Re: Lookup by week
Adrian Holovaty wrote: On 1/30/06, Jan Rademaker <[EMAIL PROTECTED]> wrote: Would it make sense to implement such a thing? An alternative would be to figure out at what date a week starts and ends and just use pub_date__range, but that's not really the point. I'm just curious how easy/hard it would be to implement. If the above isn't going to work, what's the best/most elegant way to go, __range? Yeah, I'd suggest calculating the week start/end dates in your own code and using the __range lookup. Weeks are a bit of a messy problem -- different people have different definitions of when they start, etc. Then, if weeks are implemented, you start to get requests for fortnights and scores... -- --Max Battcher-- http://www.worldmaker.net/
Re: Extending User
I've done this with the latter approach. I made a person model that had more person-specific information and made the user field a foreign key and added the edit_inline option to have the information right there in the user interface: user = meta.ForeignKey(auth.User, edit_inline=meta.STACKED, num_in_admin=1, max_num_in_admin=1) This works pretty well for me at least. -berto. On 1/30/06, char <[EMAIL PROTECTED]> wrote: > > I've seen a couple of approaches on this mailing list for adding > application specific fields to the User class. The first involves > extending auth.User but this seems a little complex, and was > specifically labeled an "advanced" method. It doesn't seem like you > should have to break out deep black magic to do something so standard. > > The other method is to have a foreign key pointing to a User. This is > easier but it makes for a messy Admin interface (resulting in both a > User class and a separate MyUser class that needs to be associated with > a User). What is the "canonical" way for extending atuh.User with > application specific fields? > >
Re: Extending User
That did it. Thanks.
Extending User
I've seen a couple of approaches on this mailing list for adding application specific fields to the User class. The first involves extending auth.User but this seems a little complex, and was specifically labeled an "advanced" method. It doesn't seem like you should have to break out deep black magic to do something so standard. The other method is to have a foreign key pointing to a User. This is easier but it makes for a messy Admin interface (resulting in both a User class and a separate MyUser class that needs to be associated with a User). What is the "canonical" way for extending atuh.User with application specific fields?
Re: error restoring project
Ah excellent - I think it's natural to presume after fixing a bunch of errors and updating paths after restoring from backup to a new dir that the fault is more than likely on your end than with product. For anyone who follows, 'ordering' should be a list and look like: class META: ordering=['title'] ... Luke Skibinski Holt
Re: possible to validate a subset of a model's fields?
You could pickle (as in, serialize) the data and save on the database. Then just unpickle when you need to view it. On 1/30/06, Oliver Rutherfurd <[EMAIL PROTECTED]> wrote: > Thanks for the suggestion, but saving to the db after every page is a > requirement of the app. If someone fills out a page and comes back a > week, or even a year later, everything they previously entered should > be in the db. -- Julio Nobrega - http://www.inerciasensorial.com.br
Re: possible to validate a subset of a model's fields?
Hi Amit, On 1/30/06, Amit Upadhyay <[EMAIL PROTECTED]> wrote: > On 1/31/06, Oliver Rutherfurd <[EMAIL PROTECTED]> wrote: > > The use case for this is a long registration process, where different > > data points from individual models are split across multiple pages. > > The application needs to save data collected after every step, so the > > user doesn't have to start from scratch if they leave the site and > > return later. As a silly example, let's say that name is collected on > > page 1, phone # on page 2. Both are required and both live in a > > 'person' table -- and I want to save the model after page 1. > > I would recommend creating your own FormManipulators, one for each page. > Validate page level manipulator and save the content in request.session, and > when you are done on the last page, pick out all the saved temproary states > and insert it in the database. Thanks for the suggestion, but saving to the db after every page is a requirement of the app. If someone fills out a page and comes back a week, or even a year later, everything they previously entered should be in the db. -Ollie
Re: Error when relating a model to another model more than once
On 1/31/06, Rudolph <[EMAIL PROTECTED]> wrote: Hi,I'm using the development version of Django (trunk) and I've got twosimple classes. The Language class should contain the two letterlanguage abbreviations and the full name of the language in English. The LanguageTranslation class should contain translated names of thelanguages:I haven't had this specific problem myself, but I think I _might_ know the cause of the problem (I don't know for sure, but it seems possible). When Django discovers a ForeignKey in a model, it puts a 'related' field in the model that is referenced. The name that Django uses for the related field is the name of the table that is referenced (in this case, 'language'). However, if you have two foreign keys referencing the same class, you end up with colliding related fields - which Django doesn't detect. The simple solution is to use related_name on the foreign key(s) in your model:... namelang = meta.ForeignKey(Language, verbose_name='Language of thetranslation', related_name='namelang')... I have recently added code to the magic-removal branch to detect this sort of model problem. In the meantime, just add the related_name. Like I said - I haven't had this specific problem myself, but I can see how a field ambiguity like this might cause all sorts of problems in admin UIs, etc.Hope this helpsRuss Magee %-)
Re: possible to validate a subset of a model's fields?
