Re: Change password / Log out doesn't work with apache2
I change the base.html, maybe not a good idea, but it works now: -{% trans 'Welcome,' %} {% if user.first_name %}{{ user.first_name }}{% else %}{{ user.username }}{% endif %}. {% block userlinks %}{% trans 'Change password' %} / {% trans 'Log out' %}{% endblock %} +{% trans 'Welcome,' %} {% if user.first_name %}{{ user.first_name }}{% else %}{{ user.username }}{% endif %}. {% block userlinks %}{% trans 'Change password' %} / {% trans 'Log out' %}{% endblock %}
Re: Change password / Log out doesn't work with apache2
On Thursday 24 Nov 2005 9:45 am, wang bin wrote: > hi, when I trying to setup my project on apache2 with mod_python, > following the document everything seems good except the Change > password / Log out doesn't work. When I click the link I get the > "The requested URL /admin/password_change/ was not found on this > server." Error, could anyone tell me the reason or how to solve > this problem ? i have the same problem. my admin site url is: localhost/myapp/admin/ but logout points to: localhost/admin/logout instead of localhost/myapp/admin/logout same with change password - there is a bug here -- regards kg http://www.livejournal.com/users/lawgon tally ho! http://avsap.org.in ಇಂಡ್ಲಿನಕ್ಸ வாழ்க!
Re: Django Trac Spam
On Nov 23, 2005, at 10:12 PM, Adrian Holovaty wrote: On 11/23/05, Tom Tobin <[EMAIL PROTECTED]> wrote: Someone's been spamming the Django Trac today, screwing up tickets and whatnot. Is there any way to fix and/or prevent this? I've been deleting 'em as they come in. Maybe we can hack Trac so that it silently ignores comments that include naughty words. I've also been keeping an eye on it. FYI, this is ticket #454 in Trac itself (http://projects.edgewall.com/trac/ticket/454); I'm about to jump in #trac and bug people about it. Jacob
Change password / Log out doesn't work with apache2
hi, when I trying to setup my project on apache2 with mod_python, following the document everything seems good except the Change password / Log out doesn't work. When I click the link I get the "The requested URL /admin/password_change/ was not found on this server." Error, could anyone tell me the reason or how to solve this problem ?
Django Trac Spam
Someone's been spamming the Django Trac today, screwing up tickets and whatnot. Is there any way to fix and/or prevent this?
'log' is not a valid tag library
hi am using the latest svn up - did init and install etc - main site works fine. When i try to log into admin, after entering username and password i get this error: 'log' is not a valid tag library Request Method: GET Request URL:http://localhost/bambanks/admin/ Exception Type: TemplateSyntaxError Exception Value:'log' is not a valid tag library Exception Location: /usr/lib/python2.3/site-packages/django/core/template/defaulttags.py in do_load, line 631 any clues? -- regards kg http://www.livejournal.com/users/lawgon tally ho! http://avsap.org.in ಇಂಡ್ಲಿನಕ್ಸ வாழ்க!
Re: Extend list template of admin app?
On 11/23/05, oggie rob <[EMAIL PROTECTED]> wrote: > How many more changes are expected on that branch? And is there an ETA > on when it might merge/if it might merge with the head? The branch is done. I'm planning on merging it in the next couple of days -- or as soon as tonight or tomorrow, if I can flee the Thanksgiving festivities effectively. Adrian -- Adrian Holovaty holovaty.com | djangoproject.com | chicagocrime.org
Re: Performance tuning tips?
