Transcript: App Engine chat time April 15
10 views
Skip to first unread message
Jason (Google)
Apr 16, 2009, 1:34:53 PM4/16/09
to Google App Engine
[09:05am] drome: Question about GAE-Java: I'm using SmartGWT + GWT in
an application and I would like to use GAE-Java for the backend. But
currently SmartGWT imposes me thousands of static files, and GAE only
allows a maximum of 1000 files. How to solve this problem?
[09:05am] drome: (Would this problem be solved, if GAE supported the
upload of a single ".war" file for each application?)
[09:06am] tobyr: RobHanson: The best thing to do is file a feature
request, and we'll see what we can do.
[09:07am] lexs2: batch delete!
[09:07am] RobHanson: tobyr: I am sure there is already a feature
request... just wondering on the reasoning why it wasn't included
initially, since it seems to break a lot of things.
[09:08am] wycats: sorry I'm late
[09:08am] schwardo: fedestabile: that should not be the case, but i
haven't personally tested it. please file a bug on the issue tracker
and we'll follow up
[09:08am] • wycats is with knowtheory
[09:08am] scudder_google: (btw jcgregorio and dan_google are also
Googlers)
[09:08am] • dan_google waves.
[09:08am] alexrudnick: scudder_google: Also alexrudnick!
[09:08am] • jcgregorio waves
[09:08am] nickjohnson: If you're looking for help on how specific APIs
work at the stub level, I and others hang out in here semi-
permanently. And there's the PYthon source.
[09:08am] knowtheory: tobyr: My goal was to try and get apps on JRuby
into an environment that was as much like GAE as possible. We can run
Ruby apps via the Rack standard. ribrdb_ pointed out that future
releases of the GAE dev server may have additional tools available
[09:08am] wycats: following up on what knowtheory said before, it
would be nice to be have documentation for hooking up to the mock
datastore without spinning up a mock server
[09:08am] tobyr: RobHanson: There's not any real context for which the
naming API is useful in our environment. Any use you tried to make of
it would result in a SecurityException later down the line
[09:09am] Jason_Google: fedestabile: Yes, please include your
application ID, and if possible, test the upload from the appspot
domain to see if it's an issue with external domains.
[09:09am] knowtheory: that we wouldn't have in a vanilla Ruby + Rack
compliant server
[09:09am] fedestabile: schwardo: yes, it surprises me too... ok, I'll
post this "issue"
[09:09am] scudder_google: drom: I don't know much about SmartGWT, but
would it be possible to zip up the files into a jar and upload with
your app?
[09:09am] knowtheory: Or alternatively, having an api so that we could
write something that's compliant without having to hook into the dev
server & spin the whole environment up (as wycats pointed out)
[09:10am] knoonan: You have oodles of dynamic languages to track now
(congratulations!)--so in case you missed this: Alan Kennedy got
servlets written in Jython running on App Engine on Modjy ("Jython
WSGI to Java servlet bridge") http://jython.xhaus.com/
[09:10am] tobyr: wycats: There's already documentation on using the
datastore without the dev_appserver - Read the docs on unit testing.
[09:10am] fedestabile: Jason_Google: I've tested the upload from the
appspot and works perfectly
[09:11am] Jason_Google: fedestabile: Good to know. Please include that
in your issue report, and we'll look into it straightaway.
[09:11am] drome: scudder_google, the thing is the static files include
javascript, png, css, html, xml... not all the files are read by
application code. Some of them are just to be read by the bowser
[09:11am] FoosMasta: Are there scenarios that we should avoid with
bigtable from the Java AppEnging APIs? It seems to handle simple Map
structures well, but what about more complex OR maps?
[09:11am] tobyr: drome: You can unzip your files on the fly using a
servlet or filter
[09:11am] fedestabile: Jason_Google: actually, http://www.myclimbs.com/uploadform
or http://myclimbs.appspot.com/uploadform is a very simple page that I
made to show this error
[09:12am] fedestabile: Jason_Google: ok, I'll post it
[09:12am] bthomson: what was recently changed in app engine's tos?
[09:12am] wycats: tobyr: those docs are pretty coupled to JUnit
[09:12am] drome: I'm mentioning SmartGWT, but I imagine there are a
lot of javascript frameworks that will suffer from the 1000 files
limitation
[09:12am] wycats: I couldn't tell what I still needed to do if I
didn't want to use JUnit
[09:12am] cylob: in the datastore how do i update an existing entity,
rather than create a new one?
[09:13am] schwardo: FooMasta: unfortunately our resident ORM expert
(maxr) couldn't make it today, so we can't rattle off a list of
caveats. Your best bet is to just try it and if you run into
problems, post on the forum.
[09:13am] nickjohnson: cylob: Java or Python?
[09:13am] cylob: python
[09:13am] nickjohnson: Just call .put() on an existing entity
[09:13am] nickjohnson: There's no difference between insertion and
update
[09:13am] cgeorg: instance = YourClass.get(key)
[09:13am] schwardo: wycats: the ApiProxy method calls are the
important part -- it should be trivial to port the JUnit test to a
different environment
[09:13am] cgeorg: instance.something = 'foo'
[09:13am] drome: tobyr: sorry, a servlet to return static content?
[09:13am] cgeorg: instance.put()
[09:13am] cylob: ok let me meditate on this for a sec ty
[09:13am] lexs2: nice error
[09:14am] lexs2: "Timeout: datastore timeout: operation took too
long."
[09:14am] scudder_google: drome: ah ok, one way you could work around
the 1000 file limit would be to store the files in the datastore and
write a servlet to handle those requests (would be good to use
memcache/jcache as well)
[09:14am] AsharLohmar: cylob: try the clasic aproach, ge tit using
getby id or something, change its props an the do a .put()
[09:14am] tobyr: wycats: None of that code is Junit specific. You
would need to do everything in the sample code as shown - setting the
ApiProxy delegates, environments, etc...
[09:14am] AsharLohmar: cylob: sorry for the misspelling
[09:14am] wycats: schwardo: trivial if you know Java
[09:14am] tobyr: drome: Yes - You would return static content via a
servlet or filter
[09:14am] scudder_google: drome: tobyr also had a great sugestion,
don't know how I missed that
[09:15am] knoonan: This is easily the busiest App Engine IRC session
I've seen (at the European-friendly timeslot)--breaking the one-
hundred-participant barrier!
[09:15am] dinguva: Googlers: I routinely clone my prod database into a
development db so I can run tests on them. I plan to run 2 different
applications for prod and dev and want to findout if i can clone the
data between two different applications in appengine. Is that
possible?
[09:15am] schwardo: wycats: sounds like a good opportunity to learn
[09:15am] drome: tobyr: ok, do happen to have a pointer for me, to
investigate further? (some examples...)
[09:16am] nickjohnson: dinguva: You'll need to write your own code to
do that. There's currently no built-in support for it.