On 1/31/06, Oliver Rutherfurd <[EMAIL PROTECTED]> wrote: The use case for this is a long registration process, where differentdata points from individual models are split across multiple pages.The application needs to save data collected after every step, so theuser doesn't have to start from scratch if they leave the site and return later. As a silly example, let's say that name is collected onpage 1, phone # on page 2. Both are required and both live in a'person' table -- and I want to save the model after page 1. I would recommend creating your own FormManipulators, one for each page. Validate page level manipulator and save the content in request.session, and when you are done on the last page, pick out all the saved temproary states and insert it in the database. -- Amit UpadhyayBlog: http://www.rootshell.be/~upadhyay+91-9867-359-701
Re: Template inheritance problem
Adrian, Thank you for your reply. I still have problems with template inheritance Can you please help me with the following? I have the Index1.html ##Index1.html# {% extends "board/test" %} {% block title %}Tittle from Index.html {% endblock %} This is a text from Index1.html {%block mine%}Hello from Index1{%endblock%} Text at the end ## and a test.html file ###test.html {% block title %}Tittle from test.html{% endblock %} This is a text from test.html {%block mine%}Hello from test.html {%endblock%} I use this view with the template def TestTemplate(request): return render_to_response('board/Index1') I would expect that the title should be Tittle from test.html but it shows still Tittle from Index.html Why? Also no text( except that in block tags in Index1.html) is printed.Why? Thank you for help Rregards, la.
Oracle support?
On the Django site it lists the currently implemented databases and say "more to come soon". Is Oracle on that list? And relatively speaking, how "soon" is soon? Robert
Re: What is save to use in __repr__
A bit of code would have clarified my problem. This is what I tried the first time: return "Testing %s %s %.2f" % (self.someForeignKey.someThing(), self.somedate.strftime("%d-%m-%Y"), self.someFloat) There were 2 problems, basically. First, I tried to access an object instance directly and not through get_(), but that's going to be fixed somehow, or so I've heard on #django. Second, the date field I use has auto_now_add=True (see Maniac's comment) that's causing problems. Would you consider this a bug? It 'feels' wrong. Right now, these kind of issues don't really bother me, I'm just trying to find out if django is right for me, but could something like this make it to 1.0? Regards, -janr
Re: FileField :: mx file size
Waylan Limberg wrote: I envision a simple script that intercepts the upload submit, checks the actual file size against the specified limit (perhaps in a hidden field or hard coded in script or even via an ajax request (but why?)) and either returns an error message or submits the upload. No, those projects use Ajax to periodically poll the server about how many bytes already uploaded. This is why the server should feed data to the framework as soon as it receives them and not after it already have got everything. Then your application code on server can really know how much data it's going to get (from Content-Length) and the store data being received to some known location in memory or on disk and tell to parallel Ajax request the progress.
