Ryan,
Tasks are repeated until they return a HTTP success code. So your task
need to be indepotent in that if they should still do the right thing
(not write things twice etc.) if your task gets run more than once
because of failure.
So if one task in the chain has an error, then that task will b
Ben,
With non-relational databases like appengine's datastore you need to
create data based on the way you will query it rather than the way you
want to store it.
In your case it looks like you would need to create a model to store
the data you want to query (essentially an index that you manage)
Hi,
I'm seeing a lot of errors in one of my applications when making
requests to the email RPC.
I am getting DeadlineExceededErrors being throw about half of the time
the api is called. The full error
is as follows.
...
File "/base/python_runtime/python_lib/versions/1/google/appengine/api/mail.p
Pamela,
Specifically Takashi is probably asking if you have any naked except blocks:
try:
# Do something
except:
# Do a dance
People often do stuff like:
try:
import json
except:
import simplejson
If you have some of these, then you could be catching
DeadlineExceededErrors tha
Ben,
It looks like filter will add a filter to the existing query object (instead
of creating a new query object like django does) but the fetch method will
retrieve the objects and return them immediately without modifying the
query. You should do something like:
items_list = items.fetch(10)
an
Andy,
I believe the project's goal was to support django 1.0 on appengine
but since 1.0 is now available there isn't much point to continuing
the project. I think that the appengine folks plan to have django
available on appengine without using google-app-engine-django
Ian
On Wed, Aug 12, 2009
The documentation says that an entity group could be as large as a single
user's data. It seems to me that to avoid the problem you are describing
that it would be a good idea to do so. I have a feeling that the number of
cases where you would need to update more than one user's data in a single
tr
Dennis,
The development appserver appears to load the local datastore file on
startup so you will need to restart the shell to see updates made in the
admin.
Ian
2009/1/27 Dennis
>
> Actually, it's the interaction between the "manage.py shell" and the
> Development Console's "Datastore Viewer"
konryd,
What kind of internal error do you get?
2009/1/27 konryd
>
> Almost every time I deploy a new version of my app and try to access
> it using http://[version].latest.myappnamae.appspot.com I get internal
> error, unless I make it default.
> >
>
--~--~-~--~~~-
The development server does not act on the datastore file like a database
and you won't see the data updated across multiple running instances of the
devserver. The datastore file is just a set of pickled objects and is loaded
into memory in it's entirety when the devserver is started.
2009/1/29 3
Have you tried using pdb? Simon Willison talked about it in his blog a long
time ago. I use it off and on and it's pretty adequate.
http://simonwillison.net/2008/May/22/debugging/
2009/1/30 Youngfe
>
> As I see, pyscripter is great for developing python app!
>
> On Jan 27, 4:52 am, dartdog wro
cw,
What's the directory where models.py with your Quiz model in it? More
accurately what is the python module with your Quiz model in it?
2009/2/3 cdub
>
> I have my application up and running with the Google App Engine Helper
> for Django.
>
> But when I went to add a simple model:
>
> models
You likely need to wait a day or so to have the quotas reset themselves. It
looks like you using too much of one of the appengine quotas.
2009/2/3 Ilia Lobsanov
>
> Correction, it's datastore_v3.Get() not Delete()
>
> The exact traceback is:
>
> Traceback (most recent call last):
> File "/base/
Yup,
The appengine_django/models.py has a funny little piece of code that
requires your models module to not be a top level module.
class ModelOptions(object):
...
def __init__(self, cls):
self.object_name = cls.__name__
self.module_name = self.object_name.lower()
model_module = s
I'm very sure that a C compiler will not be available any time in the near
future, (if ever) on appengine.
2009/2/4 minor project
> Oh.We are trying to develop a C IDE on the cloud using the
> Appengine.Apparently our program demands a gcc compiler to be a part of the
> server so as to compile t
The you are simply using too much of a quota on appengine. You will need to
check the admin interface at www.appspot.com to see what quotas you are
going over.
2009/2/4 Ilia Lobsanov
>
> What I'd like to know is what I'm doing wrong here...