[09:16am] wycats: schwardo: sure
[09:16am] schwardo: dinguva: you could use two different version
numbers, if you always want the data to be in sync
[09:16am] Jason_Google: bthomson: We usually publish a revision note
when we update the TOS but it doesn't look like this made it in last
week. It's hard to summarize, but I can tell you the sections that
changed if that will help.
[09:16am] schwardo: dinguva: if you wanted to do something more
complex like reload prod -> dev periodically then you would need to
develop a mechanism for syncing the data
[09:16am] nickjohnson: As schwardo points out, you can use the exact
same datastore if you're happy with that
[09:17am] AsharLohmar: on what dinguva said, what about an local
instance of the db ?!
[09:17am] tobyr: drome: sorry, no specific examples - but it's pretty
straightforward. This might be a common enough usecase that we could
provide an example in some docs somewhere
[09:17am] dinguva: schwardo: true. but, if i change the data with a
beta version of application, it'll change the actual prod data too !
So was wondering if I can use a clone of the data for the beta version
for say user acceptance testing.
[09:17am] AsharLohmar: i mean what if i wnat an local "clone" of the
db ?
[09:17am] dinguva: schwardo: basically looking at GAE for the
enterprise
[09:18am] knowtheory: I've heard talk that the JRuby guys have been in
contact with you, regarding the fact that GAE is running the Client
JVM?
[09:18am] nickjohnson: AsharLohmar: There are tools like gae-bar for
uploading and downloading backups
[09:18am] serickso: I've got a question. We've been having issues with
502 error pages made from time to time on our application (id: skrit).
They've been tough to track down; they seem to happen on all kinds of
requests and only on our external domain, www.skritter.com. Nothing
about them shows up in our logs. Is there anything you guys could tell
us about 502 errors that would help us handle them or figure out
what's causing them?
[09:18am] tobyr: knowtheory: yes we've been discussing it with them
[09:18am] knowtheory: And that you guys would potentially consider a
switch to the server JVM if you could compile some compelling evidence
[09:18am] knowtheory: what would you like to see in an app for testing
purposes?
[09:18am] tobyr: knowtheory: there are many, many variables involved
[09:19am] knowtheory: Yes that's why i'm asking
[09:19am] knowtheory: beacause i'll happily put one together
[09:19am] drome: tobyr: ok, I'll post into the discussion board to get
this documented somehow...
[09:19am] AsharLohmar: nickjohnson: thanks
[09:19am] knowtheory: but i'm not entirely sure what to aim at.
[09:20am] knowtheory: i can build a stateless app that does some stuff
[09:20am] schwardo: knowtheory: i don't know how useful a test app for
that would be... we did extensive load tests across a wide range of
apps to arrive at our current config
[09:20am] knowtheory: or i can find a merb app that already exists out
there in the wild
[09:20am] dinguva: schwardo: just saw your reply about prod -> dev..
If an app can reach out and access another app's data that'd be
great ! We can even have a helper application that does this
[09:20am] Reuven: guys, when is the next release with maintrunk
changes for java? it seems from the posts that there are some very
important stuff over there
[09:21am] knowtheory: Okay, well, let me be more general then, and
ask, is there any information that i could help to provide regarding
the performance of ruby applications on the JVM?
[09:21am] knowtheory: Are you interested in any particular info?
[09:21am] scudder_google: serickso: could you tell us a bit more about
the 502s, roughly what percentage of requests they occur in. I'll look
into specific issues with that app.
[09:21am] tobyr: knowtheory: I do think having some simple vs complex
jruby apps could be useful, particularly if they have "benchmarking"
built in
[09:21am] bthomson: Jason_Google: if you know which sections that
would be great, I'd rather not read the whole thing again
[09:22am] cgeorg: Erin and her kitty:
http://icanhascheezburger.files.wordpress.com/2009/03/funny-pictures-cat-and-human-argue.jpg
[09:22am] Jason_Google: Reuven: schwardo might know better than me.
Hopefully in less than a few weeks' time.
[09:22am] Jason_Google: (being purposefully vague )
[09:22am] cgeorg: wow, wrong channel
[09:22am] • cgeorg hides
[09:22am] schwardo: Reuven: i can't give you a specific date, but
there are definitely some good fixes queued up that we want to get out
[09:22am] nickjohnson: cgeorg: Cat pictures are always welcome
[09:23am] cgeorg: Who do I pay to scrub that out of the official
transcript?
[09:23am] schwardo: lol
[09:24am] Jason_Google: bthomson: As far as I can tell, sections 7.4,
10.3, 10.4, 10.5 were the only ones that changed.
[09:24am] FoosMasta: Is there any documentation or source code
available for the Java GAE Eclipse plug-in?
[09:24am] Jason_Google: And I'll make sure any future changes are
summarized in the revision history.
[09:24am] Jason_Google: cgeorg: That would be me.
[09:24am] alexrudnick: FoosMasta: There's documentation!
http://code.google.com/eclipse/docs/users_guide.html
[09:24am] fedestabile: Jason_Google: I've posted the issue, it's the
number 1312. Thanks!
[09:25am] Jason_Google: Thanks fedestabile!
[09:25am] Jason_Google: I'll star it.
[09:25am] fedestabile: great!!
[09:25am] alexrudnick: FoosMasta: It's not open-source yet. But we're
going to open source sooner or later. (hopefully sooner)
[09:25am] bthomson: thanks Jason_Google
[09:25am] RobHanson: is there a reason why you can't delete an app?
will you be able to in the future?
[09:26am] dinguva: RobHanson: delete app ++
[09:26am] tobyr: RobHanson: We're aware of it, and we're taking into
consideration that many people are asking for it.
[09:26am] cylob: script = Script.get(self.request.get('dakey'));
script.datitle = self.request.get('datitle'); script.put(); this
is still creating a new entity instead of updating
[09:27am] serickso: scudder_google: they happen infrequently, and in
clusters. I think I've seen it as a page load once or twice, and I'm
developing the site. We have a flash client though that makes many
requests to the server through PyAMF and it happens enough so it gums
up communication between the client and the server from time to time.
Nick knows more about it than me, and he wrote about it here:
http://groups.google.com/group/google-
[09:27am] bthomson: RobHanson: you can use the old app for something
else, just map it to a new domain name
[09:28am] alexrudnick: FoosMasta: Have you tried out the plugin? Any
feedback you'd like to share?
[09:28am] dinguva: bthomson: basically recycle and be friendly to the
environment
[09:28am] knowtheory: A philosophical question, why do you guys
develop and then opensource after the fact?
[09:28am] knowtheory: (That's not a troll, i'm just curious about the
reasoning)
[09:28am] RobHanson: <bthomson> yeah, I am aware of that. but
sometimes you just want it to go away.... or rename it to something
that has meaning
[09:28am] scudder_google: serickso: Thanks for the details, I'll look
into it further after we end here.
[09:29am] nickjohnson: knowtheory: I can't speak authoritatively, but
I can speculate that it's largely because open-sourcing takes non-zero
work, so we'd rather have it out there sooner and follow up with the
source, than delay everything so we can release with the source.