Re: What is save to use in __repr__
On 1/30/06, Jan Rademaker <[EMAIL PROTECTED]> wrote: > What is save to use __repr__ ? > > I've found out that I shoud use self.get_someForeignKey().someThing() > instead of self.someForeignKey.someThing(), but what about date fields > and such? > > Because, when I use self.someDateField in __repr__ django crashes the > moment _add_ a new object (using admin). This isn't a Django problem; it's a basic Python rule that any __repr__() method must return a string. For that reason, you should make sure that __repr__() returns a string, not an object instance or (in the case of date fields) a datetime object. Adrian -- Adrian Holovaty holovaty.com | djangoproject.com | chicagocrime.org
Re: What is save to use in __repr__
I ran into the same thing in a model with a ForeignKey to an auth.User object. In my __repr__ function I had: self.get_user() which would crash Django. In this instance, I wanted the username specifically, so I resorted to using: self.get_user().username I didn't try, but I think this would have worked as well: str(self.get_user()) I think the reason it crashed is because __repr__ should return a string, not an object, so if you explicitly "stringify" the object, you should be fine. In your case, try: str(self.someForeignKey().someThing()) -berto. On 1/30/06, Jan Rademaker <[EMAIL PROTECTED]> wrote: > > What is save to use __repr__ ? > > I've found out that I shoud use self.get_someForeignKey().someThing() > instead of self.someForeignKey.someThing(), but what about date fields > and such? > > Because, when I use self.someDateField in __repr__ django crashes the > moment _add_ a new object (using admin). > > The field is declard as: > datum = meta.DateTimeField(auto_now_add=True) > > Error: TemplateSyntaxError: Caught an exception while rendering. > > - janr > >
Re: FileField :: mx file size
On 1/30/06, Amit Upadhyay <[EMAIL PROTECTED]> wrote: > On 1/30/06, Maniac <[EMAIL PROTECTED]> wrote: > > > > Amit Upadhyay wrote: > > > > > It would be good to have a general django wide setting specifying the > > > maximum length allowed for POST data, if underlying server allowed it. > > > > Django just doesn't control these things. Read my answer in this very > > thread, it's http-server and browser that don't let it happen. > > > > Umm.. I have not done any research on this, but if that was the case > projects like http://sean.treadway.info/demo/upload/ and > http://encodable.com/filechucker/ wouldn't have been > possible. May be some webservers do not support it, and maybe there would be > some issues related to terminating connection if file size is larger, but > having this may be good. > > No time for investigating it right now, but existence of those projects > imply its doable, even if its a hack, atleast on some servers. > > True, but both of those projects use "AJAX", aka JavaScript. Remember, your trying to test the size of a file **before** it is uploaded. That would require doing the check on the client side - which requires **client side code;** namely, JavaScript. I'm not to up on JS myself and don't really know what capabilities it has in determining file size, if any, but I envision a simple script that intercepts the upload submit, checks the actual file size against the specified limit (perhaps in a hidden field or hard coded in script or even via an ajax request (but why?)) and either returns an error message or submits the upload. Of course, you still have those people who will turn JS off to bypass the limit, so you still want the server side code to validate file size and throw away oversized uploads. Perhaps not ideal, but its an inherit limit of http and (usually) for good reason. -- Waylan Limberg [EMAIL PROTECTED]
Re: editable in fields
Tim Terlegård wrote: To get a non-editable field you can do field = meta.CharField(editable = False). This is meant for hiding the field, right? Not to make it readonly in code? It means (as far as I know) that default manipulators will skip it and won't complain about their invalidity. So in effect it means that they are not intended to be edited by the user but rather set manually in the code. Think for example of some calculated field that can't be blank. You shouldn't show it to a user and want to calculate in, say, _pre_save. But if editable=False is not set then a manipulator would complain about it being empty before attempting to save it.
