Re: form_for_instance and form argument, empty form?
> > On 11/7/07, David Larlet <[EMAIL PROTECTED]> wrote: > >> Hi all, > >> > >> I'd spent a long time finding that bug but I want to be sure before > >> submitting it on Trac. If you pass a form argument to form_for_instance > >> like that: > >> > >> forms.form_for_instance(foo, form=FooForm) > >> > >> with an instance of foo which only contains a basic field (let's say a > >> CharField) and FooForm with a unique field too, the input rendered will > >> not be completed with the content of the foo instance. If you remove the > >> form argument: > >> > >> forms.form_for_instance(foo) > >> > >> the generated form contains the content of the foo instance. > >> > >> I'd tried to find the bug but this line (122 of newforms.models): > >> > >> return type(opts.object_name + 'InstanceForm', (form,), > >> {'base_fields': base_fields, '_model': model, > >> 'save': make_instance_save(instance, fields, 'changed')}) > >> > >> give me headaches ;-). The only thing I can say is that base_fields > >> contains the initial data before this line but my returned form not. > >> > >> Did somebody use this argument and can confirm? I can confirm that it does in fact work as you describe, but I don't believe it's a bug. The only documentation I was able to find on this argument is in the docstrings for form_for_model and form_for_instance, both of which clearly state that it's intended to receive a subclass of BaseForm, not Form. Here's the difference. BaseForm is what provides all the real functionality for the form: iterating over fields, triggering HTML output, managing validation, etc. It's not normally used outside of Django's internals, because the Form class subclasses it already. However, it doesn't know anything about how to get the fields for the form. Instead, the Form class uses a special way to pull fields out of the class definition and assign them where they belong. form_for_model and form_for_instance don't need this syntax, so they bypass that step by directly assigning the fields where they belong when they create the class. When you pass in a subclass of Form, it's already got its fields in the right place, but more importantly, it triggers that syntax checking again, where it looks for new fields. It basically copies fields from a parent class, then adds in the fields that it found in the new class. However, since form_for_model and form_for_instance don't supply any fields that way, all it gives the new form is what it found in the parent class, completely overwriting the fields it got from form_for_*. That was a bit of a long explanation, but I think it's necessary to know what's going on, and it leaves me with two more points. As for what that form argument was meant for, it's designed for a BaseForm subclass. This allows you to add *methods* to your generated form, not new fields. As for what you're trying to do, check out the 'fields' argument to form_for_instance. It creates a form with just a subset of the fields from the original model. -Gul --~--~-~--~~~---~--~~ 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, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/django-developers?hl=en -~--~~~~--~~--~--~---
Re: form_for_instance and form argument, empty form?
I thought that it was more appropriated to post it on the devlist because it sounds like a bug but ok let's move it on the userlist, sorry for the noise here. I'll be glad to hear your solution. David Marty Alchin a écrit : > This should probably be asked on django-users, as it's more about how > to use form_for_instance() than any internal development. And yes, > that means I don't believe it's a bug, and I'll gladly explain more on > django-users so more people can hear it. > > -Gul > > On 11/7/07, David Larlet <[EMAIL PROTECTED]> wrote: > >> Hi all, >> >> I'd spent a long time finding that bug but I want to be sure before >> submitting it on Trac. If you pass a form argument to form_for_instance >> like that: >> >> forms.form_for_instance(foo, form=FooForm) >> >> with an instance of foo which only contains a basic field (let's say a >> CharField) and FooForm with a unique field too, the input rendered will >> not be completed with the content of the foo instance. If you remove the >> form argument: >> >> forms.form_for_instance(foo) >> >> the generated form contains the content of the foo instance. >> >> I'd tried to find the bug but this line (122 of newforms.models): >> >> return type(opts.object_name + 'InstanceForm', (form,), >> {'base_fields': base_fields, '_model': model, >> 'save': make_instance_save(instance, fields, 'changed')}) >> >> give me headaches ;-). The only thing I can say is that base_fields >> contains the initial data before this line but my returned form not. >> >> Did somebody use this argument and can confirm? >> >> Regards, >> David >> >> >> > > > > --~--~-~--~~~---~--~~ 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, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/django-developers?hl=en -~--~~~~--~~--~--~---
Re: form_for_instance and form argument, empty form?