[09:29am] knowtheory: (rather than develop out in the open that is)
[09:29am] knowtheory: ah okay, that's interesting
[09:29am] Jason_Google: knowtheory: It was my fault last week for
publishing the source code after the release. I'll make sure that it
happens same-day for future releases.
[09:30am] serickso: scudder_google: Thanks a bunch
[09:30am] FoosMasta: alexrudnick: yea - got stuck on the 400 Bad
Request - didn't know i needed to enable java for my account in
addition to signing up
[09:30am] knowtheory: Well, grateful in either regard, since i have...
plans And i'm glad to hear you guys are interested in keeping
everyone up to speed
[09:31am] alexrudnick: FoosMasta: Oooh. So maybe the plugin could tell
you that there was a problem, in that case?
[09:31am] FoosMasta: exactly
[09:31am] alexrudnick: FoosMasta: This was when you were trying to
deploy?
[09:31am] FoosMasta: yup
[09:31am] Jason_Google: cylob: That's an interesting issue. If get
returns a valid entity, calling put to that object should update the
entity instead of creating a new one. Do you just see a new entity
with every other field the same except for the one that you modified?
[09:31am] alexrudnick: FoosMasta: That's a good idea. Writing it down
[09:32am] idrism: hey, anyone know about having a servlet with /* as
the url-mapping?
[09:32am] Reuven: Anyone experienced a problem with changing a
Serializable object using getObjectById and changing the Serializable
object state, then closing the pm and nothing happens ...
[09:32am] schwardo: idrism: sure, what about it?
[09:32am] idrism: that screws up JSP views fo rme...
[09:33am] drome: Are you planning to support upload of a single '.war'
file that holds the entire application?
[09:33am] datanucleus: Reuven : have you looked in the log ?
[09:33am] datanucleus: (at debug level)
[09:33am] • knoonan_ got randomly disconnected from the server (just
got back)
[09:33am] cylob: Jason_Google: yep, that's what i see
[09:33am] Reuven: what do you mean, I can change other objects and it
saves them but not the Serializable one
[09:33am] schwardo: idrism: in the dev environment or the server
environment (or both) ?
[09:33am] scudder_google: idrism: I would think that a /* mapping
would match everything. What does your directory structure look like?
Are you trying to send all requests to one JSP page?
[09:34am] idrism: for example if i have a controller that uses
"test.jsp", and i hit that controller, it will say 404 for "/WEB-INF/
war/test.jsp"
[09:34am] datanucleus: Reuven : JDO detects changes to objects if you
do via setters. The DataNucleus log logs all changes and state changes
[09:34am] cylob: maybe it's affected by using the required=TRUE in the
dbmodel?
[09:34am] idrism: i think internally it is trying to go through that
same servlet mapping
[09:34am] idrism: also, static files don't work when i have the /*
mapping on
[09:34am] Reuven: would a command like increaseObjectCount consider as
a Setter?
[09:34am] idrism: shouldn't spring have it fallback to the static
files if it can't find a controller to match it?
[09:35am] cylob: im kinda of newbish and just hacking spaghetti code,
so it's quite possible for a dumb error
[09:35am] FoosMasta: <alexrudnick: A buddy of mine (mele/rburton) is
thinking of doing a 101 tyle write-up on GAE
[09:35am] datanucleus: Reuven : yes. just don't update fields directly
is all im saying
[09:35am] Jason_Google: cylob: Hmm. Can you verify this on your end?
Whether removing that particular attribute fixes the issue?
[09:35am] knowtheory: tobyr & the other google guys: Thanks for the
info, i've gotta cut out for now. I'll talk to Headius about
collecting together apps that run on the JVM, and making sure that
we've got a way to tell how well they're running
[09:35am] cylob: Jason_Google: sure, brb
[09:35am] Reuven: I don't where can I see the logs?
[09:36am] Jason_Google: cylob: From my understanding of the attribute,
I don't think that should be the behavior.
[09:36am] schwardo: idrism: for static files, they should always
*override* the /* servlet. That's how the server environment
behaves. However, the local environment is not replicating that
behavior properly right now.
[09:36am] datanucleus: Reuven : don't know, DataNucleus is my side.
Maybe the Google guys can tell you where they send the log to
[09:36am] schwardo: idrism: JSP's shouldn't have issues with this,
though... can you file a bug on the issue tracker with details?
[09:36am] alexrudnick: FoosMasta: Cool! Send it out to the lists when
it's done?
[09:36am] tobyr: knowtheory: thanks
[09:37am] idrism: schwardo: okay maybe its just an issue on the local
environment. I will test it out on the server enviroment and file a
bug if necessary. Where do I file bugs?
[09:37am] knoonan_: (FYI: "Headius" is Charles Nutter, development
lead for JRuby and working at SUN.)
[09:37am] nickjohnson: idrism: http://code.google.com/p/googleappengine/issues/list
[09:37am] Reuven: datanucleus: Can you have a look at:
http://groups.google.com/group/google-appengine-java/browse_thread/thread/d9d65406c4aa9955/1903928b67623a21?lnk=gst&q=reuven#1903928b67623a21
and tell me if I'm doing anything silly in terms of JDO?
[09:37am] idrism: thanks all
[09:38am] datanucleus: Reuven : saw it before. Nothing obvious wrong.
First thing i'd do is go for the log. Alternatively but a transaction
around the update
[09:38am] tobyr: Reuven: you can change the logging level for
datanucleus using your logging.properties file.
[09:39am] cylob: Jason_Google: nope, no effect
[09:39am] Jason_Google: cylob: Is this affecting the local dev server,
production, or both?
[09:39am] tobyr: Reuven: by default, they are set to WARNING, but you
can bump them up to INFO
[09:40am] knoonan_: Googlers, are you concerned that the community
around App Engine (such as it is) will fragment into cabals, each
focused purely on its own dynamic language on the JVM?
[09:40am] Reuven: I have changed them but can't see anything weird,
[09:40am] cylob: Jason_Google: local, havent tested production
[09:41am] tobyr: Reuven: Are you seeing info level logs from
DataNucleus?
[09:41am] datanucleus: Reuven : most intersting stuff is down at DEBUG
level
[09:41am] nickjohnson: knoonan_: I'm personally not concerned. I think
that many groups taking different approaches on different languages
makes the community more interesting. There's a lot in common, too,
since they all share the same APIs
[09:41am] Reuven: tobyr: which should I change, changing all of them
to debug can be a pain
[09:41am] Jason_Google: cylob: OK, can you post a new issue in the
tracker then with your Model code and a short snippet that I can use
to try to reproduce on my end? I'll star it and try to get to it
today.
[09:41am] dan_google: knoonan_: I'm excited by the prospect,
actually. It's more likely to bring new communities to App Engine
than fragment the existing AE community.
[09:42am] cylob: Jason_Google: yeah i will post my entire source if
that helps, it's only 100 lines. where's this tracker?