Re: Error when relating a model to another model more than once
To be complete, this is the Error page I get: TypeError at /admin/reg/languages/2/ 'bool' object is not callable Request Method: GET Request URL:http://localhost:8000/admin/reg/languages/2/ Exception Type: TypeError Exception Value:'bool' object is not callable Exception Location: /usr/lib/python2.4/site-packages/django/core/meta/__init__.py in bind, line 262 Template error In template /usr/lib/python2.4/site-packages/django/contrib/admin/templates/admin/change_form.html, error at line 52 Caught an exception while rendering. 42 {% block after_field_sets %}{% endblock %} 43 {% if change %} 44 {% if bound_manipulator.ordered_objects %} 45 {% trans "Ordering" %} 46 47 {% if form.order_.errors %}{{ form.order_.html_error_list }}{% endif %} 48 {% trans "Order:" %} {{ form.order_ }} 49 50 {% endif %} 51 {% endif %} 52 {% for related_object in bound_manipulator.inline_related_objects %}{% edit_inline related_object %}{% endfor %} 53 {% block after_related_objects %}{% endblock %} 54 {% submit_row bound_manipulator %} 55 {% if add %} 56 document.getElementById("{{ bound_manipulator.first_form_field_id }}").focus(); 57 {% endif %} 58 {% if bound_manipulator.auto_populated_fields %} 59 60 {% auto_populated_field_script bound_manipulator.auto_populated_fields change %} 61 62 {% endif %} Traceback (innermost last) * /usr/lib/python2.4/site-packages/django/core/template/__init__.py in render_node 724. 725. def render_node(self, node, context): 726. return(node.render(context)) 727. 728. class DebugNodeList(NodeList): 729. def render_node(self, node, context): 730. try: 731. result = node.render(context) ... 732. except TemplateSyntaxError, e: 733. if not hasattr(e, 'source'): 734. e.source = node.source 735. raise 736. except Exception: 737. from sys import exc_info ▶ Local vars Variable Value context [{}, {'forloop': {'parentloop': {}, 'last': True, 'counter': 2, 'revcounter0': 0, 'revcounter': 1, 'counter0': 1, 'first': False}, 'related_object': }, {'block': '>, , , , '>, , , , , , , , , , , , , , , , , , , , , , , , , , '>]>}, {}, {'has_delete_permission': True, 'bound_manipulator': , 'add': False, 'app_label': 'reg', 'form_url': '', 'change': True}, {'LANGUAGES': (('bn', 'Bengali'), ('cs', 'Czech'), ('cy', 'Welsh'), ('da', 'Danish'), ('de', 'German'), ('en', 'English'), ('es', 'Spanish'), ('fr', 'French'), ('gl', 'Galician'), ('is', 'Icelandic'), ('it', 'Italian'), ('ja', 'Japanese'), ('nl', 'Dutch'), ('no', 'Norwegian'), ('pt-br', 'Brazilian'), ('ro', 'Romanian'), ('ru', 'Russian'), ('sk', 'Slovak'), ('sr', 'Serbian'), ('sv', 'Swedish'), ('zh-cn', 'Simplified Chinese'), ('zh-tw', 'Traditional Chinese')), 'LANGUAGE_CODE': 'en-us'}, {}, {'perms': , 'messages': [], 'user': rfroger}, {'is_popup': False, 'original': English, 'object_id': '2', 'form': {'error_dict': {}, 'edit_inline': True, '_inline_collections': [, ], 'manipulator': , 'order_objects': [], 'data': {'abbreviation': 'en', 'name': 'English', 'languagetranslation.0.id': None, 'languagetranslation.0.translation': '', 'languagetranslation.0.translation_language_id': None, 'id': 2}, 'original': English}, 'title': 'Change language'}] exc_info node self [ '>, , , , '>, , , , , , , , , , , , , , , , , , , , , , , , , , '>] wrapped * /usr/lib/python2.4/site-packages/django/core/template/defaulttags.py in render 104. # boolean values designating first and last times through loop 105. 'first': (i == 0), 106. 'last': (i == len_values - 1), 107. 'parentloop': parentloop, 108. } 109. context[self.loopvar] = item 110. for node in self.nodelist_loop: 111. nodelist.append(node.render(context)) ... 112. context.pop() 113. return nodelist.render(context) 114. 115. class IfChangedNode(Node): 116. def __init__(self, nodelist): 117. self.nodelist = nodelist ▶ Local vars Variable Value context [{}, {'forloop': {'parentloop': {}, 'last': True, 'counter': 2, 'revcounter0': 0, 'revcounter': 1, 'counter0': 1, 'first': False}, 'related_object': }, {'block': '>, , , , '>, , , , , , , , , , , , , , , , , , , , , , , , , , '>]>}, {}, {'has_delete_permission': True, 'bound_manipulator': , 'add': False, 'app_label': 'reg', 'form_url': '', 'change': True}, {'LANGUAGES': (('bn', 'Bengali'), ('cs', 'Czech'), ('cy', 'Welsh'), ('da', 'Danish'), ('de', 'German'), ('en', 'English'), ('es', 'Spanish'), ('fr', 'French'), ('gl', 'Galician'), ('is', 'Icelandic'), ('it', 'Italian'), ('ja', 'Japanese'), ('nl', 'Dutch'), ('no', 'Norwegian'), ('pt-br', 'Brazilian'), ('ro', 'Romanian'), ('ru', 'Russian'), ('sk', 'Slovak'), ('sr', 'Serbian'), ('sv',
Re: Lookup by week
