57 rows where "changetime" is on date 2013-03-23

View and edit SQL

Suggested facets: stage, component, type, severity, version, resolution, owner, easy, has_patch, needs_better_patch, needs_tests, needs_docs, created (date)

changetime (date)

  • 2013-03-23 · 57
id ▼ created changetime last_pulled_from_trac stage status component type severity version resolution summary description owner reporter keywords easy has_patch needs_better_patch needs_tests needs_docs ui_ux
7733 2008-07-12 15:59:25 2013-03-23 10:36:40 2019-06-24 01:38:15.014638 Design decision needed closed contrib.admin Cleanup/optimization Normal master wontfix Make clickable area of object links larger Per a [http://groups.google.com/group/django-developers/browse_frm/thread/84031974bb6e137f?hl=en django-dev] discussion, The objects in admin should provide more room for users to click on. This patch changes the link's anchor tag to display black with will fill the whole cell. nobody Mnewman nfa-someday 0 1 0 0 0 1
8135 2008-08-06 14:32:19 2013-03-23 10:48:35 2019-06-24 01:42:42.161124 Design decision needed closed Template system New feature Normal master wontfix Make RequestContext act as a dictionary When creating the context, using RequestContext to include context processors content, the next doesn't work: {{{ context = dict(example=value) context.update(RequestContext(request)) }}} instead I can use {{{ context = requestContext(request) context.update(dict(example=value)) }}} but I think that also allowing the first case would avoid some confusion. nobody garcia_marc   0 0 0 0 0 0
10701 2009-04-02 16:00:42 2013-03-23 07:50:09 2019-06-24 02:11:18.078920 Design decision needed closed Core (Management commands) New feature Normal 1.0 wontfix manage.py runfcgi dies in response to SIGHUP/SIGUSR1, disallows log rotation Thanks to #6687, daemonized fastcgi processes can now log their stderr and stdout. In our environment we would like to capture these logs, and rotate them (via the standard method of using logrotate). However, from looking in the code, it appears that runfcgi does not support closing/reopening its log file (which would allow logrotate to rotate the log out from under it). Sending a SIGUSR1 or SIGHUP to the master fcgi PID simply kills it instead of causing it to do this. Unless I'm missing something, it looks like the only way to do this at the time is to stop and then restart the processes, which is unacceptable in a production environment. A necessary patch would handle a SIGUSR1 and/or SIGHUP to close() and then re open() the files specified in errlog and outlog. I can write a patch if necessary using the python signal module. nobody robbyd runfcgi, fastcgi, SIGHUP, log rotation 0 0 0 0 0 0
10720 2009-04-03 13:33:15 2013-03-23 00:19:48 2019-06-24 02:11:30.505600 Design decision needed closed Database layer (models, ORM) New feature Normal master wontfix OrderedManyToMany API Implementation In response to my post on django-users, Russ Magee suggested I open a ticket. http://groups.google.com/group/django-users/browse_frm/thread/2593060c8f3958ed# This patch uses the standard python list interface to provide a very simple API for keeping a singly-ordered many-to-many relation field. More info, plus a simple example and unit-tests here: http://www.aryehleib.com/MutableLists.html nobody Aryeh Leib Taurog <vim@aryehleib.com> Ordered ManyToMany 0 1 0 1 1 0
11007 2009-05-04 20:23:23 2013-03-23 00:10:20 2019-06-24 02:14:41.019500 Design decision needed closed Database layer (models, ORM) New feature Normal master wontfix Add django.db.Model.changes(self) that returns a dict of differences The attached patch adds a method to the Model class that returns a dict of differences between the current instance and the values stored in the database. nobody dennis@kaarsemaker.net   0 1 1 0 0 0
11446 2009-07-09 11:25:42 2013-03-23 11:39:52 2019-06-24 02:19:21.803916 Design decision needed closed GIS New feature Normal 1.0 wontfix Adding dynamic label to GIcon It would be better if anybody can add a property like label to the GIcon, so that a label can be displayed with the custom icon. nobody anonymous   0 0 0 0 0 0
11999 2009-10-08 21:23:10 2013-03-23 11:42:05 2019-06-24 02:25:23.331293 Design decision needed closed GIS New feature Normal   wontfix Support for Google Maps encoded geometries When displaying polygons and polylines with GMaps, performance can be massively improved by using encoded geometries. It would be great for GeoDjango to support this: * a module to encode geometries and return a dictionary, so people can JSON-encode and send to the client, embed in templates, etc. * use encoded geometries in !GoogleGeoAdmin Resources: * http://code.google.com/apis/maps/documentation/polylinealgorithm.html -- official docs * http://facstaff.unca.edu/mcmcclur/GoogleMaps/EncodePolyline/ -- detailed discussion, examples, links to implementations in various languages * http://wyattbaldwin.com/2007/04/19/google-maps-encoded-polylines/ -- Python implementation. GPLv3 licensed, and code isn't tidy. rcoup rcoup gmaps 0 0 0 0 0 0
12306 2009-12-03 14:37:42 2013-03-23 22:16:37 2019-06-24 02:28:41.003363 Design decision needed closed Forms New feature Normal 1.1 wontfix Checkbox (label, field) ordering It is currently very difficult to reorder the default layout of fields: {{{ <label for="x" ...>My Label:</label> <input id="x".../> }}} For checkboxes, this should be either: {{{ <input id="x".../> <label for="x"...>My Label</label> }}} or {{{ <label><input.../> My Label</label> }}} At the moment, in a template there's no way of knowing if a field is a !BooleanField (in order to special case the html), and I don't know how I'd edit widgets etc. to override the current layout. This has been touched on as a sore point elsewhere: http://www.mail-archive.com/django-developers@googlegroups.com/msg21887.html http://stackoverflow.com/questions/572263/ ---- Where would be the best place to apply a patch for this? - At the form level: as a new argument to forms/forms.py:_html_output, called 'checkbox_row' which for e.g. table rows would look like: {{{ u'<tr><td colspan="2">%(field)s%(label)s%(help_text)s%(errors)s</td></tr>' }}} - At the field level: as a new boolean function 'is_checkbox'? - Both of the above? nobody EoghanM label checkbox BooleanField 0 0 0 0 0 0
12789 2010-02-05 08:46:48 2013-03-23 13:26:20 2019-06-24 02:33:52.807755 Design decision needed closed Core (Cache system) Bug Normal master duplicate ConditionalGetMiddleware behavior improvement. This patch is intended to improve the behavior of the django.middleware.http.ConditionalGetMiddleware? middleware. When a particular django view has at least two decorators that work with redirecting behavior, Firefox will send a ETag that seems to be valid since the response is nothing more than a Location redirect, thus the content-length is 0 and the ETag is valid for the two steps created by the two decorators that redirect the user to different pages (Example decorator : @login_required and another that acts very similarly). The bug is that the browser will receive a Not Modified header once the users passes the first decorator when he gets redirected by to the original view. The FIX: The fix changes the behavior of the django.middleware.http.ConditionalGetMiddleware? to not send a 304 response code if the content-length is 0 or undefined. I'm not sure if it's proper, but it seems good of a fix and it actually fixes my issue. So, I just wanted to shared it, hopefully it can make it in the code-base. This fix will require a little bit of documentation change to explain the extra condition. nobody penzoil ETag, ConditionalGetMiddleware, patch_response_headers 0 1 0 1 0 0