no idea why I put it there, I tried moving it to the top, the response times went back to around 3 seconds. On 11/24/05, Eugene Lazutkin <[EMAIL PROTECTED]> wrote: > > Interesting that the session middleware was the culprit. BTW, why did you > put it at the end of your middleware list? > > The default order is: > > "django.middleware.sessions.SessionMiddleware", > "django.middleware.common.CommonMiddleware", > "django.middleware.doc.XViewMiddleware", > > (see > http://www.djangoproject.com/documentation/settings/#middleware-classes). > > Thanks, > > Eugene
Re: admin suggestions
Oops! I guess I hadn't tried it as much as I thought! Unfortunately #2 there causes a problem when they aren't overridden (a TypeError if '%s' is not included in the post_url). I guess a possible fix is to do a find, e.g.: if post_url.find('%s') >= 0: post_url = post_url % pk_value return HttpResponseRedirect(post_url) but that is not as clean as I would like. Anyway it points out that the post_url_continue value may run into the same problem if it is overridden with a %s argument, however it is quite unlikely anybody would do that. -rob
Re: raw_id_admin
On 11/23/05, Grigory Fateyev <[EMAIL PROTECTED]> wrote: > class Article(meta.Model): > author = meta.ForeignKey(User, raw_id_admin=True) > > vews/articles.py generate form: > > Author: {{ form.author }} > {% if form.author.errors %}*** {{ form.author.errors|join:", " }} > {% endif % } > > If the field is filled by username, form give an error. If it filled > user_id form works correct. How to make form that can be filled by > username insted of user_id? Hi Grigory, raw_id_admin supports only entry of IDs (primary key fields). Adrian -- Adrian Holovaty holovaty.com | djangoproject.com | chicagocrime.org
Re: init fails
On 11/23/05, David Ascher <[EMAIL PROTECTED]> wrote: > I've had the same problem in cases when I forget to use named parameters to > my FloatFields, which then generates invalid SQL with "None" in it. That's > a bug, IMO. I may even try to fix it =). As of changeset 1314 (http://code.djangoproject.com/changeset/1314), the model validator (which runs automatically during "django-admin runserver") complains if you don't max_digits and decimal_places on FloatFields. Is it missing an edge case? Adrian -- Adrian Holovaty holovaty.com | djangoproject.com | chicagocrime.org
Re: Atom and web services interfaces
On 11/23/05, tonemcd <[EMAIL PROTECTED]> wrote: > I'm thinking of using Django as the front-end to an experimental > Atom-enabled 'store' as outlined by Joe Gregorio at > http://www.xml.com/pub/a/2005/09/21/atom-store-web-database.html > [...] > Any thoughts gratefully received. Hi Tone, Django's very REST-friendly, so the task should be straightforward. Feel free to post specific questions, and keep us posted on how it's coming! Adrian -- Adrian Holovaty holovaty.com | djangoproject.com | chicagocrime.org
Sequence in PostgreSQL is not updated when doing manual insert
In order to load some test data into the database, I'm using the following code: from django.core.db import db cursor = db.cursor() cursor.execute("insert into cities (id, name) values (1, 'Chicago');") cursor.execute("insert into cities (id, name) values (2, 'Miami');") ... This works fine, but then when I try to use django later to add to this table, I get an error: cities.City(name='Madison').save() ERROR: duplicate key violates unique constraint "cities_pkey" I've tracked down the problem, and it seems that when I do my manual inserts, the sequence (cities_id_seq) that PostgreSQL uses for primary keys isn't updated. Therefore, when I try to save my object with django, the primary key collides with an already existing key. I've implemented a simple fix by manually settings the sequence value after I do my inserts, but I'm wondering why django/postgresql doesn't do the right thing. Has anyone seen this before?
Model updates in early development stages
There is a thread in django-developers list talking about the way Django and Rails handle frequent changes to models on early stages of development ('prototyping' - to sound cool). Neither framwork does this transparently which is understandable because it's definitely hard and even arguably desirable at all :-). I'd like to share my method of doing this task where I use one undocumented feature that Adrian Holovaty once mentioned in this list. I'm writing here because I'm not sure if it's ready for the Cookbook in the wiki. May be everyone will reply that I'm doing it incredibly painful way and should just read this URL and that URL to do it right :-). Also feel free to correct my English since it's not my native language. So to the point. 1. I begin designing a model to the whole depths of my current domain knowledge. But I'm absolutely sure that in a near future my knowledge will change and will require changing models. 2. Significant changes to models that require database changes are: adding/removing/renaming fields, adding/removing/renaming tables, changing types and sizes of fields. But there is no point of doing something with database when adding some custom method or changing META class. 3. When database change is required I use: django-admin.py sqlreset ... which effectvely wipes out entire application structure then creates new one. 4. But this also wipes all the test data in your database and if you had to recreate it by hand each time it would defeat the whole point of this method :-). Luckily you can easily automate it. After creating the structure 'sqlreset' tries to locate and execute SQL script in /apps//sql/.sql This script should contain SQL statements that create test data for your application. Creating this file from scratch by hand is boring. Instead I create all the data using shiny Django's admin interface and then use database tools to generate such script. For PostgreSQL this utility called pg_dump and can be used like this: pg_dump -a -D > .sql Options mean: -a: dump only data, not structure (because structure will be created by sqlreset) -D: use INSERT statements with column names Then I edit the resulting file removing all core data leaving only INSERTS into my applications tables. It still sounds a bit complex but I've done it only first time. On all subsequent changes it's easier to correct this file by hand. And often it's not even required when you, for example, just change the length of a text field. 6. The 'sqlreset' outputs all SQL statements to console. To further automate the task I redirect this output into command-line database client instead of copy-pasting it by hand: django-admin.py sqlreset | psql
raw_id_admin
Hello! In model i use this field: class Article(meta.Model): author = meta.ForeignKey(User, raw_id_admin=True) vews/articles.py generate form: Author: {{ form.author }} {% if form.author.errors %}*** {{ form.author.errors|join:", " }} {% endif % } If the field is filled by username, form give an error. If it filled user_id form works correct. How to make form that can be filled by username insted of user_id? -- Всего наилучшего! greg [at] anastasia [dot] ru Григорий.