On 1/30/06, Jan Rademaker <[EMAIL PROTECTED]> wrote: > Would it make sense to implement such a thing? An alternative would be > to figure out at what date a week starts and ends and just use > pub_date__range, but that's not really the point. I'm just curious how > easy/hard it would be to implement. > > If the above isn't going to work, what's the best/most elegant way to > go, __range? Yeah, I'd suggest calculating the week start/end dates in your own code and using the __range lookup. Weeks are a bit of a messy problem -- different people have different definitions of when they start, etc. Adrian -- Adrian Holovaty holovaty.com | djangoproject.com | chicagocrime.org
editable in fields
To get a non-editable field you can do field = meta.CharField(editable = False). This is meant for hiding the field, right? Not to make it readonly in code? To me hidden is more intuitive than editable. I thought that when so much was renamed anyway... :) Tim
What is save to use in __repr__
What is save to use __repr__ ? I've found out that I shoud use self.get_someForeignKey().someThing() instead of self.someForeignKey.someThing(), but what about date fields and such? Because, when I use self.someDateField in __repr__ django crashes the moment _add_ a new object (using admin). The field is declard as: datum = meta.DateTimeField(auto_now_add=True) Error: TemplateSyntaxError: Caught an exception while rendering. - janr
Re: FileField :: mx file size
Amit Upadhyay wrote: No time for investigating it right now, but existence of those projects imply its doable, even if its a hack, atleast on some servers. Hm... Yes, they do it with some patched FCGI code.
Re: FileField :: mx file size
I think that this is doable, but need to be implemented by an Apache Hook/Handler. I've done some searches now for something related with mod_python, but not found anything, just a link with this same problem but using Perl. http://www.experts-exchange.com/Web/Web_Languages/CGI/Q_20862752.html Is a nice thing to be explored :) -Adriano Bonat On 1/30/06, Amit Upadhyay <[EMAIL PROTECTED]> wrote: > On 1/30/06, Maniac <[EMAIL PROTECTED]> wrote: > > > > Amit Upadhyay wrote: > > > > > It would be good to have a general django wide setting specifying the > > > maximum length allowed for POST data, if underlying server allowed it. > > > > Django just doesn't control these things. Read my answer in this very > > thread, it's http-server and browser that don't let it happen. > > > > Umm.. I have not done any research on this, but if that was the case > projects like http://sean.treadway.info/demo/upload/ and > http://encodable.com/filechucker/ wouldn't have been > possible. May be some webservers do not support it, and maybe there would be > some issues related to terminating connection if file size is larger, but > having this may be good. > > No time for investigating it right now, but existence of those projects > imply its doable, even if its a hack, atleast on some servers. > > > -- > Amit Upadhyay > Blog: http://www.rootshell.be/~upadhyay > +91-9867-359-701
Lookup by week
Hello, As far is I know it's not possible to query date fields for a cerain week number, eg. # Select all polls from week 5, 2006 polls.get_list( pub_date__year=2006, pub_date__week=5 ) Would it make sense to implement such a thing? An alternative would be to figure out at what date a week starts and ends and just use pub_date__range, but that's not really the point. I'm just curious how easy/hard it would be to implement. If the above isn't going to work, what's the best/most elegant way to go, __range? - janr
Re: FileField :: mx file size
On 1/30/06, Maniac <[EMAIL PROTECTED]> wrote: Amit Upadhyay wrote:> It would be good to have a general django wide setting specifying the> maximum length allowed for POST data, if underlying server allowed it.Django just doesn't control these things. Read my answer in this very thread, it's http-server and browser that don't let it happen.Umm.. I have not done any research on this, but if that was the case projects like http://sean.treadway.info/demo/upload/ and http://encodable.com/filechucker/ wouldn't have been possible. May be some webservers do not support it, and maybe there would be some issues related to terminating connection if file size is larger, but having this may be good. No time for investigating it right now, but existence of those projects imply its doable, even if its a hack, atleast on some servers.-- Amit UpadhyayBlog: http://www.rootshell.be/~upadhyay+91-9867-359-701
Re: FileField :: mx file size
Amit Upadhyay wrote: It would be good to have a general django wide setting specifying the maximum length allowed for POST data, if underlying server allowed it. Django just doesn't control these things. Read my answer in this very thread, it's http-server and browser that don't let it happen.