Marty Alchin a écrit : >>> On 11/7/07, David Larlet <[EMAIL PROTECTED]> wrote: >>> Hi all, I'd spent a long time finding that bug but I want to be sure before submitting it on Trac. If you pass a form argument to form_for_instance like that: forms.form_for_instance(foo, form=FooForm) with an instance of foo which only contains a basic field (let's say a CharField) and FooForm with a unique field too, the input rendered will not be completed with the content of the foo instance. If you remove the form argument: forms.form_for_instance(foo) the generated form contains the content of the foo instance. I'd tried to find the bug but this line (122 of newforms.models): return type(opts.object_name + 'InstanceForm', (form,), {'base_fields': base_fields, '_model': model, 'save': make_instance_save(instance, fields, 'changed')}) give me headaches ;-). The only thing I can say is that base_fields contains the initial data before this line but my returned form not. Did somebody use this argument and can confirm? > > I can confirm that it does in fact work as you describe, but I don't > believe it's a bug. The only documentation I was able to find on this > argument is in the docstrings for form_for_model and > form_for_instance, both of which clearly state that it's intended to > receive a subclass of BaseForm, not Form. > You're absolutely right on that point. I just found it in the documentation too: http://www.djangoproject.com/documentation/models/model_forms/ > Here's the difference. > > BaseForm is what provides all the real functionality for the form: > iterating over fields, triggering HTML output, managing validation, > etc. It's not normally used outside of Django's internals, because the > Form class subclasses it already. However, it doesn't know anything > about how to get the fields for the form. Instead, the Form class uses > a special way to pull fields out of the class definition and assign > them where they belong. form_for_model and form_for_instance don't > need this syntax, so they bypass that step by directly assigning the > fields where they belong when they create the class. > > When you pass in a subclass of Form, it's already got its fields in > the right place, but more importantly, it triggers that syntax > checking again, where it looks for new fields. It basically copies > fields from a parent class, then adds in the fields that it found in > the new class. However, since form_for_model and form_for_instance > don't supply any fields that way, all it gives the new form is what it > found in the parent class, completely overwriting the fields it got > from form_for_*. > > That was a bit of a long explanation, but I think it's necessary to > know what's going on, and it leaves me with two more points. > > As for what that form argument was meant for, it's designed for a > BaseForm subclass. This allows you to add *methods* to your generated > form, not new fields. > > As for what you're trying to do, check out the 'fields' argument to > form_for_instance. It creates a form with just a subset of the fields > from the original model. > Wow, thanks for this interesting explanation! I always wondered what's the difference between Form and BaseForm without digging too deep into the code. Now that's really clear. Unfortunately, what happens with my Form is exactly what I'd like to do: create a generic form from a Form class which define fields (with widgets, etc) and methods and which is initialized with the content of the instance. That seemed more natural to me instead of dealing with fields, form and eventually form_for_fields... I know that those functions are just shortcuts but it works very well in terms of genericity and my main concern here is to integrate a generic way to handle forms for resources in restapi. If you have any suggestion about that I'll be very happy to learn a bit more :) David --~--~-~--~~~---~--~~ 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, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/django-developers?hl=en -~--~~~~--~~--~--~---
Re: form_for_instance and form argument, empty form?
This should probably be asked on django-users, as it's more about how to use form_for_instance() than any internal development. And yes, that means I don't believe it's a bug, and I'll gladly explain more on django-users so more people can hear it. -Gul On 11/7/07, David Larlet <[EMAIL PROTECTED]> wrote: > > Hi all, > > I'd spent a long time finding that bug but I want to be sure before > submitting it on Trac. If you pass a form argument to form_for_instance > like that: > > forms.form_for_instance(foo, form=FooForm) > > with an instance of foo which only contains a basic field (let's say a > CharField) and FooForm with a unique field too, the input rendered will > not be completed with the content of the foo instance. If you remove the > form argument: > > forms.form_for_instance(foo) > > the generated form contains the content of the foo instance. > > I'd tried to find the bug but this line (122 of newforms.models): > > return type(opts.object_name + 'InstanceForm', (form,), > {'base_fields': base_fields, '_model': model, > 'save': make_instance_save(instance, fields, 'changed')}) > > give me headaches ;-). The only thing I can say is that base_fields > contains the initial data before this line but my returned form not. > > Did somebody use this argument and can confirm? > > Regards, > David > > > > > --~--~-~--~~~---~--~~ 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, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/django-developers?hl=en -~--~~~~--~~--~--~---
Re: form_for_instance and form argument, empty form?
