frank wrote:
> Sure, I could create a download button for each
> item, but that would be an extra step for the user. 'It would be nice'
> if that weren't necessary.
> BTW the downloads are static files, served by Apache. That extra
> button
> is looking more and more likely...
At some point
Malcolm Tredinnick wrote:
> The minor problem with that approach is that it means people can avoid
> the Django counter and directly fetch the URL. So if counting is a
> requirement, that solution needs an extra tweak to ensure that you have
> first visited the counter page. This can be done wit
ashwoods wrote:
> does anybody have experience of sharing session information from a
> plone site. I want our intranet users that are authenticaded in plone
> to be able to access django without having to relog.
Haven't done this directly, but what you'd have to do is get the __ac
cookie that p
This sounds like you want to use my 'single row database' trickery
pokery:
http://groups.google.com/group/django-developers/browse_frm/thread/2caa976249783fae/#
Just create a Django model 'Properties' with your 'threshold' as one
field, then access it just like a normal Django model. The databas
[EMAIL PROTECTED] wrote:
> I could create a model and then only have one row, but that seems
> clunky.
I did exactly this and overrode the .save() method so that the model
was constrained to only have one row by resetting self.id to 1 in the
.save() (I think...). Effect was that if you added a
Ah, and you also have to override the .delete() method!
http://groups.google.com/group/django-developers/browse_frm/thread/2caa976249783fae/#
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups
"Django users" group.
T
My slightly graceless method is to override .save and .delete:
http://groups.google.com/group/django-developers/browse_frm/thread/2caa976249783fae/#
Then whatever model you define, it can only ever have one row (well, it
will have zero rows before you put anything it).
Barry
--~--~-~-
ashwoods wrote:
a simple open and free CMS would be good for a lot of people, specially
the newcomers. the problem is how to design a CMS that doesn't limit
the freedom you learn to love when you choose to use django over some
cms.
Someone looks like they tried to start a project like this,
I've had a few thoughts on this:
1. on the main admin page, just list the names of the apps and not the
db tables as well.
2. clicking an app takes you to a per-app admin page, by default it
would just be the one pretty blue-headed box for the app.
3. then it would be easy to have a custom per-
I dont want my users to have to bother with a username. I want them to
authenticate with their email as their username. But django wont allow
valid @ signs (and other stuff) in a username. I could patch that but
that could bite me badly. So I found a better way.
First I create a manipulator:
fro
This is to avoid duplicate usernames? My idea is to call them all
userXXX where XXX is the id value (the primary key in the User table).
Another thought that came to me yesterday was whether there are any
issues if a user wants to change his or her email address. I don't
think it causes any probl
just use PasswordField() in your model.
Might be a recent introduction. Its literally a two-liner in
django.forms:
class PasswordField(TextField):
input_type = "password"
http://code.djangoproject.com/browser/django/trunk/django/forms/__init__.py
Barry
--~--~-~--~
Wouldn't it be better to setup 'procmail' to process incoming emails as
they arrive. Then your procmail script could update the database. As
long as you are using a decent database server (*cough*POSTGRES*cough*)
then concurrency shouldnt be a problem.
--~--~-~--~~~--
Ground up would be my way of doing it. Lets break it down...
You need an 'update' view that only shows the fields you want changing.
That means a custom manipulator that only mentions the fields you want
to modify. Protect this view by making sure the logged-in user is the
user mentioned in the U
DavidA wrote:
> An alternative that stays inside Django is to setup a trigger in cron
> (or NT's Task Scheduler) that gets a URL every few minutes and the view
> code for that URL does the email processing you mention. You avoid
> creating a true daemon/service just by "waking" up periodically a
Jay Parlar wrote:
> The thing is, I'm not sure how to fit that into Django. Currently, my
> view function just blocks on the generation/compilation. To do this
> with AJAX, it feels like I'd have to spawn a thread from Django to do
> the processing, and let the view return immediately. But when
fyleow wrote:
> keyword = books.objects.get(headline__icontains=keyword)
I would then split/reshape this list into a list of (keyword,
list_of_books) tuples, so you have a structure like this:
keywordsearch = [('python', [,]), ('coding',[,])]
then in the template you loop "for kwmatch in
I'd rather go the other way - generate a class diagram from django
model source!