12846 2010-02-11 23:23:32 2013-03-23 18:52:21 2019-06-24 02:34:33.021348 Accepted closed Core (Other) New feature Normal master wontfix [PATCH] Support for compiled pyc files I am using django bundled with my app in an RPM file, and that RPM only ships pyc files, not py code. Version 1.0.4 had problems with that, and although I haven't tried later versions, browsing the code in trunk it looks like it is still an issue. I will attach a small patch that fixes it. Thanks, David Parker nobody dap   0 1 1 0 0 0
13313 2010-04-09 19:33:09 2013-03-23 00:38:09 2019-06-24 02:39:45.152288 Accepted new Database layer (models, ORM) Bug Normal master   Custom Default Manager with extra __init__ arguments fails if model is used in a ManyToManyField Suppose I have the following classes: {{{ class SpecialManager(models.Manager): def __init__(self, special_arg): self.special_arg=special_arg def random_method(): pass class SpecialModel(models.Model): objects = SpecialManager('hello') pass class InnocentModel(models.Model): my_friends = models.ManyToManyField(SpecialModel) }}} This is all well and fine, until I actually try to use the model: {{{ >>> obj = InnocentModel.objects.all()[0] >>> obj.my_friends Traceback (most recent call last): File "<console>", line 1, in <module> File "d:\code\env\src\django\django\db\models\fields\related.py", line 716, in __get__ reverse=False File "d:\code\env\src\django\django\db\models\fields\related.py", line 469, in __init__ super(ManyRelatedManager, self).__init__() TypeError: __init__() takes at least 2 arguments (1 given) }}} This is because related.py creates a subclass of the manager's default model for relation purposes: http://code.djangoproject.com/browser/django/trunk/django/db/models/fields/related.py?rev=12908#L706 {{{ # Dynamically create a class that subclasses the related # model's default manager. rel_model=self.field.rel.to superclass = rel_model._default_manager.__class__ RelatedManager = create_many_related_manager(superclass, self.field.rel) manager = RelatedManager( model=rel_model, core_filters={'%s__pk' % self.field.related_query_name(): instance._get_pk_val()}, instance=instance, symmetrical=self.field.rel.symmetrical, source_field_name=self.field.m2m_field_name(), target_field_name=self.field.m2m_reverse_field_name(), reverse=False ) }}} This manager is then instantiated, and calls super(): http://code.djangoproject.com/browser/django/trunk/django/db/models/fields/related.py?rev=12908#L469 {{{ 465 class ManyRelatedManager(superclass): 466 … nobody CBWhiz@gmail.com ManyToManyField, Manager 0 0 0 0 0 0
13401 2010-04-22 20:45:20 2013-03-23 21:45:50 2019-06-24 02:40:41.554850 Design decision needed closed Forms Bug Normal 1.1 wontfix Forms RadioInput produces invalid HTML django.forms RadioInput widget produces invalid HTML. The input tag is rendered inside the label tag which is semantically invalid. It makes it difficult to apply styles to the RadioInput and RadioSelect (multiple radioinputs) widget. The problem can be fixed by modifying the last row from RadioInput widgets __unicode method: original: return mark_safe(u'<label%s>%s %s</label>' % (label_for, self.tag(), choice_label)) fixed: return mark_safe(u'<label%s>%s</label>%s' % (label_for, choice_label, self.tag())) nobody pajatso   0 0 0 0 0 0
13581 2010-05-20 21:52:33 2013-03-23 11:32:51 2019-06-24 02:42:37.924363 Design decision needed closed GIS Cleanup/optimization Normal 1.1 wontfix improve naming of default coordinates for GeoModelAdmin¶ The GeoModelAdmin class currently has properties of default_lon and default_lat. While these are reasonable names when using an unprojected coordinate system, they are misleading if a subclass with another projection is used (such as the spherical mercator projection used by the class OSMGeoAdmin). This may lead to confused users, for example: http://code.djangoproject.com/ticket/11094. I would suggest changing the default coordinate names to default_x and default_y. nobody tylere GeoModelAdmin 0 0 0 0 0 0
13683 2010-06-02 11:28:19 2013-03-23 13:12:55 2019-06-24 02:43:42.844197 Design decision needed closed contrib.admin Cleanup/optimization Normal 1.2 needsinfo Don't instantiate empty inline form when max_num is 0 Django 1.2 added dynamic addition of Inline forms in the admin interface. In order to support this functionality, InlineAdminFormSet constructs an extra InlineAdminForm using self.formset.empty_form. However, we have an Inline that displays a fixed set of controls and should never be constructed without the appropriate parameters. The "Add another" javascript control can be hidden by specifying max_num = 0 on the Inline class. However, this does not prevent InlineAdminFormSet attempting to construct the blank form object. I am attaching a patch that changes this behaviour. dalorin dalorin sprintdec2010 0 1 0 1 0 1
13688 2010-06-02 19:54:36 2013-03-23 00:25:14 2019-06-24 02:43:46.124813 Design decision needed closed Database layer (models, ORM) New feature Normal 1.2 wontfix Ability to specify the field type for extra select {{{ Entry.objects.extra(select={'is_recent': "pub_date > '2006-01-01'"}) }}} should look like: {{{ Entry.objects.extra(select={BooleanField('is_recent'): "pub_date > '2006-01-01'"}) }}} Ideally this would map over the existing fields. The idea is that some of the extra fields might need preparation before using them in python (like custom fields). This is done by to_python() method, but for extra fields there's no such thing. They will always contain the raw DB data. nobody jonozzz extra select field types 0 0 0 0 0 0
13899 2010-07-06 18:25:18 2013-03-23 07:34:14 2019-06-24 02:46:10.039043 Design decision needed closed GIS New feature Normal 1.2 wontfix Geolocation Support for GeoDjango admin This adds geolocation support to GeoDjango admin maps, using the W3C Geolocation Standrad and as a fallback, Google Gears. Tested under Firefox 3.2, Chrome 5 and Safari 5. Patch works fine with the default settings, but it might fail if there is a custom projection applied. I am not confident enough with OpenLayers to make the script more aware of the map's current settings. More information on this here: http://taylanpince.com/blog/posts/adding-geolocation-support-to-geodjango-admin/ nobody taylanpince   0 1 0 0 0 0