>
> On Feb 3, 4:17 am, Ian
You will probably need to convert the data for each table into CSV files and
use the bulkload_client.py script that is present in the root directory of
the google appengine SDK. You can look up online how to use it or run
"python bulkload_client.py" to see the options.
Ian
2009/2/4 S.I
>
> Hi,
BTW, there is a server side component to the bulk uploader so you will need
to look up how to set that up.
Ian
2009/2/4 Ian Lewis
> You will probably need to convert the data for each table into CSV files
> and use the bulkload_client.py script that is present in the root directory
&g
Hi all,
The documentation seemed to suggest that the more root entities you have the
more it can distribute your query so it should be faster, but I'm curious
about entity groups and what, if any, advantage there is to putting entities
in an entity group besides being able to use them in a transac
Michael,
Hmm, it seems to be running ok here for me.
On Tue, Feb 10, 2009 at 7:00 PM, mcobrien wrote:
>
> Hi,
>
> now that sdk 1.1.9 disallows access to "skipped" files, the
> LoadAppengineEnvironment() function in the django helper isn't
> working. It makes a call to LoadAppConfig, which tries
Michael,
Ah, I may not have been running the current version from svn.
Matt,
Thanks for fixing it,
On Tue, Feb 10, 2009 at 7:35 PM, Matt Brown wrote:
>
> Hi Michael,
>
> mcobrien wrote:
> > now that sdk 1.1.9 disallows access to "skipped" files, the
> > LoadAppengineEnvironment() function in
I'm getting an error similar to the errors about not being able to access
skipped files with appengine-django but this time I'm getting an error about
not being able to access the PIL module file Image.py
I'm getting a different error pertaining to PIL now. This may or may not be
related to appeng
profile_image.put()
if old_profile_image:
old_profile_image.delete()
...
Ian
On Wed, Feb 11, 2009 at 3:52 AM, Matthew Blain wrote:
>
> Hello Ian,
> How are you accessing PIL? Are you using the Images API* , or are you
> importing from PIL directly?
>
> --Mat
Matthew,
I also wanted like to mention that this happens to me on every request which
is why I suspected appengine-django. I don't actually import the image api
unless I am saving a profile image.
On Wed, Feb 11, 2009 at 11:34 AM, Ian Lewis wrote:
> Matthew,
>
> I'm using
I found that this seems unrelated to the message about the PIL module and
was related to the local datastore I was using.
On Wed, Feb 11, 2009 at 12:36 PM, Ian Lewis wrote:
> Matthew,
>
> I also wanted like to mention that this happens to me on every request
> which is why I suspect
Is there any way I can import and run pdb locally? Currently I have
problems with importing pdb. I do a import pdb;pdb.set_trace() and I get the
following error:
DEBUG2009-02-11 10:10:15,174 dev_appserver.py] Could not import
"readline": Disallowed C-extension or built-in module
DEBUG200
Should have searched my mail first:
http://groups.google.com/group/google-appengine/browse_thread/thread/4f4f9626fb22d453/b693677ad779fa64?hl=en&lnk=gst&q=pdb#b693677ad779fa64
On Wed, Feb 11, 2009 at 7:12 PM, Ian Lewis wrote:
> Is there any way I can import and run pdb locally?
I'm guessing the gmail shared namespace limitation has to do with being able
to the forthcoming release that allows apps to recieve mail (presumably at
@gmail.com). This is probably something they thought about in the
beginning but just didn't implement.
On Sun, Feb 15, 2009 at 7:56 AM, Alexander
I think your get_key method needs a return in there.
class DerefModel(db.Model):
def get_key(self, prop_name):
getattr(self.__class__, prop_name).get_value_for_datastore(self)
On Fri, Feb 27, 2009 at 3:51 PM, Bill wrote:
>
> Alex,
> Do you mean something like this:
>
> http://appe
Satish,
Did you enter your country code correctly? How did you enter your phone
number?
Ian
On Sat, Feb 28, 2009 at 9:09 PM, satishbudholiya
wrote:
>
> I tried so many times but haven't received any SMS for verification of
> Google App Engine.
> Please help me.