If you generate model code from a diagram you'll probably still have to
edit the source to add new methods or tweak verbose_names, or add Admin
meta classes, so then if you change the model you will lose all those
ch
Viktor wrote:
> But when I delete a round, with round.delete(), all games conected with
> that round are also deleted, but the Score table isn't updated, that is,
> the delete method from the game objects is not called?!?!
Looking at the code I see that when an objects is .delete()d all the
rel
Indeed! Perhaps nobody thought anyone would ever override the delete()
method...
I'm not sure why the strategy is to gather together all the related
objects and then do the SQL rather than call the delete() method on
each of them. Perhaps its more efficient. Perhaps it avoids possible
loops where
> On Thu, 2006-06-22 at 12:47 +, Paul Childs wrote:
> > I have tried a number of ways of doing this but can't come up with a
> > solution. I don't think pre-processing the data will help since the
> > template still doesn't know how many keys there will be. Changing the
> > structure of the
Well, you could encode the information that is normally transferred in
the cookie by adding a parameter to the URL all the time, so that your
URLs are always something like:
/foo/bar/info?session_id=0873556323
BUT if anyone gets that URL they get that person's session. Which is a
BAD thing. So d
ihomestore wrote:
> A simple case would be: once users login, if we do not keep track of
> user session, how do we know it is the same user once he / she diverts
> to other pages (we do not want to ask user to login for every new page
> s/he visits).
Okay, so duplicating the functionality of d
Felix Ingram wrote:
> Well that's true but 128 bits still gives you 3 x 10^38 possible
> strings, so you could probably give everyone on the planet a key an
> still not have a collision.
Understatement of the week! The human population is currently about
6.5x10^9. You could give every teaspoon
[EMAIL PROTECTED] wrote:
> I just tried two loops "for" imbricated, it does'nt seems to work.
Imbricated? That's a new one on me. I just looked it up and it means
overlapping like roof tiles, edge over edge, which makes me think of
two for loops like this:
for i = 1 to 10
for j = 1 to 10
print
一首诗 wrote:
> I have a table like this
>
> id | name | parentid
>
> And parentid is a Foreign Key to ID.
>
> So I have a tree in my database. My question is that, how can I
> represent it on the web as a tree.
If you can rework your database slightly you can use Modified Preorder
Tree Traversal
Someone is working on a Django CMS:
http://groups.google.com/group/dynamite-developers?lnk=li
but there's no code available yet...
Barry
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups
"Django users" group.
To po
[EMAIL PROTECTED] wrote:
> suppose i have a model with 2 keywords like
>
> class Test(models.Model):
>result_id = models.IntegerField(maxlength=10)
> lab_id = models.IntegerField(maxlength=10)
>
> is there a method to extract all the keywords from this class.
> So in this example i
[EMAIL PROTECTED] wrote:
> thanks,
>
> that worked
>
> richard
Ah, now I see my confusion. What you wanted was the 'field names', not
'keywords', and you said: [result_id,lab_id] when you really meant
['result_id','lab_id'].
Barry
--~--~-~--~~~---~--~~
You rec
And (obviously?) dont rely on the correct make and model coming back
from your form. You'll eventually get someone constructing their own
POST data for a laugh and seeing what happens if they had selected
Renault Impala... :)
--~--~-~--~~~---~--~~
You received thi
Sean Schertell wrote:
> Am I alone on this? If I created such a repository would anyone use it?
Why not just use SourceForge[1] as repository and just keep a list of
django-related projects on the django main site?
Barry
[1] Or similar.
--~--~-~--~~~---~--~~
I use Windows at home for video and audio editing, and getting Django
things to work on there can be annying at times - its usually file
paths to python, and things like #!/bin/env and all that, and whether
you run from cygwin or DOS shell. Aaargh, the horror. I wrote my poker
blog at home on Wind
I did something like this in perl using HTML::Mason. However I used
'system' to start the background process, and files for
communication...
Its an on-demand backup system. The HTML::Mason code creates a temp
file and then uses 'system' to run the backup job, passing the
filename. The backup then
33 matches
Mail list logo