13917 2010-07-09 22:32:28 2013-03-23 14:00:09 2019-06-24 02:46:21.538319 Accepted new contrib.admin New feature Normal 1.2   Multiple popup window feature of related objects popup through id_to_windowname In an app that has a cascading set of objects we required the ability to have multiple popup windows. In tabbed browsers this can work out well, especially when the tabs can cascade like the firefox tree-style view. To bring this support to django all that is required is an override for the two methods id_to_windowname and windowname_to_id at line 18 of django/contrib/admin/media/js/admin/RelatedObjectLookups.js, such as this: {{{ var uniqueness=(new Date()).valueOf(); function id_to_windowname(text) { text = text.replace(/\./g, '__dot__'); text = text.replace(/\-/g, '__dash__'); return uniqueness + text; } function windowname_to_id(text) { text = text.replace(uniqueness, ''); text = text.replace(/__dot__/g, '.'); text = text.replace(/__dash__/g, '-'); return text; } }}} This trivial change to the 'name' of the window allows multiple caller-callee connections, even involving the same models, as used as the id. Dave nobody davidcooper   0 1 1 0 0 1
13984 2010-07-23 07:05:44 2013-03-23 22:09:02 2019-06-24 02:47:04.384510 Design decision needed closed Template system New feature Normal master duplicate template.loader.select_template should return template_name as well. The whole purpose of this function is to find out which possible location of a certain template is valid. It make sense to return the compiled template as well as its location. Here's a common use case: A template is found by select_template, however, this template extends an "base.html" whose is located in the same directory, which is unknown until runtime as well. So the extends tag fails. If select_template returns the location of the template it finds, then this info could be passed into an context and thus make the extends tag useful again. JoshStegmaier DaNmarner   0 1 0 0 0 0
14121 2010-08-17 00:10:19 2013-03-23 08:18:51 2019-06-24 02:48:31.658767 Design decision needed closed Database layer (models, ORM) New feature Normal 1.2 wontfix Provide object's primary key as a hint for database router method db_for_read() when filtering objects by primary key '''Problem:''' Setting up a custom database router with db_for_read() method to allow for sharding with MySQL. Need to be able to control what database an object is read from on a per-object basis. [[BR]] [[BR]] '''Example Case:''' I want to retrieve User objects from the database cluster as follows: * If the username starts with a number, read from slave_db_1 * If the username starts with a letter, read from slave_db_2 [[BR]] Currently, the only way to achieve this behavior is to specify "using" on each condition, and per each call to the database. [[BR]] [[BR]] '''Proposed Solution:''' [[BR]] [[BR]] When filtering objects by primary key, pass the primary key to db_for_read() as a hint. [[BR]] [[BR]] '''Example Code:''' {{{ #View Code: """Assume username is pk for User model""" users = User.objects.get(username="testuser") #DB Router Code class UserReadRouter(object): def db_for_write(self, model, **hints): return 'master' def db_for_read(self, model, **hints): """Assume pk of User object is a string""" if model.__name__ == "User" and 'object_pk' in hints: if hints['object_pk'][0].isalpha(): return 'slave_db_2' else: return 'slave_db_1' }}} [[BR]] [[BR]] The end result is that "testuser" is retrieved from slave_db_1. nobody sandy@mahalo.com router, hints 0 0 0 0 0 0
14186 2010-08-28 21:22:12 2013-03-23 11:36:18 2019-06-24 02:49:13.559620 Design decision needed closed GIS New feature Normal master wontfix Adding GDirections wrapper to overlays.py Today I needed to add support to GoogleMap GeoDjango abstraction so it could draw directions in a map. So I coded GDirections class, modified gmap.py to add a directions parameter and modified the template. Directions are drawn in a path within the map, adding a directions div and getting direction steps would be trivial with this. Take in mind only one GDirections object is passed to the GoogleMap object, as only one route will be drawn in the map. You can find an example of use in the Doc strings of the class: {{{ from django.shortcuts import render_to_response from django.contrib.gis.maps.google.overlays import GDirections def sample_request(request): route = GDirections(LineString(POINT(40.44 -3.77), POINT(42.33 -3.66)) return render_to_response('mytemplate.html', {'google' : GoogleMap(directions=route)}) }}} You can still pass other overlays and they will get drawn in the map. Just be careful with zoom, because GDirections will set automatically the zoom to show the whole path. I hope somebody finds this useful, Best regards Miguel Araujo maraujop maraujop   0 1 0 0 0 0
14332 2010-09-23 16:28:09 2013-03-23 08:01:08 2019-06-24 02:50:47.412310 Design decision needed closed Core (Management commands) New feature Normal 1.2 wontfix manage testserver and manage runserver should have similiar address:port options Hi, $ ./manage.py runserver -h Usage: ./manage.py runserver [options] [optional port number, or ipaddr:port] Starts a lightweight Web server for development. [...] $ ./manage.py testserver -h Usage: ./manage.py testserver [options] [fixture ...] In order to get the testserver to listen on a particular port, you need to specify the --addrport option. It would be useful if that option was also understood by the normal runserver command. Or if testserver interpreted a bare port number as the runserver command does. fheinz wildfire@progsoc.org runserver, testserver 0 1 1 0 0 0
14586 2010-10-29 17:46:00 2013-03-23 21:30:00 2019-06-24 02:53:32.053047 Accepted new Internationalization Cleanup/optimization Normal 1.2   Make interpolate JS function not require the named parameter. It is completely unnecessary and annoying to have to specify an additional parameter in Javascript when using a dictionary to fill in the strings. Switching back and forth between python and js, this can be often times forgotton, with bad results. With this fix, compatibility is maintained with all current utilizations. nobody aminland   0 1 0 0 1 0
14874 2010-12-09 10:01:14 2013-03-23 17:01:52 2019-06-24 02:56:36.768627 Design decision needed closed contrib.admin Bug Normal master wontfix remove js-confirms on changelist see http://code.djangoproject.com/browser/django/trunk/django/contrib/admin/media/js/actions.js: just delete line 104 to 126. why? because[[BR]] a) trusted editors should know what they do.[[BR]] b) triggering the change-function of form-fields may not capture all changes.[[BR]] c) and probably most important: we don´t do that on the change-form. it´s all about coherence. nobody sehmaschine@gmail.com   0 0 0 0 0 1
15313 2011-02-16 11:05:34 2013-03-23 07:45:29 2019-06-24 03:01:23.967818 Design decision needed closed Database layer (models, ORM) Uncategorized Normal 1.2 wontfix Oracle DB backend should not overwrite NLS_LANG I already define the correct NLS_LANG for the database: os.environ["NLS_LANG"] = "SPANISH_SPAIN.WE8ISO8859P1" but then, it is overwritten here: http://code.djangoproject.com/browser/django/trunk/django/db/backends/oracle/base.py Workaround: set NLS_LANG after database definition. nobody manelclos oracle 0 0 0 0 0 0
