multiple sca releases with the same sdo version
[09:12] <jboynes> does anyone see a problem with that?
[09:13] <dkulp> No, but until things are available in the PUBLIC maven
repositories, a new user needs to be able to run a single command to get
everything built and runnable.
[09:14] <dkulp> Right now, if you wipe out your ~/.m2/repository directory, you
CANNOT build the sandbox.
[09:14] <jboynes> perhaps, although we should be able to fix that
[09:15] <dkulp> Even if you could build it, the fact is you need to first built
parts of trunk, then go to the sandbox, build the spec parts, then the sca
parts, etc...
[09:16] <jboynes> sure, although we can fix that without creating a complete
branch
[09:16] <cr22rc> wouldn't we have different poms referencing different levels
depending what people were interested in? How do we check those in?
[09:16] <jboynes> not sure what you mean "different poms referencing different
levels depending what people were interested in" ?
[09:17] <dkulp> I guess my point is that if we are epecting developers to have
a look at the sandbox code, let's get it to a point where it is buildable
without going through a lot of hoops, checking out different versions/areas of
the repository, etc....
[09:18] <jboynes> ok I'm lost
[09:18] <jboynes> we need to add the new spec stuff in I agree
[09:18] *** kevin quit ( )
[09:19] <jboynes> for the rest, now m1 is in the maven repo do we just need to
add that repo to the pom?
[09:19] *** kevin ([EMAIL PROTECTED]) joined
[09:20] <dkulp> You also need to update all the version numbers in the
dependencies. They currently just look for SNAPSHOT.
[09:21] <jboynes> for sdo?
[09:21] <dkulp> For the "PARENT" of the sandbox/sca/pom.xml as well.
[09:21] <jboynes> ,ok
[09:22] <jboynes> so we need to change the sca pom to change the version and to
add the snapshot repo
[09:23] <jboynes> and we need to change the version in the sdo databinding pom
[09:23] <jboynes> and we need to figure out what to do about the spec changes
[09:23] <jboynes> anything else?
[09:24] <dkulp> Well, all the versions in the sandbox should probably be
incremented to 1.0-incubating-M2-SNAPSHOT.
[09:25] <jboynes> perhaps something else: sandbox-SNAPSHOT I would suggest
[09:26] <dkulp> Sure. In anycase, I think we have to make sure the sandbox is
buildable from a "fresh" checkout and a "fresh" ~/.m2/repository
[09:26] *** jsdelfino ([EMAIL PROTECTED]) joined
[09:28] <jboynes> so what should we do with the spec changes?
[09:28] *** kevin quit ( )
[09:28] <jboynes> leave them where they are, move them to a branch, move them
to trunk?
[09:28] *** gwinn quit (Read error: 104 (Connection reset by peer))
[09:29] <jboynes> move them somewhere else?
[09:29] *** gwinn ([EMAIL PROTECTED]) joined
[09:29] <dkulp> Where they are is fine, as long as the version number is
incremented and we create a top-level pom that will include it in the build.
[09:30] *** kevin ([EMAIL PROTECTED]) joined
[09:30] <jboynes> ok
[09:31] <jboynes> anything else?
[09:33] <dkulp> Not that I can think of right now. :-)
[09:34] <jboynes> ok
[09:34] <jboynes> anyone else?
[09:34] <jboynes> or is this a wrap?
[09:34] <kgoodson> no
[09:35] *** gwinn_ ([EMAIL PROTECTED]) joined
[09:36] <dkulp> I think we're OK.
[09:36] <jboynes> ok - thanks folks, talk to y'all later
[09:36] <kevin> bye
[09:36] <dkulp> If we can make it an action item for the week to get the
sandbox buildable, that's great.
[09:36] *** jliu ([EMAIL PROTECTED]) left ()
[09:36] *** kgoodson changed nick to kgoodson_away
[09:37] <jboynes> how's it go, oh yeah - "thanks for volunteering" :-)
[09:37] <jboynes> jk - I'll work on it
[09:37] <dkulp> :-)
[09:37] <dkulp> One of the guys here has a saying: "ask, and ye shall receive
the action item"
[09:38] <cr22rc> LOL
[09:38] *** gwinn quit (Read error: 104 (Connection reset by peer))
[09:38] *** gwinn_ changed nick to gwinn---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