>
> Satish Budholiya
>
> >
>
--
S,
Did you download it from here?
http://code.google.com/appengine/downloads.html
Ian
On Sun, Mar 1, 2009 at 1:30 PM, saurabhshe...@gmail.com <
inthedeepsile...@gmail.com> wrote:
>
>
> the google app engine given on site is not working both on linux and
> windows
>
> from where we can download
Aurimas,
Pretty slick. How did you create the pdfs? I'd like to make a copy of the
Japanese documentation in the same way.
Ian
On Sat, Feb 28, 2009 at 5:57 PM, Aurimas wrote:
>
> 1. Intro - 37 pages
> http://aurimas.rimsa.name/App-Engine-Docs/1-Intro.pdf
> 2. Python - 34 pages
> http://aurimas
Krishna,
Anything you see in the data viewer is created by some application code. It
wouldn't be put there by google appengine specifically. That said it looks
like something is using a sharded counter to keep counts of objects in the
datastore or something similar. This data would not likely be r
datastore". Best not to meddle although I was hoping to clean
> things up if I could.
>
> Thank you for responding.
> Krishna
>
> On Mar 1, 11:31 pm, Ian Lewis wrote:
> > Krishna,
> >
> > Anything you see in the data viewer is created by some application code
I was learning.
> But the next version of Bloog will likely use a single integer
> property for the counter and scrap the sharded counter.
> -Bill
>
> On Mar 1, 9:11 pm, Ian Lewis wrote:
> > Krishna,
> >
> > The datastore can only return or count up to 1000
Aha, I was hoping you had worked up an html scraper or something to build
the pdfs.
On Mon, Mar 2, 2009 at 5:41 PM, Aurimas wrote:
>
> Nothing special here, just an old school copy/paste.
>
> On Mar 1, 4:11 pm, Ian Lewis wrote:
> > Aurimas,
> >
> > Pretty slic
I have a model set up with a LinkProperty which is not required however when
trying to save the entity with a value of None I get a BadValue error saying
the Link cannot be empty. Is this by design and/or documented somewhere?
...
class Profile(db.Model):
...
homepage = db.LinkProperty(verbose
quot; marker like "http://nolink"; or "not specified" depending on
> validation of the property, but I find this less appealing.
>
> -Bill
>
> On Mar 2, 4:22 am, Ian Lewis wrote:
> > I have a model set up with a LinkProperty which is not required however
> w
starting an issue:
>
> http://code.google.com/p/googleappengine/issues/detail?id=1117
>
> Star it if you'd like the default behavior to change.
> -Bill
>
> On Mar 2, 8:42 pm, Ian Lewis wrote:
> > Bill,
> >
> > Agreed.
> >
> > But if it's by design then
Are you doing an event.put() to save each event?
On Wed, Mar 4, 2009 at 1:32 PM, disorderdev wrote:
>
> Hi, I'm a newbie here, both python and app engine. a few questions:
> I store some fields in DB, but when show message on Web, I need more
> fields, most of them are calculated according to f
Gustavo,
google.appengine.ext contains a number of things not the least of which is
the datastore api. You will probably have the most success searching the
group archives.
Ian
On Thu, Mar 19, 2009 at 8:18 AM, Lord Gustavo Miguel Angel <
goosfanc...@gmail.com> wrote:
> Hi,
>
> I need informati
David,
Please see this issue,
http://code.google.com/p/googleappengine/issues/detail?id=901&q=order%20by&colspec=ID%20Type%20Status%20Priority%20Stars%20Owner%20Summary%20Log%20Component
On Mon, Mar 23, 2009 at 10:38 AM, David I. Murray wrote:
>
> Consider these 2 queries:
>
> Query 1:
> SELEC
Tom,
You can't use the remote api from appengine. It's for use in accessing
the data from applications outside appengine.
AFAICT, trying to connect multiple apps to a single datasource to get
around resource limitations is a violation of google's terms of
service. You might want to take a look a
Robbie,
Please check out the documentation. The java applications are deployed
as a java servlet container (war file et. al.)
http://code.google.com/appengine/docs/java/gettingstarted/introduction.html
GAE is server side so it has nothing to do with GWT or client side
applications but the new G
Ben,
The short answer is no.