15319 2011-02-16 15:15:53 2013-03-23 09:04:08 2019-06-24 03:01:27.784283 Design decision needed closed Database layer (models, ORM) New feature Normal master wontfix _get_next_or_previous_by_FIELD populated from Meta Right now date field and datetime fields use contribute_to_class to add a 'get_next_by_<field name>'. If you look at the code in _get_next_or_previous_by_FIELD, it is pretty general, and should work with any field with a stable sort order. Maybe instead of creating get_next_by_FIELD and get_previous_by_FIELD based on field type, we could create them based on a Meta attribute in the class. I'm thinking we could call the Meta attribute seekable_fields or something like that. For instance: {{{#!python class Person(models.Model): first_name = models.CharField(max_length=255) last_name = models.CharField(max_length=255) class Meta: seekable_fields = ('last_name', ) }}} The resulting person objects would have methods named 'get_next_by_last_name' and 'get_previous_by_last_name'. I think the benefit here is that we are able to specify the fields we want to seek by, instead of only being able to use date fields. Eventually, we could drop the contribute_to_class methods for the date field, since it is just a special case that could be handled by including the name of any date field you wanted to be able to sort by. nobody alexr   0 0 0 0 0 0
15562 2011-03-06 19:11:48 2013-03-23 08:11:06 2019-06-24 03:04:04.824775 Design decision needed closed contrib.staticfiles New feature Normal 1.3-rc wontfix include_patterns in staticfiles finder For completeness and also to be able to include only files like *.gif or the like, I've added include_patterns to the list() method in the staticfiles finders. I'm attaching a patch for that. nobody Kronuz   0 0 1 1 1 0
15616 2011-03-15 15:13:14 2013-03-23 13:04:54 2019-06-24 03:04:39.482147 Design decision needed closed Core (Other) Cleanup/optimization Normal master needsinfo django.core.exceptions.ValidationError is poorly implemented and documented django.core.exceptions.ValidationError has a horrific interface. It's publicly accessible attributes differ depending upon how the object is created, as do the pre/post conditions for it's publicly accessible methods. If constructed with a dictionary of error messages, then it gets an attribute 'message_dict'. If constructed with a string, or a list of strings, then it gets an attribute 'messages'. The documentation on Model.clean() talks about ValidationError, but does not mention this at all, and instructs the user to look at messages_dict. [1] If constructed with a dictionary of error messages, then the method update_error_dict(self, error_dict) will happily accept error_dict=None - it checks for and handles this condition. If constructed with a string or list, then not supplying error_dict is an error. This sort of inconsistent behaviour is hard to explain. At the least, this behaviour should be documented clearly, but ideally the interface should be consistent, regardless of how it is constructed. [1] http://docs.djangoproject.com/en/dev/ref/models/instances/#django.db.models.Model.clean nobody tomevans223 validation validationerror 0 0 0 0 0 0
15641 2011-03-18 13:06:43 2013-03-23 11:25:41 2019-06-24 03:04:55.607003 Design decision needed closed File uploads/storage New feature Normal master wontfix Allow FileSystemStorage(base_url=None) to disable url access changeset:9946 changed the behaviour of passing base_url=None to the FileSystemStorage contructor. Before the change it would disable url access whereas now it enables access through MEDIA_URL The attached patch restores the old behaviour. nobody Harm Geerts <hgeerts@osso.nl>   0 1 0 0 1 0
15711 2011-03-29 18:17:40 2013-03-23 13:57:43 2019-06-24 03:05:43.804078 Design decision needed closed Database layer (models, ORM) New feature Normal 1.2 wontfix Abstract model managers should return subclass instances Suppose a base: {{{ class Document(Model): title = CharField() class Meta: abstract = True class Presentation(Document): num_kitten_pictures = IntegerField() class Spreadsheet(Document): num_pie_charts = IntegerField() }}} It would be useful if Documents.objects.filter(title='x') returned instances of both Presentation and Spreadsheet whose title was 'x'. I'd expect support only for base-class fields. This should be safe, since inheritance should follow an is-a relationship. nobody jdunck pony 0 0 0 0 0 0
15729 2011-03-31 15:52:27 2013-03-23 13:24:24 2019-06-24 03:05:55.252619 Design decision needed closed Core (Cache system) Bug Normal 1.3 needsinfo Cache middleware issues after restart when using mod_wsgi with PyLibMC I'm running into a cache error using the PyLibMC cache backend in conjunction with mod_wsgi. I have a strong feeling it's related to #15655 and is probably an issue for the pylibmc maintainers and not Django. I do feel, however, that the issue is worth reporting here because it is occurring on a pretty standard installation with the recommended web server (mod_wsgi). It also makes me wonder whether more robust error handling might need to be built into the PyLibMC backend to account for ways things could go wrong. So, the gist of the problem: I have memcached running on a separate server and PyLibMC as the memcached backend. Things work great...until I restart memcached. After doing so I get the following traceback: {{{ Traceback (most recent call last): File "/var/www/****/src/django/django/core/handlers/base.py", line 89, in get_response response = middleware_method(request) File "/var/www/*****/src/django/django/middleware/cache.py", line 143, in process_request cache_key = get_cache_key(request, self.key_prefix, 'GET', cache=self.cache) File "/var/www/*****/src/django/django/utils/cache.py", line 190, in get_cache_key headerlist = cache.get(cache_key, None) File "/var/www/*****/src/django/django/core/cache/backends/memcached.py", line 58, in get val = self._cache.get(key) UnknownReadFailure: error 7 from memcached_get: UNKNOWN READ FAILURE }}} Restarting Apache resolved this error. The issue does not occur when using MemcachedCache backend. nobody jsdalton   0 0 0 0 0 0
16089 2011-05-24 17:29:45 2013-03-23 15:56:38 2019-06-24 03:09:51.862596 Design decision needed closed contrib.admin New feature Normal 1.3 fixed Adding/Removing admin Inlines on the fly Inlines are very useful for creating/updating related objects at the same time. However, sometimes you only want to create/update one of the objects instead of both. For example, when creating a User you may not want to add the associated content in a related profile model, but want to do so later. The current implementation defines and instantiates the inlines when a admin.ModelAdmin class is instantiated, but does not allow addition/removal of more inlines anywhere(or anytime) else. I propose two functions register_inline_instance() and unregister_inline_instance() that allow inlines to be added and removed from a model on the fly. Example usage: {{{ class UserAdmin(admin.ModelAdmin): inlines = [StudentInline,] ... def add_view(self, request, form_url='', extra_context=None): self.unregister_inline_instance(StudentInline) return super(UserAdmin, self).add_view(request, form_url, extra_context) def change_view(self, request, form_url='', extra_context=None): self.register_inline_instance(StudentInline) return super(UserAdmin, self).change_view(request, form_url, extra_context) ... }}} nobody sheep2   0 1 1 1 1 0