Re: error restoring project
Luke Skibinski Holt wrote: my hd failed the other day taking my db with it. I'm trying to restore my project now, but I'm getting this error in the admin app: ERROR: missing FROM-clause entry for table "auth_users" SELECT "blogs_blogs"."user_id","blogs_blogs"."title","blogs_blogs"."desc","blogs_blogs"."entriespp","blogs_blogs"."disabled" FROM "blogs_blogs" ORDER BY "auth_users"."username" DESC This is a known issue (http://code.djangoproject.com/ticket/930) dealing with the One-To-One relationship. The SQL is not creating the proper JOIN for the default ordering and you need to explicitly create an ordering='title' in your Meta class so that is isn't ordering based on a field in the other class. I'm assuming this isn't being fixed right now because the Subclassing system will replace it. -- --Max Battcher-- http://www.worldmaker.net/
Re: Template inheritance problem
Adrian,thank you for your reply. Yes, they both are in the same (board) directory. So, what shall I try to find out the reason why it doesn't work? Thank you Regards, L.
Re: FileField :: mx file size
On Sunday 29 January 2006 20:29, arthur debert wrote: > is there a way to check for this BEFORE the file > is uploaded? On Monday 30 January 2006 08:15, Adrian Holovaty wrote: > On 1/30/06, tonemcd <[EMAIL PROTECTED]> wrote: > > So does this happen *before* the file is uploaded - I mean when the > > submit button is pressed? It seems that Django will be getting > > information from the browser about the length of the data to be > > uploaded - and that's something I've always thought HTTP could not > > do... > > As far as the validator I posted, it does the validation *before* the > file is saved to the filesystem, when it is still in memory. But that is *after* all the file bits have been sent from the browser to the server. Right? My understanding is that the op wants to keep bandwidth under control by refusing to allow files greater than N bytes from ever entering memory. I've not yet found a way to do so; in the past, I've had to examine the size of the uploaded file after it has arrived at the server in its entirety. It works, but the consumed bandwidth was wasted for files that exceeded the size limit as the bits were simply tossed.
Re: FileField :: mx file size
Cheers Adrian, that's very handy to know.. Tone
validators combined
I have two text input fields in a form. At least one of them is required. My question is, how can I check for this? class MyManipulator = (formfields.Manipulator): def __init__(self, request=None): self.fields = [ formfields.TextField( field_name='apples', length=5, ), self.fields = [ formfields.TextField( field_name='pears', length=5, ), ] The problem is, I cannot use is_required=True. I can use another field and make it is_required=True and define another validator - this is what I'm doing now: self.fields = . formfields.TextField( field_name='basket', length=5, is_required=True, validator_list=[apples_or_pears], ), def apples_or_pears(self, field_data, all_data): if not all_data.get('apples', '') and not all_data.get('pears', ''): raise validator.ValidationError, _('bummer - what now?') This is a hack and has several drawbacks, such as it does not belong to the correct field, and is within an unexpected field's validator list. Thank you for your suggestions. Jiri
Re: mod_python installation problems
On 1/30/06, treelife <[EMAIL PROTECTED]> wrote: > Firstly, Debian 3.1 does not have a python2.4 version of the software > for Apache2.0 and when I install the 2.3 version Apache is giving me > some ".. Segmentation fault (11)". Try disabling the PHP Apache module. There are some known issues when PHP and mod_python are enabled at the same time. Adrian -- Adrian Holovaty holovaty.com | djangoproject.com | chicagocrime.org
Re: FileField :: mx file size
On 1/30/06, tonemcd <[EMAIL PROTECTED]> wrote: > So does this happen *before* the file is uploaded - I mean when the > submit button is pressed? It seems that Django will be getting > information from the browser about the length of the data to be > uploaded - and that's something I've always thought HTTP could not > do... As far as the validator I posted, it does the validation *before* the file is saved to the filesystem, when it is still in memory. Adrian -- Adrian Holovaty holovaty.com | djangoproject.com | chicagocrime.org