[09:42am] tobyr: koonan_ The dynamic language folks themselves are
very co-operative and work together. Check out the jvm-languages
google group. (http://groups.google.com/group/jvm-languages)
[09:43am] nickjohnson: cylob: http://code.google.com/p/googleappengine/issues/list
[09:43am] Jason_Google: Ah, Nick beat me to it.
[09:43am] nickjohnson: already had it in my scrollback
[09:43am] Jason_Google: It's also where we commit the SDK source, etc.
[09:43am] tobyr: koonan_ I think we should see the same kind of
attitude from dynamic languages App Engine users
[09:43am] datanucleus: Reuven : Max doesn't seem to be logging a lot
for updates so may be hard to debug currently. I'd just try all at
DEBUG and try to spot your update
[09:44am] knoonan_: Let many flowers bloom! I love seeing Java and
Ruby, etc. on App Engine. There's a danger of a "Tower of Babel"
though--Pythonistas (who don't know Ruby and Java) may be put off by
the talk of servlets and Rails, etc. (especially on these IRC
sessions).
[09:45am] cylob: Jason_Google: ty. http://code.google.com/p/googleappengine/issues/detail?id=1314
[09:46am] Reuven: Can't I have class A with is Persistent that
includes class B and List<C> where C has B in it?
[09:46am] cylob: but it's probably going to be an embarassing newb
mistake on my part
[09:46am] datanucleus: Reuven : are class B and C persistent ?
[09:47am] Reuven: I get an exception that B is defined as One-To-One
and has more then one (due to C)
[09:47am] Reuven: Yes, B and C are persistent
[09:47am] datanucleus: You can have relations like that
[09:47am] Reuven: So what does the error means?
[09:47am] Jason_Google: knoonan_: We're playing it by ear. We already
have runtime-specific groups set up, and we'll pay attention if we
feel part of App Engine's user base feels like it's being disregarded,
although hopefully that won't be the case.
[09:48am] Jason_Google: cylob: Great, I'll look into it today.
[09:48am] Reuven: I only moved C to be Serializable to solve this
issue
[09:48am] Reuven: Sorry B
[09:48am] datanucleus: Depends on the exact error message etc. Would
have to see more
[09:50am] FoosMasta: Do you guys use any Atlassian software
internally?
[09:53am] picalolabu: quick urlfetch question. i believe it can only
open port 80 but my dev server is running on 8080 (my 80 is taken) -
is there an easy workaround or do i have to switch my local ports to
run the devserver on 80?
[09:53am] Reuven: I can't seem to find anything in the logs, is there
any workaround that you think can help me solve the issue at:
http://groups.google.com/group/google-appengine-java/browse_thread/thread/d9d65406c4aa9955
[09:54am] knoonan_: Are you seeing more adoption of App Engine by
large companies? Ones that might not wish to be advertised on
http://code.google.com/appengine/casestudies.html.
[09:55am] datanucleus: Reuven : Are UserVotes Persistence ?
[09:55am] bibac_: will you support maven, e.g. plugins for common gae
tasks, putting the gae dependencies in repos, or are u counting on the
community to come up with this?
[09:55am] scudder_google: picalolabu: The port number restrictions are
for the host you are connecting to (not from). Are you trying to
connect from an app running in App Engine back to your dev server?
[09:55am] jcgregorio: picalolabu: are you trying to fetch from the
devserver back to devserver?
[09:56am] Reuven: no just Sirializable now
[09:56am] datanucleus: So it can't detect any change in them, hence no
update
[09:56am] tobyr: bibac_: Right now, we'd like to see where the
community takes maven support on their own
[09:56am] bibac_: tobyr: fair enough
[09:57am] picalolabu: jcgregorio: yes, for unit testing, i'd like to
have a /unittest/ url that tests some of my other handlers
[09:57am] Reuven: I can't make it Persistent due to the One-to-Many
problem
[09:57am] jcgregorio: knoonan_: obviously we can't comment if they
asked no to appear there
[09:58am] jcgregorio: picalolabu: the python dev_appserver is single
threaded, so that won't work, you will need to start two instances
[09:58am] datanucleus: Reuven : call JDOHelper.makeDirty(item,
"votes"); This marks the field in the item as "dirty" (changed)
[09:58am] picalolabu: scudder_google: yes, i'm trying to call myself
for unit tests
[09:58am] bibac_: jcgregorio: common, it's just us her
[09:58am] knoonan_: @jcgregorio You can't name them, obviously. You
could comment though on adoption. I'd expect you'll see greater
adoption with the launch of Java--and after all the work on task
queues and long-lived processing is in place. Batch processing matters
to large enterprises.
[09:58am] picalolabu: jcgregorio: thx - if i start two instances, do i
still need to switch my local ports so that one of my instances runs
on 80?
[09:59am] jcgregorio: bibac_
[09:59am] FoosMasta: I second knoonan_'s comment
[10:00am] jcgregorio: picalolabu: you will have to change one of them
to serve on a different port, so they don't conflict
[10:01am] Reuven: datanucleus: thanks a lot you saved me waiting for
someone to figure out how to solve this issue, I can continue working
now in the meanwhile ...
[10:01am] picalolabu: jcgregorio: and does one of them HAS to be on
port 80? or is there any other workaround when calling urlfetch?
[10:01am] jcgregorio: picalolabu: but in general the dev_server should
be able to send requests to any URI and that includes one with a port#
specifier, i.e. http://localhost:8080
[10:02am] picalolabu: jcgregorio: ah ... thanks
[10:03am] alexrudnick: So, just curious -- is anybody using the
Eclipse plugin? Any thoughts or feedback or issues with it?
[10:04am] drome: I'm using it
[10:04am] drome: so far so good
[10:04am] drome: maybe a 'remember password' field for deployment
purposes?
[10:04am] bibac_: me too, works as advertised
[10:04am] bibac_: yeah, pls
[10:05am] AsharLohmar: alexrudnick: until now evry thing is smooth
[10:05am] alexrudnick: drome: Cool, duly noted. It's on our list of
things to look at.
[10:05am] Reuven: datanucleus: last question: how do I makeDirty an
inner vote? (Item has List<Tag> and a Tag has UserVote) Just do: Tag
tag = item.getTags().get(0); JDOHelper.makeDirty(tag, "votes");
[10:05am] AsharLohmar: but on thing
[10:05am] AsharLohmar: i run one project
[10:05am] datanucleus: JDOHelper.makeDirty(object, "field");
[10:05am] datanucleus: the object is persistable
[10:05am] AsharLohmar: in hosted mode
[10:06am] AsharLohmar: after i close the 2 windows, and i run another
project from the same workspace
[10:06am] drome: A problem I'm having in eclipse
[10:06am] AsharLohmar: it start with the older one
[10:06am] drome: by default there's no 'test' source directory
[10:07am] AsharLohmar: drome: not anymore
[10:07am] Jason_Google: OK Everyone. Thanks for another great hour.
Chat Time is officially ending, although you're more than welcome to
keep on discussing.