The long answer is no, MySQL doesn't scale automatically so GAE only
supports the datastore which is based on Google's scalable Big Table
implementation.
Ian
On Wed, Apr 8, 2009 at 1:26 PM, Ben wrote:
>
> Can we run a mysql database in the engine?
>
> >
>
--
=
This is also not currently working on deployed appengine.
On Fri, May 1, 2009 at 5:44 AM, Waldemar Kornewald wrote:
>
> On Apr 30, 10:27 pm, dalenewman wrote:
> > Looks like the java community already has this search business all
> > figured out :-)
> >
> > http://www.kimchy.org/searchable-googl
You will need to provide a way for appengine to email to your e-mail account
if you want to be able to have it send e-mail to you. This is done by
providing the information about your SMTP server in the settings.py or by
passing the --enable_sendmail or --smtp_host, --smtp_port, --smtp_user, and
--
Nick,
Are you using something like appengine-patch or appengine_django to write
your application?
On Sat, May 30, 2009 at 3:49 AM, Nick Johnson (Google) <
nick.john...@google.com> wrote:
>
> Hi roderic,
>
> You can't use a an App Engine datastore ReferenceProperty to refer to
> a Django model in
Sorry, I meant roderic,
2009/5/30 Ian Lewis
> Nick,
>
> Are you using something like appengine-patch or appengine_django to write
> your application?
>
>
> On Sat, May 30, 2009 at 3:49 AM, Nick Johnson (Google) <
> nick.john...@google.com> wrote:
>
>>
>
Is there a particular reason you need to do something like this?
ReferenceProperty should work whether you use an integer or key_name as the
key to your entities.
On Mon, Jun 8, 2009 at 9:09 AM, Tom Wu wrote:
> Thanks Chris.
>
> This is what I need. You the Man.
>
> Have this run in GAE instead
did something like self.request.FILES["myfile"].name not work?
Since uploading files is a fairly common use case, would it be a good idea
to get this added to the documentation?
On Mon, Jun 8, 2009 at 3:59 AM, Todd wrote:
>
> It took me hours to discover this, and I wanted to share, since I
> c
Just a related thought, this kind of thing should be a non-issue once the
background processes api is released which should be soon. Long running
processes should probably be migrated to that once it's released rather than
using cron since you're not limited to how often cron runs.
On Mon, Jun 8,
Greg,
Please star this issue.
http://code.google.com/p/googleappengine/issues/detail?id=1116
One person noted in the comments that you could cache a system wide value
when the first put request fails so that you don't have to do a test put()
on every request to your application.
It would be inte
Anders,
For ajax calls you might consider creating a very simple handler that uses
as few modules as possible when performing ajax calls. Then put that as a
separate url in your app.yaml. In that case appengine would only load the
modules it needs to serve that request and would hopefully not take
AFAIK you could just do this instead of doing an eval and setting the
properties using setattr:
m = p_model(**kwargs)
where kwargs is the dictionary of properties that you are using to create
your model instance.
On Mon, Jun 8, 2009 at 2:33 PM, Ethan Post wrote:
> I figured this out, so no nee
Issac,
You cannot include a space after the | for filters. Try {{ event.code|escape
}}.
BTW: django does auto-escaping and you shouldn't need the escape filter
unless you have turned auto-escaping off. Turning auto-escaping off is
something I personally wouldn't recommend.
On Mon, Jun 8, 2009 at
And restore to original app, then you will find out the ReferenceProperty(
> db.key) will be a problem.
>
>
> Best Regards
> Tom Wu
>
>
>
> 2009/6/8 Ian Lewis
>
>> Is there a particular reason you need to do something like this?
>> ReferenceProperty should
Pranav,
The only way I can think of is to use the date/time in your key names in
some way. That would allow you to save the same items in a completely time
separate space. You could also use an indexed property on the entity to
allow you to filter by time.
On Tue, Jun 9, 2009 at 11:13 AM, Pranny
What kind of exceptions are you getting?
On Tue, Jun 9, 2009 at 7:34 PM, capoista wrote:
>
> I am trying to set up a simple 'hello world' test of pyamf working on
> the app-engine-patch but my simple gateway.py file causes major
> exceptions.