16112 2011-05-28 09:04:00 2013-03-23 13:43:31 2019-06-24 03:10:06.530256 Design decision needed closed Database layer (models, ORM) Bug Normal 1.3 fixed Excluding some value in a related model excludes objects without that model I have models that look something like: {{{#!python class Group(models.Model): name = models.CharField(max_length=100) activity_category = models.ForeignKey('ActivityCategory', null=True, blank=True, ) class ActivityCategory(models.Model): name = models.CharField(max_length=50) }}} I'm trying to find each Group that doesn't have some ActivityCategory: {{{#!python qobj = Q(activity_category__name='Dorm') queryset = groups.models.Group.objects.filter(~qobj) }}} Intuitively, I would expect this to find both groups with no ActivityCategory and groups where the ActivityCategory exists and it isn't "Dorm". Reading https://docs.djangoproject.com/en/dev/topics/db/queries/#lookups-that-span-relationships seems to support this interpretation --- "If you are filtering across multiple relationships and one of the intermediate models doesn't have a value that meets the filter condition, Django will treat it as if there is an empty (all values are NULL), but valid, object there." Unfortunately, the ORM generates a query using INNER JOIN instead of LEFT OUTER JOIN, so my Group with a null ActivityCategory will not be returned. I can work around this by explicitly indicating that null is possible, but it's annoying and unintuitive: {{{#!python qobj = Q(activity_category__isnull = True) | ~(Q(activity_category__name='Dorm') | Q(activity_category__name='FSILG')) queryset = groups.models.Group.objects.filter(qobj) }}} nobody adehnert   0 0 0 0 0 0
16226 2011-06-10 18:56:44 2013-03-23 09:08:10 2019-06-24 03:11:22.897597 Design decision needed closed contrib.admin Bug Normal master wontfix [patch] prepopulated_fields only work with slug fields and break on non-char fields While analyzing ticket #15922, it became apparent that `prepopulate_from` field in Django admin unconditionally assumes that all prepopulated fields are of type `SlugField` and thus the `URLify` function in javascript is always called. In addition it is assumed that each target field has a `max_length` attribute. However, only fields that inherit from a CharField have `max_length`. For all other field types, the javascript just fails. Besides samples in #15922, here is another sample case in which any text in the `starting_price` field will be prepopulated in `price` field slugified: {{{ # models.py from django.db import models class Book(models.Model): starting_price = models.CharField(max_length=100) price = models.CharField(max_length=100) # admin.py from django.contrib import admin import models class BookAdmin(admin.ModelAdmin): prepopulated_fields = { 'price': ['starting_price'], } admin.site.register(models.Book, BookAdmin) }}} peter_kese peter_kese   0 1 0 0 0 1
16505 2011-07-24 03:41:44 2013-03-23 09:03:43 2019-06-24 03:14:23.139376 Accepted new Database layer (models, ORM) New feature Normal 1.3   Consider a different interface for get_next_by_FOO and get_previous_by_FOO Recently I was looking to use these methods along with a deferred query. As I quickly found out, it's impossible - in order to defer I'll have to implement all of the get_next_by logic separately. That's because these methods are tied directly to a model. You can't defer because it's not part of a manager. I'm no django expert (yet), but I'm wondering if these methods might be better suited, in the long term, off of .objects instead - so something like the following would work: {{{ some_blog = Blog.objects.get(pk=1) # Get the next one by passing in a model older_blog = Blog.objects.get_next_by_date_added(some_blog) # Get the next one by passing in a datetime object itself older_blog = Blog.objects.get_next_by_date_added(some_blog.date_added) # Get the next blog, but don't get pull out content because it's unnecessary. deferred_older_blog = Blog.objects.defer('content').get_next_by_date_added(some_blog) }}} Any thoughts here? nobody umbrae@gmail.com   0 0 0 0 0 0
16663 2011-08-20 23:06:18 2013-03-23 13:00:49 2019-06-24 03:16:05.826369 Design decision needed closed Database layer (models, ORM) Bug Normal 1.3 needsinfo Model field choices should accept empty iterator The base `Field` class constructor uses the following shorthand to initialize the `choices` for a field (line 96 of `django/db/model/fields/__init__.py`): {{{ self._choices = choices or [] }}} In Python, an empty iterator evaluates to False, so if we did: {{{ class MyModel(models.Model): my_field = models.IntegerField(choices=my_list_like_object) }}} `_choices` will default to an empty list if the `my_list_like_object` is permanently or temporarilly empty at time of class definition. However, the documentation clearly states the intent of allowing [https://docs.djangoproject.com/en/dev/ref/models/fields/#choices any iterable] as a `choices` kwarg: Finally, note that choices can be any iterable object -- not necessarily a list or tuple. This lets you construct choices dynamically. But if you find yourself hacking choices to be dynamic, you're probably better off using a proper database table with a ForeignKey. choices is meant for static data that doesn't change much, if ever. So I propose making a small change to the above line: {{{ if choices == None: self._choices = [] else: self._choices = choices }}} nobody danielnaab   0 1 0 0 0 0
17473 2011-12-28 16:08:48 2013-03-23 08:16:59 2019-06-24 03:24:52.922227 Design decision needed closed Database layer (models, ORM) New feature Normal 1.3 wontfix Add like and ilike lookups Add like and ilike [http://docs.djangoproject.com/en/dev/topics/db/queries/#field-lookups lookups]. Should be possible to do something like this: {{{ MyModel.objects.filter(field_name__like='xx%YY%zz') MyModel.objects.filter(field_name__ilike='xx%yy%zz') User.objects.filter(username__like='a%in') }}} Instead of: {{{ MyModel.objects.filter(field_name__regex='^xx.*YY.*zz$') }}} It is more quick and more readable. I create the next [http://pypi.python.org/pypi/django-like app], after read the next [http://stackoverflow.com/questions/8644146/django-query-how-to-write-where-field-like-10-8-0 question] I didn't understand why this is not coded nobody pmartin   0 1 0 0 0 0
17474 2011-12-28 17:14:42 2013-03-23 07:12:21 2019-06-24 03:24:53.554900 Design decision needed closed HTTP handling Bug Normal 1.4 wontfix Problem when the request doesn't have Content-Type. I don't know exactly why, but some ajax requests don't have Content-Type header. When this happens, when I try to get `request.POST`, I got the following exception: `'NoneType' object has no attribute 'startswith'`, raised by method `_load_post_and_files` at line 270 of `django/http/__init__.py`. The problem is that when the request doesn't have Content-Type, `self.META.get('CONTENT_TYPE', '')` returns `None`. To fix the problem I just did it: {{{ content_type = self.META.get('CONTENT_TYPE', None) or '' if content_type.startswith('multipart'): }}} nobody Marcelo Salhab Brogliato <msbrogli@vialink.com.br> content-type 0 1 0 0 0 0
17557 2012-01-17 19:40:58 2013-03-23 12:48:26 2019-06-24 03:25:46.352523 Design decision needed closed Generic views Cleanup/optimization Normal 1.4-alpha-1 wontfix Refactor form CBV to remove some unnecessary complexity While creating a CBV to work with inline formsets, I ran into a few places where copy-n-paste was the only way to adjust the behavior that I was after and keep the same feel of the internal Django CBV. This refactor removes the need to copy-n-paste out of `get` in order to add additional data by moving the generation of the `form` context data into `get_context_data`. nobody tswicegood   0 1 0 0 0 0