[10:07am] drome: so I create one, but then the pluggin tries to
compile my code as part of the application
[10:07am] Reuven: Thanks a lot, again.
an application and I would like to use GAE-Java for the backend. But
currently SmartGWT imposes me thousands of static files, and GAE only
allows a maximum of 1000 files. How to solve this problem?
[09:05am] drome: (Would this problem be solved, if GAE supported the
upload of a single ".war" file for each application?)
[09:06am] tobyr: RobHanson: The best thing to do is file a feature
request, and we'll see what we can do.
[09:07am] lexs2: batch delete!
[09:07am] RobHanson: tobyr: I am sure there is already a feature
request... just wondering on the reasoning why it wasn't included
initially, since it seems to break a lot of things.
[09:08am] wycats: sorry I'm late
[09:08am] schwardo: fedestabile: that should not be the case, but i
haven't personally tested it. please file a bug on the issue tracker
and we'll follow up
[09:08am] • wycats is with knowtheory
[09:08am] scudder_google: (btw jcgregorio and dan_google are also
Googlers)
[09:08am] • dan_google waves.
[09:08am] alexrudnick: scudder_google: Also alexrudnick!
[09:08am] • jcgregorio waves
[09:08am] nickjohnson: If you're looking for help on how specific APIs
work at the stub level, I and others hang out in here semi-
permanently. And there's the PYthon source.
[09:08am] knowtheory: tobyr: My goal was to try and get apps on JRuby
into an environment that was as much like GAE as possible. We can run
Ruby apps via the Rack standard. ribrdb_ pointed out that future
releases of the GAE dev server may have additional tools available
[09:08am] wycats: following up on what knowtheory said before, it
would be nice to be have documentation for hooking up to the mock
datastore without spinning up a mock server
[09:08am] tobyr: RobHanson: There's not any real context for which the
naming API is useful in our environment. Any use you tried to make of
it would result in a SecurityException later down the line
[09:09am] Jason_Google: fedestabile: Yes, please include your
application ID, and if possible, test the upload from the appspot
domain to see if it's an issue with external domains.
[09:09am] knowtheory: that we wouldn't have in a vanilla Ruby + Rack
compliant server
[09:09am] fedestabile: schwardo: yes, it surprises me too... ok, I'll
post this "issue"
[09:09am] scudder_google: drom: I don't know much about SmartGWT, but
would it be possible to zip up the files into a jar and upload with
your app?
[09:09am] knowtheory: Or alternatively, having an api so that we could
write something that's compliant without having to hook into the dev
server & spin the whole environment up (as wycats pointed out)
[09:10am] knoonan: You have oodles of dynamic languages to track now
(congratulations!)--so in case you missed this: Alan Kennedy got
servlets written in Jython running on App Engine on Modjy ("Jython
WSGI to Java servlet bridge") http://jython.xhaus.com/
[09:10am] tobyr: wycats: There's already documentation on using the
datastore without the dev_appserver - Read the docs on unit testing.
[09:10am] fedestabile: Jason_Google: I've tested the upload from the
appspot and works perfectly
[09:11am] Jason_Google: fedestabile: Good to know. Please include that
in your issue report, and we'll look into it straightaway.
[09:11am] drome: scudder_google, the thing is the static files include
javascript, png, css, html, xml... not all the files are read by
application code. Some of them are just to be read by the bowser
[09:11am] FoosMasta: Are there scenarios that we should avoid with
bigtable from the Java AppEnging APIs? It seems to handle simple Map
structures well, but what about more complex OR maps?
[09:11am] tobyr: drome: You can unzip your files on the fly using a
servlet or filter
[09:11am] fedestabile: Jason_Google: actually, http://www.myclimbs.com/uploadform
or http://myclimbs.appspot.com/uploadform is a very simple page that I
made to show this error
[09:12am] fedestabile: Jason_Google: ok, I'll post it
[09:12am] bthomson: what was recently changed in app engine's tos?
[09:12am] wycats: tobyr: those docs are pretty coupled to JUnit
[09:12am] drome: I'm mentioning SmartGWT, but I imagine there are a
lot of javascript frameworks that will suffer from the 1000 files
limitation
[09:12am] wycats: I couldn't tell what I still needed to do if I
didn't want to use JUnit
[09:12am] cylob: in the datastore how do i update an existing entity,
rather than create a new one?
[09:13am] schwardo: FooMasta: unfortunately our resident ORM expert
(maxr) couldn't make it today, so we can't rattle off a list of
caveats. Your best bet is to just try it and if you run into
problems, post on the forum.
[09:13am] nickjohnson: cylob: Java or Python?
[09:13am] cylob: python
[09:13am] nickjohnson: Just call .put() on an existing entity
[09:13am] nickjohnson: There's no difference between insertion and
update
[09:13am] cgeorg: instance = YourClass.get(key)
[09:13am] schwardo: wycats: the ApiProxy method calls are the
important part -- it should be trivial to port the JUnit test to a
different environment
[09:13am] cgeorg: instance.something = 'foo'
[09:13am] drome: tobyr: sorry, a servlet to return static content?
[09:13am] cgeorg: instance.put()
[09:13am] cylob: ok let me meditate on this for a sec ty
[09:13am] lexs2: nice error
[09:14am] lexs2: "Timeout: datastore timeout: operation took too
long."
[09:14am] scudder_google: drome: ah ok, one way you could work around
the 1000 file limit would be to store the files in the datastore and
write a servlet to handle those requests (would be good to use
memcache/jcache as well)
[09:14am] AsharLohmar: cylob: try the clasic aproach, ge tit using
getby id or something, change its props an the do a .put()
[09:14am] tobyr: wycats: None of that code is Junit specific. You
would need to do everything in the sample code as shown - setting the
ApiProxy delegates, environments, etc...
[09:14am] AsharLohmar: cylob: sorry for the misspelling
[09:14am] wycats: schwardo: trivial if you know Java
[09:14am] tobyr: drome: Yes - You would return static content via a
servlet or filter
[09:14am] scudder_google: drome: tobyr also had a great sugestion,
don't know how I missed that
[09:15am] knoonan: This is easily the busiest App Engine IRC session
I've seen (at the European-friendly timeslot)--breaking the one-
hundred-participant barrier!
[09:15am] dinguva: Googlers: I routinely clone my prod database into a
development db so I can run tests on them. I plan to run 2 different
applications for prod and dev and want to findout if i can clone the
data between two different applications in appengine. Is that
possible?
[09:15am] schwardo: wycats: sounds like a good opportunity to learn
[09:15am] drome: tobyr: ok, do happen to have a pointer for me, to
investigate further? (some examples...)
[09:16am] nickjohnson: dinguva: You'll need to write your own code to
do that. There's currently no built-in support for it.
[09:16am] wycats: schwardo: sure
[09:16am] schwardo: dinguva: you could use two different version
numbers, if you always want the data to be in sync
[09:16am] Jason_Google: bthomson: We usually publish a revision note
when we update the TOS but it doesn't look like this made it in last
week. It's hard to summarize, but I can tell you the sections that
changed if that will help.