On 11/7/07, David Larlet <[EMAIL PROTECTED]> wrote: > Unfortunately, what happens with my Form is exactly what I'd like to do: > create a generic form from a Form class which define fields (with > widgets, etc) and methods and which is initialized with the content of > the instance. That seemed more natural to me instead of dealing with > fields, form and eventually form_for_fields... I know that those > functions are just shortcuts but it works very well in terms of > genericity and my main concern here is to integrate a generic way to > handle forms for resources in restapi. Well, from what you posted before, your forms aren't really all that "generic", so I'm not sure exactly what you're trying to accomplish. If you're just looking for a simple way to populate a form's initial values with those from a matching model, try this: foo = Foo.objects.get(id=1) form = FooForm(initial=foo.__dict__) Now, I haven't tested that, so I make no guarantees, but at a quick glance, it looks like it should do what you're looking for. Remember, form_for_instance is for creating a form *and* populating it, not just for populating an existing form. Also, keep in mind that you won't have the save() method on FooForm unless you add it yourself, so I'm really not sure how this would be any better than just using form_for_instance as-is. -Gul --~--~-~--~~~---~--~~ 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, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/django-developers?hl=en -~--~~~~--~~--~--~---
Re: form_for_instance and form argument, empty form?
On 11/7/07, SmileyChris <[EMAIL PROTECTED]> wrote: > Actually, I see no good reason why you shouldn't be able to use a Form > as the SuperClass of your form_for_* form. Yeah, I was trying avoid the issue of "should" of "shouldn't" and just point out what was going on. -Gul --~--~-~--~~~---~--~~ 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, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/django-developers?hl=en -~--~~~~--~~--~--~---
Re: form_for_instance and form argument, empty form?
On Nov 8, 4:30 am, "Marty Alchin" <[EMAIL PROTECTED]> wrote: > When you pass in a subclass of Form, it's already got its fields in > the right place, but more importantly, it triggers that syntax > checking again, where it looks for new fields. It basically copies > fields from a parent class, then adds in the fields that it found in > the new class. However, since form_for_model and form_for_instance > don't supply any fields that way, all it gives the new form is what it > found in the parent class, completely overwriting the fields it got > from form_for_*. Actually, I see no good reason why you shouldn't be able to use a Form as the SuperClass of your form_for_* form. I've just attached a patch to #3815 which makes it work. --~--~-~--~~~---~--~~ 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, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/django-developers?hl=en -~--~~~~--~~--~--~---
Re: form_for_instance and form argument, empty form?
Marty Alchin a écrit : > On 11/7/07, David Larlet <[EMAIL PROTECTED]> wrote: > >> Unfortunately, what happens with my Form is exactly what I'd like to do: >> create a generic form from a Form class which define fields (with >> widgets, etc) and methods and which is initialized with the content of >> the instance. That seemed more natural to me instead of dealing with >> fields, form and eventually form_for_fields... I know that those >> functions are just shortcuts but it works very well in terms of >> genericity and my main concern here is to integrate a generic way to >> handle forms for resources in restapi. >> > > Well, from what you posted before, your forms aren't really all that > "generic", so I'm not sure exactly what you're trying to accomplish. > Ok, my previous example is really basic to extract the issue but I need to explain the background now. restapi [1] let you define resources which match a queryset. Those resources are python class (just read at the 5 steps on the homepage of the project to understand it well). When a request is done to the url, the class call the __call__ function and dispatch this one to the right function given the HTTP verb (to sum up). Now let's focus on the form part. The python class (Collection) can be instantiated with a form_class argument which is used in all functions that require form_for_* and this is a key element here. You can have a look at the bottom of the responder file [2] to see what I mean. I didn't know that a BaseForm was required here when I'd proposed to add this argument [3] and I thought that it was a good way to have a generic form which can be customized. Since your previous email, I'm a bit confused about that because I realize that it's a bit useless if I can't get my forms from instances rendered with instances' data. I hope this is more clear now (even if my english is not that good...). [1] http://code.google.com/p/django-rest-interface/ [2] http://django-rest-interface.googlecode.com/svn/trunk/django_restapi/responder.py [3] http://groups.google.com/group/django-developers/browse_thread/thread/f86bd8ea6579baaa/709fdf8e092084ea?lnk=gst SmileyChris a écrit : > On Nov 8, 4:30 am, "Marty Alchin" <[EMAIL PROTECTED]> wrote: > >> When you pass in a subclass of Form, it's already got its fields in >> the right place, but more importantly, it triggers that syntax >> checking again, where it looks for new fields. It basically copies >> fields from a parent class, then adds in the fields that it found in >> the new class. However, since form_for_model and form_for_instance >> don't supply any fields that way, all it gives the new form is what it >> found in the parent class, completely overwriting the fields it got >> from form_for_*. >> > > Actually, I see no good reason why you shouldn't be able to use a Form > as the SuperClass of your form_for_* form. > > I've just attached a patch to #3815 which makes it work. > > Did this patch populate form fields with instance data? I can't figure out this given the tests. Regards, David PS: RajeshD let me know if you're on this list, otherwise I'll forward this email to you, I don't want to cross post anymore. --~--~-~--~~~---~--~~ 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, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/django-developers?hl=en -~--~~~~--~~--~--~---