17781 2012-02-27 09:51:46 2013-03-23 20:33:45 2019-06-24 03:28:11.598289 Design decision needed closed Core (Cache system) Bug Normal master wontfix Django cache middleware does not set "Age" header as described in rfc2616 Django cache middleware set "Cache-Control: max-age=" and "Expires", but does not set "Age". This can be an issue, because "max-age" in frozen, so you can only rely on "Expires" for expiration time. Having the "Age" header set would allow the client to now when this page was cached by django. This feature would require adding a cache.age(key) function in all cache backends, which is far from trival, especially with backends such as memcached. nobody corentin.chary@gmail.com   0 0 0 0 0 0
17878 2012-03-12 13:25:04 2013-03-23 14:32:57 2019-06-24 03:29:15.563418 Design decision needed closed Database layer (models, ORM) New feature Normal 1.4-beta-1 wontfix Recursive order_with_respect_to Unsure of the version in which this occurs, as I'm running the development version. I have a simple a straightforward model that looks something like this: {{{ class Section(models.Model): pass class Passage(models.Model): section=models.ForeignKey(Section,related_name='passages') class Meta: order_with_respect_to = 'section' ANSWERS = ((0,''),(1,'A'),(2,'B'),(3,'C'),(4,'D'),(5,'E')) class Question(models.Model): passage=models.ForeignKey(Passage,related_name='questions') correct_answer = models.IntegerField(max_length=1,choices=ANSWERS) class Meta: order_with_respect_to = 'passage' }}} Note that I have omitted the properties of the model not pertinent to this question. Now, my problem is that although it is quite easy to get an ordered list of correct answers for a given passage: (p is a passage instance) {{{ p.questions.values_list('correct_answer',flat=True) }}} It is actually much more difficult than you'd expect to get the same answer set, now for the entire section instead. Essentially, I want the ordering on the question model to look like this: {{{ ordering = ['passage___order','_order'] }}} instead of the assumed: {{{ ordering = ['_order'] }}} However, even if I try to set this manually in the Meta class for Question, the actual ordering of a fetch like this: (where s is an instance of a section) {{{ Question.objects.filter(passage__section=s).values_list('correct_answer',flat=True) }}} The ordering that's returned here is quite useless, as it just sorts the questions based on the '_order' attribute with no regard to the fact that now it'd be much more useful to order by the parent passage's order_with_respect_to attribute first, and then this one in succession. As far as I know, it seems like setting the order_with_respect_to takes precedent over any custom ordering. However, now I don't really have a way of getting the data I want without looping through every passage and doing the joins in… nobody eric@lubin.us   0 0 0 0 0 0
17984 2012-03-26 15:08:24 2013-03-23 10:21:48 2019-06-24 03:30:23.868572 Design decision needed closed contrib.admin Bug Normal 1.4 wontfix admin list_filter security fix doesn't allow 'pk' lookups in query string Let's say I have a Payment model containing a received_by field that is a !ForeignKey to a Customer model. Now I add this to my Payment !ModelAdmin: {{{ list_filter = ["received_by"] }}} If I use a query string like "?received_by!__id!__exact=2" on the payment changelist view, there is no problem. This is the form that the admin itself uses for the filter. Using something more implicit like "?received_by!__id=2" works fine too. However, using "?received_by!__pk=2" raises !SuspiciousOperation. This was introduced by the security fix in r15031. The lookup_allowed method in admin/options.py explicitly makes an exception for "!__id" (or whatever the primary key attribute of the model is explicitly called), but it doesn't allow direct use of the 'pk' shortcut. The point of the security fix was to only allow lookups specified by list_filter. However, in its current form it also causes the query string syntax to diverge from the documented lookup syntax https://docs.djangoproject.com/en/1.4/topics/db/queries/#the-pk-lookup-shortcut . I suggest allowing use of the pk shortcut in the query string. I would be happy to write a patch for this if the developers agree it should be fixed. nobody 3point2   0 0 0 0 0 0
18000 2012-03-29 00:10:00 2013-03-23 12:29:35 2019-06-24 03:30:34.364172 Ready for checkin closed contrib.formtools New feature Normal 1.4 fixed Add Save and Goto Step to Form Wizard I've been working with the new form wizard in 1.4 and I discovered that apparently the goto form step feature was designed to not store data from the current step before going to the step specified. This seems like a situation that will easily lead to confusion and frustration. I suggest that in addition to the current cancel and goto step feature a save and goto step feature is needed. I created a patch that changes the behavior of wizard_goto_step into a save and goto, which seems safer to me, and then adds a wizard_cancel_step_goto. Please feel free to tinker with the code. I'm happy as long a I get a little more control over how data is handled when going to another step. Toodle-looooooooooo.......... creecode steph creecode form wizard save goto step 0 1 0 0 0 0
18185 2012-04-21 19:35:48 2013-03-23 08:51:41 2019-06-24 03:32:33.362288 Design decision needed closed Documentation Bug Normal master wontfix inlineformset_factory: formfield_callback does not execute for custom fields on modelForm hi, I'm on django 1.3 using class based generic views. So in my use case, i have a modelForm defined with custom fields. something along the lines of {{{ class MF_B(forms.ModelForm): stuff = forms.MultipleChoiceField(queryset=None, required=False) class Meta: model=B }}} as the code above implies, I want to populate the stuff choices here with a queryset. Model B also has a foreign key to Model A such that i would call inlineformset_factor like so: {{{ AB_Formset = inlineformset_factory(A, B, form=MF_B, max_num=3, formfield_callback=self.get_field_qs) }}} Just pretend get_field_qs is in scope and what we are trying to do here is something like: http://stackoverflow.com/questions/7310861/django-passing-parameters-to-inline-formset My expectation here is that get_field_qs would get called for ALL fields on the form (i.e. including my custom fields on the ModelForm). That doesn't happen. Looking at the django code, if i go to django/forms/models.py -> def fields_for_model -> line 146 It appears that formfield_callback will get executed for fields on model ONLY. Because the name of the parameter is formfield_callback, I would expect get_field_qs would be called for all the fields. As a result I feel this is a bug. nobody kush420 inlineformset_factory modelform formfield_callback custom fields 0 0 0 0 0 0
