On 3/2/06, Adrian Holovaty <[EMAIL PROTECTED]> wrote:
>
> On 3/1/06, Russell Keith-Magee <[EMAIL PROTECTED]> wrote:
> > Obviously, this is a name clash, so the model validator complains.
>
> Russ, any way you can improve that error message to explicitly say
> "Add related_name" or something along
Eugene Lazutkin wrote:
>I just posted patches (trunk and magic) for *Subj*. You can find details
>here: http://code.djangoproject.com/ticket/1442. It supersedes tickets
>#463, #900, and #1237. The fix was inspired by ticket #1268.
>
>
"class DatabaseWrapper(local)" looks nice :-)
Though it ma
Ivan Sagalaev wrote:
>However this patch revealed pretty nasty show-stopper bug: we leak
>connections somewhere. For even a simple view that does only a single
>select connection is opened, then closed and then opened again and never
>closed.
>
Found it. Both modpython.py and wsgi.py have this
On 3/2/06, Rock <[EMAIL PROTECTED]> wrote:
>
> My other changeset developed at this week's Django Sprint is much more
> extensive and is best explained by the new section that I am adding to
> the DB API docs.
>
> Feedback is welcome.
Truth be told, a desire to improve aggregates in Django is the
On 3/2/06, limodou <[EMAIL PROTECTED]> wrote:
> I know that. I'v subscribe the trac update maillist. But there are too
> much changes. And I want just an outline. I think the change logs also
> need as the new version released. :^)
I've taken a break from doing the "Week in review" blog entries
b
Russell Keith-Magee posted this in a separate thread:
> Article.objects.all().count() reads quite obviously as the count of
> all articles. However, Article.objects.all().count('title') reads (to
> me) like an inelegant way of saying 'a count of all titles'. This
> isn't what your proposal would r
在 2006-3-2 12:28:04,Eugene Lazutkin <[EMAIL PROTECTED]> 写道:
>
> I just posted patches (trunk and magic) for *Subj*. You can find details
> here: http://code.djangoproject.com/ticket/1442.
It's good and pythonic. I like the solution.
Should we use the code
"""
def close( self
Ivan Sagalaev wrote:
>>
> Found it. Both modpython.py and wsgi.py have this bit:
>
> try:
> request = ModPythonRequest(req)
> response = self.get_response(req.uri, request)
> finally:
> db.db.close()
>
> # Apply response middleware
> for middleware_method
Ivan Sagalaev wrote:
>
> Though it mandates opening new connection for each request which would
> kill even "hello world" if it happens to use Oracle for backend :-).
> However this problem is better solved with some external connection pool
> solution.
I am all up for a generic connection po
张骏 wrote:
>
> Should we use the code
> """
> def close( self ):
> if setting.USING_THREAD_POOL:
> return
> else:
> self.connection.close()
> """
> instead of the oranginal "DatabaseWrap.close" function for
Eugene Lazutkin wrote:
>If I understood you correctly, you proposes to introduce new
>USING_THREAD_POOL setting. If it is set, we don't close connection
>hoping that it is going to be reused, if the thread itself was reused
>for other request. It leaves the connection open until thread dies and
>
On Thursday 02 March 2006 04:49, xamdam wrote:
> class Book(models.Model):
> name = models.CharField(maxlength=200)
> author = models.ManyToManyField(Person)
> def set_author(self, a): self.author = a # is this (having to
> define the set_ method) fixed yet?
> def __repr__(self):
Ivan Sagalaev wrote:
>It's not at all that simple in practice
>
I just recalled one more thing.
Cheap pooling (by just not closing connections) wouldn't let you control
separately lifetime and quantity of Apache child processes and open db
connections. And this is usually needed since these a
Hi,
the behaviour of django for fields that contain html markup ("<" ...)
is wrong, at least in version 0.91.
First, the admin pages do not html escape these. There might be
corner cases when this is intended, but it seems very dangerous.
Second, regarding Template, I think the default not t
Ivan Sagalaev wrote:
>
> It's not at all that simple in practice. You can't just leave the
> connection for later reuse because it holds the whole environment:
> transactions, locale settings, authentication, may be something else.
> Each bit of this environment should be handled separately:
>
>
What if we changed terminology so that "project" became "site"
throughout the Django code and documentation?
$ django-admin.py startsite myblog
Seems like this would make things a bit easier for the new user, who
has no real idea what "project" means but surely knows what "site"
means.
I was go
a 'site' in django terminology refers to a instance of your
application running with a given group of settings.
ie.. there is a LJW site and a lawerence site both running the
ellington app. (possibly on the same machine)
On 3/3/06, pbx <[EMAIL PROTECTED]> wrote:
>
> What if we changed terminolog
The whole concept of projects in Django is that they aren't tied to
sites. You can use the same project across multiple sites and use
multiple projects on one site.