[09:16am] schwardo: dinguva: if you wanted to do something more
complex like reload prod -> dev periodically then you would need to
develop a mechanism for syncing the data
[09:16am] nickjohnson: As schwardo points out, you can use the exact
same datastore if you're happy with that
[09:17am] AsharLohmar: on what dinguva said, what about an local
instance of the db ?!
[09:17am] tobyr: drome: sorry, no specific examples - but it's pretty
straightforward. This might be a common enough usecase that we could
provide an example in some docs somewhere
[09:17am] dinguva: schwardo: true. but, if i change the data with a
beta version of application, it'll change the actual prod data too !
So was wondering if I can use a clone of the data for the beta version
for say user acceptance testing.
[09:17am] AsharLohmar: i mean what if i wnat an local "clone" of the
db ?
[09:17am] dinguva: schwardo: basically looking at GAE for the
enterprise
[09:18am] knowtheory: I've heard talk that the JRuby guys have been in
contact with you, regarding the fact that GAE is running the Client
JVM?
[09:18am] nickjohnson: AsharLohmar: There are tools like gae-bar for
uploading and downloading backups
[09:18am] serickso: I've got a question. We've been having issues with
502 error pages made from time to time on our application (id: skrit).
They've been tough to track down; they seem to happen on all kinds of
requests and only on our external domain, www.skritter.com. Nothing
about them shows up in our logs. Is there anything you guys could tell
us about 502 errors that would help us handle them or figure out
what's causing them?
[09:18am] tobyr: knowtheory: yes we've been discussing it with them
[09:18am] knowtheory: And that you guys would potentially consider a
switch to the server JVM if you could compile some compelling evidence
[09:18am] knowtheory: what would you like to see in an app for testing
purposes?
[09:18am] tobyr: knowtheory: there are many, many variables involved
[09:19am] knowtheory: Yes that's why i'm asking
[09:19am] knowtheory: beacause i'll happily put one together
[09:19am] drome: tobyr: ok, I'll post into the discussion board to get
this documented somehow...
[09:19am] AsharLohmar: nickjohnson: thanks
[09:19am] knowtheory: but i'm not entirely sure what to aim at.
[09:20am] knowtheory: i can build a stateless app that does some stuff
[09:20am] schwardo: knowtheory: i don't know how useful a test app for
that would be... we did extensive load tests across a wide range of
apps to arrive at our current config
[09:20am] knowtheory: or i can find a merb app that already exists out
there in the wild
[09:20am] dinguva: schwardo: just saw your reply about prod -> dev..
If an app can reach out and access another app's data that'd be
great ! We can even have a helper application that does this
[09:20am] Reuven: guys, when is the next release with maintrunk
changes for java? it seems from the posts that there are some very
important stuff over there
[09:21am] knowtheory: Okay, well, let me be more general then, and
ask, is there any information that i could help to provide regarding
the performance of ruby applications on the JVM?
[09:21am] knowtheory: Are you interested in any particular info?
[09:21am] scudder_google: serickso: could you tell us a bit more about
the 502s, roughly what percentage of requests they occur in. I'll look
into specific issues with that app.
[09:21am] tobyr: knowtheory: I do think having some simple vs complex
jruby apps could be useful, particularly if they have "benchmarking"
built in
[09:21am] bthomson: Jason_Google: if you know which sections that
would be great, I'd rather not read the whole thing again
[09:22am] cgeorg: Erin and her kitty:
http://icanhascheezburger.files.wordpress.com/2009/03/funny-pictures-cat-and-human-argue.jpg
[09:22am] Jason_Google: Reuven: schwardo might know better than me.
Hopefully in less than a few weeks' time.
[09:22am] Jason_Google: (being purposefully vague )
[09:22am] cgeorg: wow, wrong channel
[09:22am] • cgeorg hides
[09:22am] schwardo: Reuven: i can't give you a specific date, but
there are definitely some good fixes queued up that we want to get out
[09:22am] nickjohnson: cgeorg: Cat pictures are always welcome
[09:23am] cgeorg: Who do I pay to scrub that out of the official
transcript?
[09:23am] schwardo: lol
[09:24am] Jason_Google: bthomson: As far as I can tell, sections 7.4,
10.3, 10.4, 10.5 were the only ones that changed.
[09:24am] FoosMasta: Is there any documentation or source code
available for the Java GAE Eclipse plug-in?
[09:24am] Jason_Google: And I'll make sure any future changes are
summarized in the revision history.
[09:24am] Jason_Google: cgeorg: That would be me.
[09:24am] alexrudnick: FoosMasta: There's documentation!
http://code.google.com/eclipse/docs/users_guide.html
[09:24am] fedestabile: Jason_Google: I've posted the issue, it's the
number 1312. Thanks!
[09:25am] Jason_Google: Thanks fedestabile!
[09:25am] Jason_Google: I'll star it.
[09:25am] fedestabile: great!!
[09:25am] alexrudnick: FoosMasta: It's not open-source yet. But we're
going to open source sooner or later. (hopefully sooner)
[09:25am] bthomson: thanks Jason_Google
[09:25am] RobHanson: is there a reason why you can't delete an app?
will you be able to in the future?
[09:26am] dinguva: RobHanson: delete app ++
[09:26am] tobyr: RobHanson: We're aware of it, and we're taking into
consideration that many people are asking for it.
[09:26am] cylob: script = Script.get(self.request.get('dakey'));
script.datitle = self.request.get('datitle'); script.put(); this
is still creating a new entity instead of updating
[09:27am] serickso: scudder_google: they happen infrequently, and in
clusters. I think I've seen it as a page load once or twice, and I'm
developing the site. We have a flash client though that makes many
requests to the server through PyAMF and it happens enough so it gums
up communication between the client and the server from time to time.
Nick knows more about it than me, and he wrote about it here:
http://groups.google.com/group/google-
[09:27am] bthomson: RobHanson: you can use the old app for something
else, just map it to a new domain name
[09:28am] alexrudnick: FoosMasta: Have you tried out the plugin? Any
feedback you'd like to share?
[09:28am] dinguva: bthomson: basically recycle and be friendly to the
environment
[09:28am] knowtheory: A philosophical question, why do you guys
develop and then opensource after the fact?
[09:28am] knowtheory: (That's not a troll, i'm just curious about the
reasoning)
[09:28am] RobHanson: <bthomson> yeah, I am aware of that. but
sometimes you just want it to go away.... or rename it to something
that has meaning
[09:28am] scudder_google: serickso: Thanks for the details, I'll look
into it further after we end here.
[09:29am] nickjohnson: knowtheory: I can't speak authoritatively, but
I can speculate that it's largely because open-sourcing takes non-zero
work, so we'd rather have it out there sooner and follow up with the
source, than delay everything so we can release with the source.
[09:29am] knowtheory: (rather than develop out in the open that is)
[09:29am] knowtheory: ah okay, that's interesting
[09:29am] Jason_Google: knowtheory: It was my fault last week for
publishing the source code after the release. I'll make sure that it
happens same-day for future releases.