>
> Does anyone know if it is possible to use pyamf o
ogterran,
Unfortunately the SearchableModel indexes all StringProperties and
TextProperties in the model. You will need to create another model that
contains only the items you want to index. Something like:
class Product(db.Model):
pid = db.StringProperty(required=True)
title = db.
While memcache data is volitile and there is no guarantee that the data that
you put into the cache will be there, memcache shouldn't be giving you
different values for different requests. Are you sure the key is the same
and that it's not being set somewhere else in your code?
On Fri, Jun 19, 200
It's a limit of the datastore that you cannot have multiple inequality
filters spanning multiple properties. You'll simply need to structure your
data so that it is easier to retrieve.
On Fri, Jun 19, 2009 at 2:58 AM, mclovin wrote:
>
> Hey,
>
> I have entities stored with x and y values. What I
ogterran,
It should do one for the search and then one for each item in the search
result. If you are worried performance on the calls to the datastore you can
modify this code to make the ProductSearchIndex entity be a child of the
Product entity and use a key only query to retrieve only the key
Nick,
But he is also doing v.product which will do a get() by key for each Product
entity will it not?
On Tue, Jun 23, 2009 at 7:14 PM, Nick Johnson (Google) <
nick.john...@google.com> wrote:
> Hi ogterran,
>
> On Tue, Jun 23, 2009 at 9:59 AM, ogterran wrote:
>
>>
>> Hi Ian,
>>
>> Thanks for th
The longer the keys are the worse performance is so I would just add a
single known character prefix to your keys.
A single character shouldn't have any noticable impact on performance.
On Tue, Jun 23, 2009 at 7:14 PM, Koen Bok wrote:
>
> I want my datastore keys to be uuid's (hex presentation)
read/thread/73dc1dc31bfd497b
>>
>> Do you know if this was fixed in a recent release?
>>
>> Adrian
>>
>>
>> On Jun 23, 11:33 am, "Nick Johnson (Google)"
>> wrote:
>> > 2009/6/23 Ian Lewis
>> >
>> > > ogterra
If you used numeric keys then you can do pagination by key in order to get
all items from the datastore and count them. Basically this involves
querying the datastore several times, each time filtering on keys that are
greater than the highest key from the results of the previous query. You
won't w
Thats the thing about Django. It tends to not work well unless you use the
whole framework. INSTALLED_APPS is a setting in the settings.py in django.
You'll probably need to create a settings.py and put
'django.contrib.humanize' in the INSTALLED_APPS list.
see: http://docs.djangoproject.com/en/dev
Ritesh,
Keys can either have a numeric id or a key_name. In order to access the key
name for entities that you saved with a key_name you need to use the
key_name() method.
instead of obj.key().id(), try obj.key().name()
Incedentally you can also use obj.key().id_or_name() to get either the id o
a key_name.
So to answer your question, yes, within a model/Kind you can have some
entities that have a numeric id and some that have key names.
Ian
2009/6/24 Ian Lewis
> Ritesh,
>
> Keys can either have a numeric id or a key_name. In order to access the key
> name for entities th
Yes, that is correct.
2009/6/24 Ritesh Nadhani
>
> Actually, what I meant was:
>
> I cannot have an entity with both key_name and id.
>
> They seem to be mutually exclusive for an entity.
>
> 2009/6/23 Ian Lewis :
> > Ritesh,
> >
> > Also, BTW, The re
2009/6/24 Nick Johnson (Google)
> Pagination by key works regardless of if you're using key names or IDs.
>
Good to know. Sorry about the misinformation. I had forgotten that entities
are ordered in the datastore by their key.
BTW, here is an article in the documentation about pagination:
http:
Hi all,
I just wanted to let everyone know about a site I created for sharing code
snippets in a social way. You can check it out at http://www.smipple.net/.
Smipple allows you to save, organize, and share snippets with others easily.
The best part though is that you can follow other people who ha
ags to other peoples' snippets, or
> otherwise mark collections in a way that I can link someone to a
> collection I've made would be great.