On 3/2/06, Ian Holsman <[EMAIL PROTECTED]> wrote:
>
> a 'site' in django terminology refers to a instance of your
> application runn
On Fri, 2006-03-03 at 09:43 +1100, Ian Holsman wrote:
> a 'site' in django terminology refers to a instance of your
> application running with a given group of settings.
>
> ie.. there is a LJW site and a lawerence site both running the
> ellington app. (possibly on the same machine)
Now, Ian, d
On 3/3/06, Luke Plant <[EMAIL PROTECTED]> wrote:
fields. For ManyToMany, you do: book.author.add(author1, author2...)I guess it would be good if the __set__ descriptor was there, and eitherdid the right thing or threw some exception.
I've been working with the descriptor protocol on RelatedManage
On 3/2/06, Malcolm Tredinnick <[EMAIL PROTECTED]> wrote:
> My initial thought is that I'm -0 on the proposal, since I'm not sure it
> makes that much of a difference. It is kind of logical, since a project
> is a collection of applications and the configuration that ties them all
> together. But,
pbx wrote:
> I was going to file a ticket but wanted to make sure there isn't some
> important conceptual distinction that I'm missing. Is there?
>
I am definitely -2 on this. Malcolm does a good job of starting the ball
rolling as to why.
For any admin / web type guy a 'site' is a machine / v
Russell Keith-Magee wrote:
> I can see the merit in finishing the descriptor protocol
> for all m2m objects. So:
>
> book.authors = [author1, author2]
> would be equivalent to
> book.author.clear(); book.author.add(author1, author2)
Assigning a list implies assigning an order. How about:
book.au
OK, thanks for all the clarifications and corrections. A couple quick
responses:
(Sean)
> For any admin / web type guy a 'site' is a machine / virtual server. It
> contains many projects / apps / whatever.
This actually isn't true for me -- my virtual server has a dozen
domains on it, and for me
On 3/2/06, Sean Perry <[EMAIL PROTECTED]> wrote:
> I am definitely -2 on this. Malcolm does a good job of starting the ball
> rolling as to why.
The range is -1..+1 ;-)
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Group
On 3/3/06, kmh <[EMAIL PROTECTED]> wrote:
Assigning a list implies assigning an order. How about:book.authors.set([author1, author2])
This doesn't address the problem - what is the 'set' behaviour of the descriptor protocol. The start of the _expression_ has to be 'book.authors ='; the question is
On Thu, 2006-03-02 at 15:59 +1100, Malcolm Tredinnick wrote:
[...]
> So, I guess my questions are as follows:
>
> (1) Why are we only filling in auto_now_add and auto_now in save(),
> rather than prior to validation? Purely accidental, or is there a more
> noble purpose?
>
> (2) Is the right pla
Russell Keith-Magee wrote:
> On 3/3/06, kmh <[EMAIL PROTECTED]> wrote:
>
> > Assigning a list implies assigning an order. How about:
>
>
> book.authors.set([author1, author2])
>
>
> This doesn't address the problem - what is the 'set' behaviour of the
> descriptor protocol.
>The start of the expre
On 3/3/06, kmh <[EMAIL PROTECTED]> wrote:
If it were a 'set' method assignment would not be possible and anexception hinting to use 'set' would be raised. (Damn English language... setting the set with a set...)If I am understanding you correctly, you are opposed to having assignment notation for
Russell Keith-Magee wrote:
> On 3/3/06, kmh <[EMAIL PROTECTED]> wrote:
> >
> >
> > If it were a 'set' method assignment would not be possible and an
> > exception hinting to use 'set' would be raised.
>
>
> (Damn English language... setting the set with a set...)
>
> If I am understanding you corr
Just synced the latest magic-removal this morning;
$python manage.py init
yields
Error: Your action, 'init', was invalid
$python manage.py install admin
installs the django_admin_log table ONLY, so nothing really works (is
it dependent on the previous init failure?)
thanks,
max
--~--~
init has been replaced with syncdb
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups
"Django developers" group.
To post to this group, send email to django-developers@googlegroups.com
To unsubscribe from this group, s
On 3/2/06, Malcolm Tredinnick <[EMAIL PROTECTED]> wrote:
> Ellington is a project. It contains applications.
Just for the record, World Online (and I, for my stuff at
washingtonpost.com and personal use) don't even use "projects". I have
a "holovaty" package that contains a bunch of apps and *no*
On 3/3/06, Adrian Holovaty <[EMAIL PROTECTED]> wrote:
> Anyhow, I'm not crazy about the term "project" either, so we can
> definitely change it if there's a better term. How about
> "extravaganza" or "experience"?
-1
I think it is change for the sake of change.
From a documentation perspective y
34 matches
Mail list logo