[09:30am] serickso: scudder_google: Thanks a bunch
[09:30am] FoosMasta: alexrudnick: yea - got stuck on the 400 Bad
Request - didn't know i needed to enable java for my account in
addition to signing up
[09:30am] knowtheory: Well, grateful in either regard, since i have...
plans And i'm glad to hear you guys are interested in keeping
everyone up to speed
[09:31am] alexrudnick: FoosMasta: Oooh. So maybe the plugin could tell
you that there was a problem, in that case?
[09:31am] FoosMasta: exactly
[09:31am] alexrudnick: FoosMasta: This was when you were trying to
deploy?
[09:31am] FoosMasta: yup
[09:31am] Jason_Google: cylob: That's an interesting issue. If get
returns a valid entity, calling put to that object should update the
entity instead of creating a new one. Do you just see a new entity
with every other field the same except for the one that you modified?
[09:31am] alexrudnick: FoosMasta: That's a good idea. Writing it down
[09:32am] idrism: hey, anyone know about having a servlet with /* as
the url-mapping?
[09:32am] Reuven: Anyone experienced a problem with changing a
Serializable object using getObjectById and changing the Serializable
object state, then closing the pm and nothing happens ...
[09:32am] schwardo: idrism: sure, what about it?
[09:32am] idrism: that screws up JSP views fo rme...
[09:33am] drome: Are you planning to support upload of a single '.war'
file that holds the entire application?
[09:33am] datanucleus: Reuven : have you looked in the log ?
[09:33am] datanucleus: (at debug level)
[09:33am] • knoonan_ got randomly disconnected from the server (just
got back)
[09:33am] cylob: Jason_Google: yep, that's what i see
[09:33am] Reuven: what do you mean, I can change other objects and it
saves them but not the Serializable one
[09:33am] schwardo: idrism: in the dev environment or the server
environment (or both) ?
[09:33am] scudder_google: idrism: I would think that a /* mapping
would match everything. What does your directory structure look like?
Are you trying to send all requests to one JSP page?
[09:34am] idrism: for example if i have a controller that uses
"test.jsp", and i hit that controller, it will say 404 for "/WEB-INF/
war/test.jsp"
[09:34am] datanucleus: Reuven : JDO detects changes to objects if you
do via setters. The DataNucleus log logs all changes and state changes
[09:34am] cylob: maybe it's affected by using the required=TRUE in the
dbmodel?
[09:34am] idrism: i think internally it is trying to go through that
same servlet mapping
[09:34am] idrism: also, static files don't work when i have the /*
mapping on
[09:34am] Reuven: would a command like increaseObjectCount consider as
a Setter?
[09:34am] idrism: shouldn't spring have it fallback to the static
files if it can't find a controller to match it?
[09:35am] cylob: im kinda of newbish and just hacking spaghetti code,
so it's quite possible for a dumb error
[09:35am] FoosMasta: <alexrudnick: A buddy of mine (mele/rburton) is
thinking of doing a 101 tyle write-up on GAE
[09:35am] datanucleus: Reuven : yes. just don't update fields directly
is all im saying
[09:35am] Jason_Google: cylob: Hmm. Can you verify this on your end?
Whether removing that particular attribute fixes the issue?
[09:35am] knowtheory: tobyr & the other google guys: Thanks for the
info, i've gotta cut out for now. I'll talk to Headius about
collecting together apps that run on the JVM, and making sure that
we've got a way to tell how well they're running
[09:35am] cylob: Jason_Google: sure, brb
[09:35am] Reuven: I don't where can I see the logs?
[09:36am] Jason_Google: cylob: From my understanding of the attribute,
I don't think that should be the behavior.
[09:36am] schwardo: idrism: for static files, they should always
*override* the /* servlet. That's how the server environment
behaves. However, the local environment is not replicating that
behavior properly right now.
[09:36am] datanucleus: Reuven : don't know, DataNucleus is my side.
Maybe the Google guys can tell you where they send the log to
[09:36am] schwardo: idrism: JSP's shouldn't have issues with this,
though... can you file a bug on the issue tracker with details?
[09:36am] alexrudnick: FoosMasta: Cool! Send it out to the lists when
it's done?
[09:36am] tobyr: knowtheory: thanks
[09:37am] idrism: schwardo: okay maybe its just an issue on the local
environment. I will test it out on the server enviroment and file a
bug if necessary. Where do I file bugs?
[09:37am] knoonan_: (FYI: "Headius" is Charles Nutter, development
lead for JRuby and working at SUN.)
[09:37am] nickjohnson: idrism: http://code.google.com/p/googleappengine/issues/list
[09:37am] Reuven: datanucleus: Can you have a look at:
http://groups.google.com/group/google-appengine-java/browse_thread/thread/d9d65406c4aa9955/1903928b67623a21?lnk=gst&q=reuven#1903928b67623a21
and tell me if I'm doing anything silly in terms of JDO?
[09:37am] idrism: thanks all
[09:38am] datanucleus: Reuven : saw it before. Nothing obvious wrong.
First thing i'd do is go for the log. Alternatively but a transaction
around the update
[09:38am] tobyr: Reuven: you can change the logging level for
datanucleus using your logging.properties file.
[09:39am] cylob: Jason_Google: nope, no effect
[09:39am] Jason_Google: cylob: Is this affecting the local dev server,
production, or both?
[09:39am] tobyr: Reuven: by default, they are set to WARNING, but you
can bump them up to INFO
[09:40am] knoonan_: Googlers, are you concerned that the community
around App Engine (such as it is) will fragment into cabals, each
focused purely on its own dynamic language on the JVM?
[09:40am] Reuven: I have changed them but can't see anything weird,
[09:40am] cylob: Jason_Google: local, havent tested production
[09:41am] tobyr: Reuven: Are you seeing info level logs from
DataNucleus?
[09:41am] datanucleus: Reuven : most intersting stuff is down at DEBUG
level
[09:41am] nickjohnson: knoonan_: I'm personally not concerned. I think
that many groups taking different approaches on different languages
makes the community more interesting. There's a lot in common, too,
since they all share the same APIs
[09:41am] Reuven: tobyr: which should I change, changing all of them
to debug can be a pain
[09:41am] Jason_Google: cylob: OK, can you post a new issue in the
tracker then with your Model code and a short snippet that I can use
to try to reproduce on my end? I'll star it and try to get to it
today.
[09:41am] dan_google: knoonan_: I'm excited by the prospect,
actually. It's more likely to bring new communities to App Engine
than fragment the existing AE community.
[09:42am] cylob: Jason_Google: yeah i will post my entire source if
that helps, it's only 100 lines. where's this tracker?
[09:42am] tobyr: koonan_ The dynamic language folks themselves are
very co-operative and work together. Check out the jvm-languages
google group. (http://groups.google.com/group/jvm-languages)
[09:43am] nickjohnson: cylob: http://code.google.com/p/googleappengine/issues/list
[09:43am] Jason_Google: Ah, Nick beat me to it.