Re: assign or rename a variable in templates
Bryan, If you can handle Zope PageTemplates, the plugin from Stefan might be useful: http://www.zope.org/Members/shh/DjangoPageTemplates That allows for variable creation, but of course you don't get the django templates... Cheers, Tone
Re: FileField :: mx file size
Amit, I think that's more a progress 'barber pole' widget than something which will detect the length of a file and drop the connection if it's over a certain size (although it's hard to tell...) Cheers, Tone
Re: FileField :: mx file size
Adrian Holovaty wrote: def isSmallFile(self, field_data, all_data): if len(field_data["content"] > 1: # 10,000 bytes raise validators.ValidationError, "Please enter a smaller file." The trick here is to operate on field_data['content'] instead of field_data. File-upload fields are a special-case; they come across as a dictionary of {'filename', 'content-type', 'content'}. If you can do len(filed_data['content']) then this content is already has been uploaded on the server. At present the only way to know the size of data before they are uploaded is the Content-Length header sent by browsers. It also depends on server environment if it passes the control to a script before all the POST data has been received. I have an impression that Apache doesn't do it (thought I didn't check properly). And even if it did then the only way to prevent user to download anything is to break the connection. Which also generally impossible with http servers and would look ulgy in a browser that just sends data and doesn't expect anything sane to happen before upload finishes. So it would just show some network error message. I saw some discussions about specifying content size restriction in future browsers on a WHAT WG list (http://listserver.dreamhost.com/pipermail/whatwg-whatwg.org/) but it didn't end up in anything yet.
Re: image field update impossible.
thank you mary ! MEDIA_ROOT = '/home/greg/Projects/woot/' with 'postimgs' as the path in imagefield allow editing without error.
error restoring project
my hd failed the other day taking my db with it. I'm trying to restore my project now, but I'm getting this error in the admin app: ERROR: missing FROM-clause entry for table "auth_users" SELECT "blogs_blogs"."user_id","blogs_blogs"."title","blogs_blogs"."desc","blogs_blogs"."entriespp","blogs_blogs"."disabled" FROM "blogs_blogs" ORDER BY "auth_users"."username" DESC I'm getting it for both of my apps installed, both having one-to-one relationships with the user. Other models within those apps that don't have any relationship to the user (but to other models) are not failing. The steps I took to restore my project were create database manage.py init manage.py createsuperuser manage.py install admin manage.py install I'm using the latest revision as of this posting. Any ideas on what is going wrong? Did I miss something or do something in the wrong order? Luke Skibinski Holt
mod_python installation problems
Many people seem to have some difficulties configuring apache mod_python, at least initially to work with Django. I have also had these initial configuration difficulties, but have successfully configured it on my local box, which runs Suse 9.3. However, I am having real difficulties just installing it on my live web server which runs Debian 3.1 and Apache 2.0.x ( I believe x=.54) Firstly, Debian 3.1 does not have a python2.4 version of the software for Apache2.0 and when I install the 2.3 version Apache is giving me some ".. Segmentation fault (11)". I would be grateful for any help, as I really want to get back to doing some web development, but find myself spending all my time trying to get the requisite software installed. Sincerely
Re: image field update impossible.
u have to check your media_root that is in settings.py cause i think the problem is that he can't find the right path for your media files and the right path had to be media_root path+ the path of upload_to [EMAIL PROTECTED] wrote: > It doesn't solve the problem, i can't still override the image with a > new one, an invalid filename error rise. > see there http://img95.imageshack.us/img95/4536/see9za.png.