18242 2012-04-30 06:29:26 2013-03-23 14:03:07 2019-06-24 03:33:10.565802 Design decision needed closed Forms New feature Normal 1.3 wontfix Making widget type accessible in template Hi there, Is there a possibility to add a `type` attribute to form widgets? The goal is to make it easy for designers to apply different styles to each widget type while keeping the template code DRY. Django's current source code already does something similar for "<input>" based widgets such as `TextInput` and `DateInput` (though for a different purpose). These widgets have `input_type` attribute attached to them so we can already do: {{{ <ul> {% for field in form %} <li class = '{{ field.field.widget.input_type }}'> {{ field }} </li> {% endfor %} </ul> }}} But the above means we can only target `input` based widgets and not `select`, `textarea` etc. If we add a more generic `type` (e.g "select", "checkbox", "textarea") attribute to widgets, we'll be able to do: {{{ <ul> {% for field in form %} <li class = '{{ field.field.widget.type }}'> {{ field }} </li> {% endfor %} </ul> }}} Or even better (if we attach widgets to `BoundField`): {{{ <ul> {% for field in form %} <li class = '{{ field.widget.type }}'> {{ field }} </li> {% endfor %} </ul> }}} If it sounds like a good idea, I'd be more than willing to try writing a patch for this. nobody selwin   0 0 0 0 0 0
18265 2012-05-03 15:28:22 2013-03-23 20:19:18 2019-06-24 03:33:25.351180 Design decision needed closed contrib.auth New feature Normal master wontfix request.user cached value does not track request.session change During the Single-Sign On technique implementation I came to the need of replacing request.session with a new session or just setting request.session = None. The descriptor in AuthenticationMiddleware nevertheless does not remember the session for which it sets the request._cached_user. IMHO, the descriptor should cache the session key when it saves the _cached_user and allow _cached_user to be used only when the session key did not unchange. Currently everyone who changes the session during the request processing HAS to dig into the AuthenticationMiddleware internals and modify the _cached_user to get valid request.user. That's hard to maintain. I can offer the patch, but would like to get some response first. nobody Tuttle   1 0 0 0 0 0
18272 2012-05-04 17:03:38 2013-03-23 15:30:14 2019-06-24 03:33:29.849862 Design decision needed closed Forms Bug Normal master wontfix ModelChoiceField's queryset attribute persists between two form instances == Setup == Using the models from the tutorial: {{{#!python from django.db import models class Poll(models.Model): question = models.CharField(max_length=200) pub_date = models.DateTimeField('date published') class Choice(models.Model): poll = models.ForeignKey(Poll) choice = models.CharField(max_length=200) votes = models.IntegerField() }}} And the following form: {{{#!python from django import forms from polls.models import Poll class ChoiceForm(forms.Form): poll = forms.ModelChoiceField(queryset=Poll.objects.all()) def __init__(self, *args, **kwargs): """Change the choices attribute of the field `poll`.""" super(ChoiceForm, self).__init__(*args, **kwargs) queryset = self.fields['poll'].queryset choices = [(poll.pk, poll.pk) for poll in queryset] self.fields['poll'].choices = choices }}} == Bug == When doing so, the queryset gets "stuck" on what it was the first time the form was instanciated: added Poll objects don't show up in it, and deleted polls are still present. {{{#!python from datetime import datetime from polls.forms import ChoiceForm from polls.models import Poll Poll.objects.all().delete() # Just making sure. f = ChoiceForm() print f['poll'].as_widget() # '<select name="poll" id="id_poll">\n</select>' Poll.objects.create(question="What is your favourite colour?", pub_date=datetime.now()) f = ChoiceForm() print f['poll'].as_widget() # Expected: '<select name="poll" id="id_poll">\n<option value="1">1</option>\n</select>' # Result: '<select name="poll" id="id_poll">\n</select>' }}} == Workaround == One workaround is to copy the queryset before iterating over it: {{{#!python queryset = self.fields['poll'].queryset.all() }}} == Analysis == I believe the issue lies in {{{ModelChoiceField.__deepcopy__()}}}, in particular the line: {{{#!python result.queryset = result.queryset }}} {{{queryset}}} being a property, this ends up doing (among other things which I… Spark23 bmispelon modelchoicefield deepcopy 0 0 0 0 0 0
18502 2012-06-21 18:25:19 2013-03-23 08:50:03 2019-06-24 03:35:59.733076 Design decision needed closed Documentation New feature Normal 1.4 wontfix django.forms.models.fields_for_model() missing from online docs The `fields_for_model()` function is exported in `django.forms.models`, and has a nice docstring. It is useful e.g. for building custom Forms that combine fields from multiple models. Is there a reason why this function is not mentioned anywhere in the online docs? https://docs.djangoproject.com/search/?q=fields_for_model&release=1 nobody cberzan@gmail.com   0 0 0 0 0 0
19251 2012-11-05 14:08:09 2013-03-23 07:26:55 2019-06-24 03:44:02.412286 Design decision needed closed Forms New feature Normal master wontfix Normalize newlines from Textarea widgets Currently, newlines from Textarea may come in at least two different formats. Most browsers send them as CRLF. But when sent from JS (AJAX), it depends on the browser (Firefox and WebKit seem to send LF-only). I would suggest to normalize them to LF in all cases. This may also solve the issue about different char counting between client-side and server-side (http://groups.google.com/group/django-users/browse_thread/thread/3161244a2f119444/5fad134a89039d7b?lnk=gst&q=textarea%2C+max_length#5fad134a89039d7b). nobody claudep   0 1 0 0 0 0
19345 2012-11-22 19:16:50 2013-03-23 08:08:06 2019-06-24 03:45:03.042263 Design decision needed closed contrib.contenttypes Bug Normal master wontfix ContentType.objects.get_for_model shouldn't create missing content types This behavior is very unexpected, given the name of the method. In general, this isn't an issue because content types always exist — they're created in `post_syncdb`. I encountered the problem while debugging multiple databases issues involving `syncdb` (#16039). The `post_syncdb` handler that creates permissions looks for content types, creating them as a side effect. Since I was attempting to test that the content types weren't created, my test failed in a very unexpected fashion. nobody aaugustin   0 0 0 0 0 0
19433 2012-12-05 18:41:29 2013-03-23 12:23:09 2019-06-24 03:46:01.048979 Ready for checkin closed contrib.formtools Cleanup/optimization Normal 1.4 fixed Allow declaration form_list and condition_dict in WizzardView subclass, not as params to as_view Hi, every standard django CBV allows for declaring some attributesin the subclass to avoid passing them to as_view function in urls.py module. Unfortunately this does not apply to WizzardView, which does not allow a declaration of form_list nor condition_dict in a WizzardView subclass - each of these must be passed as an argument to as_view function. Since we need to subclass the view anyway (to define form processing function) AND the condition_dict contains a view logics, I believe the proper place for those attributes would be views.py - WizardView subclass, NOT urls.py - as_view params. example code provided here http://stackoverflow.com/q/13669033/146248 steph migajek cbv wizzard wizzardview 0 0 0 0 0 0