Interesting. I'll consider it.
> - Are you considering releasing the source code?
Yes. Probably in the next month or so.
2009/6/30 Nick Johnson (Google)
>
> 2009/6/29 Ian Lewis :
> > Nick,
> >
> > Thanks for the suggestions. I really appreciate that you spent the time
> to
> > write them out. I could track them better if you left them as feedback
> > though. That's
On Fri, Nov 12, 2010 at 8:27 AM, dflorey wrote:
> Hi,
> please have a look at "trecontacts", the 1-minute cron jobs are still
> not getting executed.
> ~500 apps of our customers are affected and stopped working ~12 hours
> ago.
>
> Daniel
It seems it was out for us as well from 22:17 to 08:12
Hi all,
I've written an application that does some work every X minutes/seconds for
each entity of an entity type in the datastore. Currently this is done by
running an cron job and calculating the next runtime for each entity and
enqueing tasks with delays so that they are run at the appropriate
While the task name thing should generally work, they don't provide a
guarantee that inserting the task
won't succeed more than once. Also, I don't particularly want to have to add
a task to check if it is inserted.
There is also the problem of tombstoning names, so I would need to come up
with uni
On Mon, Nov 22, 2010 at 9:14 AM, Nick Johnson (Google) <
nick.john...@google.com> wrote:
> Hi Ian,
>
> On Mon, Nov 22, 2010 at 11:10 AM, Ian Lewis wrote:
>
>> While the task name thing should generally work, they don't provide a
>> guarantee that inserting
hanks,
Ian
--
===
株式会社ビープラウド Ian Lewis
〒150-0021
東京都渋谷区恵比寿西2-3-2 NSビル6階
email: i...@beproud.jp
TEL:03-6416-9836
FAX:03-6416-9837
http://www.beproud.jp/
===
--
You received this message because you are subscribed to the Google Groups
&
Engine
>> Blogger: http://googleappengine.blogspot.com
>> Reddit: http://www.reddit.com/r/appengine
>> Twitter: http://twitter.com/app_engine
>>
>>
>>
>> On Wed, Nov 24, 2010 at 5:21 PM, Ian Lewis wrote:
>>
>>> Hey all,
>>>
>>> We
Hi!
We just released Kay 1.1 rc1!
Kay is a web framework made specifically for Google App Engine. The basic
design of Kay is based on the Django framework, like middleware, settings,
pluggable applications, etc.
Kay uses Werkzeug as a lower level framework, Jinja2 as template engine, and
babel fo
Hi,
The number if indexes isn't affected by the number of properties on an
entity. It only has to do with the number of indexes for "complex queries"
that are needed in your application (i.e. Those in index.yaml). So the
number of indexes created automatically (by properties on indexes) are
probab
Speaking from experience, when the datastore is in maintenance mode,
memcache has also always been in maintenance mode. Given that it supports
namespacing the same way the datastore does and is almost always in
maintenance mode at the same time it's likely that memcache is tied to the
datastore in
Hi!
The Kay team just released Kay 1.0.1 rc1!
This is a release candidate for users running Kay 1.0.0 and includes
bug fixes backported from the 1.1 series. We would love for Kay users
to test it out and report any bugs you may find.
Download here: http://code.google.com/p/kay-framework/download
Hi,
I think that the Appengine team is recommending that folks use the HR
datastore for new applications.
That said the only benefit seems to be that you can write to the
datastore even during maintenance so if that is not important to you,
and your data is likely to be very large then master/sla
Hi everyone,
We just released Kay 1.1 rc2!
The release target of last Monday came and went but work on Kay 1.1
hasn't halted. A few bugs were fixed and a couple more patches were
included in this release. If there are no unexpected problems we plan
to release 1.1 on Monday March 7th. This is stil
Hi everyone,
We just released Kay 1.1 and 1.0.1 final!
Kay 1.1 is the next minor release of the Kay Framework for Appengine.
It includes new features as well as bug fixes.
Kay 1.0.1 includes bug fixes in Kay 1.1 backported to the 1.0.x branch
of Kay. It is for users of Kay who do not want to
upg
88 matches
Mail list logo