[09:43am] nickjohnson: already had it in my scrollback
[09:43am] Jason_Google: It's also where we commit the SDK source, etc.
[09:43am] tobyr: koonan_ I think we should see the same kind of
attitude from dynamic languages App Engine users
[09:43am] datanucleus: Reuven : Max doesn't seem to be logging a lot
for updates so may be hard to debug currently. I'd just try all at
DEBUG and try to spot your update
[09:44am] knoonan_: Let many flowers bloom! I love seeing Java and
Ruby, etc. on App Engine. There's a danger of a "Tower of Babel"
though--Pythonistas (who don't know Ruby and Java) may be put off by
the talk of servlets and Rails, etc. (especially on these IRC
sessions).
[09:45am] cylob: Jason_Google: ty. http://code.google.com/p/googleappengine/issues/detail?id=1314
[09:46am] Reuven: Can't I have class A with is Persistent that
includes class B and List<C> where C has B in it?
[09:46am] cylob: but it's probably going to be an embarassing newb
mistake on my part
[09:46am] datanucleus: Reuven : are class B and C persistent ?
[09:47am] Reuven: I get an exception that B is defined as One-To-One
and has more then one (due to C)
[09:47am] Reuven: Yes, B and C are persistent
[09:47am] datanucleus: You can have relations like that
[09:47am] Reuven: So what does the error means?
[09:47am] Jason_Google: knoonan_: We're playing it by ear. We already
have runtime-specific groups set up, and we'll pay attention if we
feel part of App Engine's user base feels like it's being disregarded,
although hopefully that won't be the case.
[09:48am] Jason_Google: cylob: Great, I'll look into it today.
[09:48am] Reuven: I only moved C to be Serializable to solve this
issue
[09:48am] Reuven: Sorry B
[09:48am] datanucleus: Depends on the exact error message etc. Would
have to see more
[09:50am] FoosMasta: Do you guys use any Atlassian software
internally?
[09:53am] picalolabu: quick urlfetch question. i believe it can only
open port 80 but my dev server is running on 8080 (my 80 is taken) -
is there an easy workaround or do i have to switch my local ports to
run the devserver on 80?
[09:53am] Reuven: I can't seem to find anything in the logs, is there
any workaround that you think can help me solve the issue at:
http://groups.google.com/group/google-appengine-java/browse_thread/thread/d9d65406c4aa9955
[09:54am] knoonan_: Are you seeing more adoption of App Engine by
large companies? Ones that might not wish to be advertised on
http://code.google.com/appengine/casestudies.html.
[09:55am] datanucleus: Reuven : Are UserVotes Persistence ?
[09:55am] bibac_: will you support maven, e.g. plugins for common gae
tasks, putting the gae dependencies in repos, or are u counting on the
community to come up with this?
[09:55am] scudder_google: picalolabu: The port number restrictions are
for the host you are connecting to (not from). Are you trying to
connect from an app running in App Engine back to your dev server?
[09:55am] jcgregorio: picalolabu: are you trying to fetch from the
devserver back to devserver?
[09:56am] Reuven: no just Sirializable now
[09:56am] datanucleus: So it can't detect any change in them, hence no
update
[09:56am] tobyr: bibac_: Right now, we'd like to see where the
community takes maven support on their own
[09:56am] bibac_: tobyr: fair enough
[09:57am] picalolabu: jcgregorio: yes, for unit testing, i'd like to
have a /unittest/ url that tests some of my other handlers
[09:57am] Reuven: I can't make it Persistent due to the One-to-Many
problem
[09:57am] jcgregorio: knoonan_: obviously we can't comment if they
asked no to appear there
[09:58am] jcgregorio: picalolabu: the python dev_appserver is single
threaded, so that won't work, you will need to start two instances
[09:58am] datanucleus: Reuven : call JDOHelper.makeDirty(item,
"votes"); This marks the field in the item as "dirty" (changed)
[09:58am] picalolabu: scudder_google: yes, i'm trying to call myself
for unit tests
[09:58am] bibac_: jcgregorio: common, it's just us her
[09:58am] knoonan_: @jcgregorio You can't name them, obviously. You
could comment though on adoption. I'd expect you'll see greater
adoption with the launch of Java--and after all the work on task
queues and long-lived processing is in place. Batch processing matters
to large enterprises.
[09:58am] picalolabu: jcgregorio: thx - if i start two instances, do i
still need to switch my local ports so that one of my instances runs
on 80?
[09:59am] jcgregorio: bibac_
[09:59am] FoosMasta: I second knoonan_'s comment
[10:00am] jcgregorio: picalolabu: you will have to change one of them
to serve on a different port, so they don't conflict
[10:01am] Reuven: datanucleus: thanks a lot you saved me waiting for
someone to figure out how to solve this issue, I can continue working
now in the meanwhile ...
[10:01am] picalolabu: jcgregorio: and does one of them HAS to be on
port 80? or is there any other workaround when calling urlfetch?
[10:01am] jcgregorio: picalolabu: but in general the dev_server should
be able to send requests to any URI and that includes one with a port#
specifier, i.e. http://localhost:8080
[10:02am] picalolabu: jcgregorio: ah ... thanks
[10:03am] alexrudnick: So, just curious -- is anybody using the
Eclipse plugin? Any thoughts or feedback or issues with it?
[10:04am] drome: I'm using it
[10:04am] drome: so far so good
[10:04am] drome: maybe a 'remember password' field for deployment
purposes?
[10:04am] bibac_: me too, works as advertised
[10:04am] bibac_: yeah, pls
[10:05am] AsharLohmar: alexrudnick: until now evry thing is smooth
[10:05am] alexrudnick: drome: Cool, duly noted. It's on our list of
things to look at.
[10:05am] Reuven: datanucleus: last question: how do I makeDirty an
inner vote? (Item has List<Tag> and a Tag has UserVote) Just do: Tag
tag = item.getTags().get(0); JDOHelper.makeDirty(tag, "votes");
[10:05am] AsharLohmar: but on thing
[10:05am] AsharLohmar: i run one project
[10:05am] datanucleus: JDOHelper.makeDirty(object, "field");
[10:05am] datanucleus: the object is persistable
[10:05am] AsharLohmar: in hosted mode
[10:06am] AsharLohmar: after i close the 2 windows, and i run another
project from the same workspace
[10:06am] drome: A problem I'm having in eclipse
[10:06am] AsharLohmar: it start with the older one
[10:06am] drome: by default there's no 'test' source directory
[10:07am] AsharLohmar: drome: not anymore
[10:07am] Jason_Google: OK Everyone. Thanks for another great hour.
Chat Time is officially ending, although you're more than welcome to
keep on discussing.
[10:07am] drome: so I create one, but then the pluggin tries to
compile my code as part of the application
[10:07am] Reuven: Thanks a lot, again.
Reply all
Reply to author
Forward
0 new messages