19512 2012-12-24 02:00:46 2013-03-23 13:34:51 2019-06-24 03:46:52.080982 Design decision needed closed Database layer (models, ORM) New feature Normal 1.4 wontfix Aggregate or Annotaion on previously Annotated values {{{ from django.db.models import Count, Sum, F myModel.objects \ .annotate(a_count=Count('field_a'), b_count=Count('field_b')) \ .annotate(total=F('a_count')+F('b_count')) \ .order_by('total') }}} This query is going to raise `AttributeError: 'ExpressionNode' object has no attribute 'split'` I don't find any method in django ORM to support such kind of operations (in which two fields are involved). nobody eye-see-you@live.com django, annotate, aggregate 0 0 0 0 0 0
19624 2013-01-16 17:33:59 2013-03-23 15:11:04 2019-06-24 03:48:05.755568 Design decision needed closed Generic views Cleanup/optimization Normal 1.5-beta-1 wontfix Generic Detail View context contains object twice. The (simplified and commented) code excerpt below, taken from django/views/generic/detail.py shows that the object is put into context twice, under different keywords. Is this intentional? The whole construction is most certainly not elegant and makes inheriting from SingleObjectMixin somewhat cumbersome as I have to explicitely set self.object if I don't have BaseDetailView in my class inheritance or override BaseDetailView.get() {{{ #!python class SingleObjectMixin(ContextMixin): def get_context_data(self, **kwargs): # this puts self.object into context under keyword context_object_name context[context_object_name] = self.object context.update(kwargs) return super(SingleObjectMixin, self).get_context_data(**context) class BaseDetailView(SingleObjectMixin, View): def get(self, request, *args, **kwargs): # this puts self.object into context under 'object' --> is this redundancy intentional? self.object = self.get_object() context = self.get_context_data(object=self.object) return self.render_to_response(context) }}} nobody mail@room2web.de generic view 0 0 0 0 0 0
20048 2013-03-14 08:02:37 2013-03-23 02:02:58 2019-06-24 03:52:37.476361 Ready for checkin closed contrib.flatpages Bug Release blocker 1.5 fixed Test failures in flatpages app when custom User model is in use contrib.flatpages contains a number of tests that depend upon authenticated users. However, these tests fail if your app uses a custom User model. These tests should be skipped if a custom User is defined. An audit of other contrib apps should be performed for similar failures. matiasb russellm   1 1 0 0 0 0
20060 2013-03-16 16:00:49 2013-03-23 02:02:58 2019-06-24 03:52:45.087188 Accepted closed contrib.sitemaps Bug Release blocker 1.5 fixed Test failures in sitemaps app when using custom User model contrib.sitemaps tests depends on the default User model. They all fail if run with a custom User model. A refactoring may be needed to fix/skip when appropriate. nobody matiasb   0 0 0 0 0 0
20087 2013-03-19 08:59:57 2013-03-23 10:03:06 2019-06-24 03:53:02.406134 Unreviewed closed contrib.syndication Uncategorized Normal 1.4 duplicate RSS updating issue Hi, some RSS readers cannot update the latest post because of the Last-Modified header not shown in Django version < 1.5. I find Django1.5 has fix this issue. Can you also make this in Django1.4(or even 1.3)? Please refer to the code below: class Feed(object): ... def __call__(self, request, *args, **kwargs): try: obj = self.get_object(request, *args, **kwargs) except ObjectDoesNotExist: raise Http404('Feed object does not exist.') feedgen = self.get_feed(obj, request) response = HttpResponse(mimetype=feedgen.mime_type) if hasattr(self, 'item_pubdate'): # if item_pubdate is defined for the feed, set header so as # ConditionalGetMiddleware is able to send 304 NOT MODIFIED response['Last-Modified'] = http_date( timegm(feedgen.latest_post_date().utctimetuple())) feedgen.write(response, 'utf-8') return response nobody julyzergcn1@gmail.com RSS updating 1 0 0 0 0 0
20115 2013-03-23 06:28:46 2013-03-23 10:05:52 2019-06-24 03:53:20.244901 Unreviewed closed Database layer (models, ORM) Bug Normal 1.5 duplicate Error with MySQL database using % in direct SQL Hello everybody! Привет всем русскоговорящим в отдельности! I found problem in Django 1.4.2 and current version in main branch of git repo. I try to do something like this: (found names that starts with string "Jo", like Joe,John etc) {{{ #!python from django.db import connection ... cursor=connection.cursor() cursor.execute('SELECT * from `test`.`test_table` WHERE name LIKE "Jo%";') }}} and take exception: '''Exception raised: <type 'exceptions.TypeError'> not enough arguments for format string''' looks very strange... after time, i found the root of problem. when execute() method start he invoke this: {{{ django/db/backends/util.py:36 def execute(self, sql, params=()): }}} and some levels bottom it invoke this: {{{ MySQLdb/cursors.py:139 def execute(self, query, args=None): ... if args is not None: query = query % db.literal(args) }}} Look with attention on default values of argument of functions. When i do in my code execute() additional args set by default to turple (), BUT MySQLdb in his execute method waiting None value as default. Next, if my SQL code contain a % symbol (not in placeholder meaning), that MySQLdb execute() method think that i send additional arguments to paste in SQL and do formatting. At this moment raise error because this is wrong Python code: {{{#!python 'SELECT * from `test`.`test_table` WHERE name LIKE "Jo%";'%() Traceback (most recent call last): File "/usr/lib/python2.7/site-packages/IPython/core/interactiveshell.py", line 2538, in run_code exec code_obj in self.user_global_ns, self.user_ns File "<ipython-input-1-4ae0cf876369>", line 1, in <module> 'SELECT * from `test`.`test_table` WHERE name LIKE "Jo%";'%() TypeError: not enough arguments for format string }}} In confirmation of the foregoing, if change my code to {{{#!python from django.db import connection ... cursor=connection.cursor() cursor.execute('SELECT * from `test`.`test_table` WHERE name LIKE "Jo%";', None) }}} that it will b… nobody EvgeniyMakhmudov db, error 1 1 0 0 0 0
20119 2013-03-23 17:08:53 2013-03-23 19:02:20 2019-06-24 03:53:22.831352 Unreviewed closed Documentation Uncategorized Normal 1.5 fixed typo error "in it's definition" > "in its definition" https://docs.djangoproject.com/en/dev/topics/auth/customizing/#django.contrib.auth.models.CustomUser.USERNAME_FIELD It should be written "in its definition" nobody vinyll   0 1 0 0 0 0

Advanced export

JSON shape: default, array, newline-delimited, object

CSV options:

CREATE TABLE "tickets_full" (
        id int primary key,
        created datetime,
        changetime datetime,
        last_pulled_from_trac datetime,
        stage text,
        status text,
        component text,
        type text,
        severity text,
        version text,
        resolution text,
        summary text,
        description text,
        owner text,
        reporter text,
        keywords text,
        easy boolean,
        has_patch boolean,
        needs_better_patch boolean,
        needs_tests boolean,
        needs_docs boolean,
        ui_ux boolean
    );