Re: init fails
On 11/23/05, Kenneth Gonsalves <[EMAIL PROTECTED]> wrote: On Wednesday 23 Nov 2005 2:46 pm, Kenneth Gonsalves wrote:> hi> i have a working django app. i did an svn checkout of the app> into another machine. When i try to run init, i am getting this> error: >> Error: The database couldn't be initialized.> name 'true' is not defined> this is with postgres - any clues?ok - solved, that was a syntax error in my modelI've had the same problem in cases when I forget to use named parameters to my FloatFields, which then generates invalid SQL with "None" in it. That's a bug, IMO. I may even try to fix it =). --david
Re: Sequence in PostgreSQL is not updated when doing manual insert
On Nov 23, 2005, at 11:15 AM, pgross wrote: I've tracked down the problem, and it seems that when I do my manual inserts, the sequence (cities_id_seq) that PostgreSQL uses for primary keys isn't updated. Therefore, when I try to save my object with django, the primary key collides with an already existing key. I've implemented a simple fix by manually settings the sequence value after I do my inserts, but I'm wondering why django/postgresql doesn't do the right thing. Has anyone seen this before? Yeah, this is an annoying postgresql wart. We've run into it enough that we've rolled the fix into the django-admin utility. "django- admin sqlsequencereset " will print out the statements to fix the sequences. Programmatically, you can get at this SQL from the django.core.management.get_sql_sequence_reset() function. Jacob
Re: Performance tuning tips?
Interesting that the session middleware was the culprit. BTW, why did you put it at the end of your middleware list? The default order is: "django.middleware.sessions.SessionMiddleware", "django.middleware.common.CommonMiddleware", "django.middleware.doc.XViewMiddleware", (see http://www.djangoproject.com/documentation/settings/#middleware-classes). Thanks, Eugene
Atom and web services interfaces
Hello, I'm thinking of using Django as the front-end to an experimental Atom-enabled 'store' as outlined by Joe Gregorio at http://www.xml.com/pub/a/2005/09/21/atom-store-web-database.html I'm a real newbie to Django, having spent the last 7 years in Zope-land, and am looking at Django as an adjunct to the many many Zope systems we run here (I just had reason to play around with PageTemplates and PythonScripts this morning and the frustration was intense!) Any thoughts gratefully received. Tone
Re: Paginating complex queries
On 23 Nov 2005, at 15:07, Simon Willison wrote: The aim with the template system has always been to keep it de- coupled from the request/response stuff, so it can be used as a standalone component. It's hard to see how the request object could be exposed to custom tags without breaking that separation. That said, it would be enable some very neat custom template tag tricks. Right. I think that decoupling is a Good Thing. I use template systems for a variety of purposes outside of web app development and it's nice to have a standard option for all requirements. I personally haven't bashed my head against Django that much yet so I feel I'm getting a bit out of my depth here. Perhaps the best thing is just to keep explicitly passing in the pagination arguments. Afternoon, man about the Internet -- http://aftnn.org/
Re: Paginating complex queries
On 23 Nov 2005, at 14:59, Afternoon wrote: Is there a way that the request object could be exposed to custom tag code, but not the template itself? The aim with the template system has always been to keep it de- coupled from the request/response stuff, so it can be used as a standalone component. It's hard to see how the request object could be exposed to custom tags without breaking that separation. That said, it would be enable some very neat custom template tag tricks.
Re: Paginating complex queries
On 23 Nov 2005, at 14:42, Colin Howlett wrote: +1 to putting the request object into the context as standard, or at least into DjangoContext. I've only been playing with Django for a few days, and it's already obvious that this single change would make writing useful custom tags a whole lot easier. This also means that there's more of a temptation to use the request object directly in template logic. Is there a way that the request object could be exposed to custom tag code, but not the template itself? Afternoon, man about the Internet -- http://aftnn.org/
Re: Bi-Directional ManyToMany
Its not that needs to be done asap, i'll wait until new-admin merges and then give it a try.
Extend list template of admin app?
Hi, I would like to add some custom mass-editing capabilities to the admin app, and for this I would like to know if it's possible to extend the list template of the admin app. Looking at the code in django/contrib/admin/views/main.py I see, that the template is entirely dynamically created (which is obvious as the admin app doesn't know which model there will be), which makes extending the template impossible. But maybe there's some hook or some other trick to allow this. Thanks in advance Benjamin
Re: Performance tuning tips?
Yes I am running openload on localhost, I don't trust my DSL enough to get accurate results :( my MIDDLEWARE_CLASSES: "django.middleware.common.CommonMiddleware", "django.middleware.doc.XViewMiddleware", "django.middleware.sessions.SessionMiddleware", I'm caching on a view by view basis (hope that's the correct terminology) using this code. index = cache_page(index, 180) After looking at those middleware classes I removed the session middleware and it has dropped response times to a very respectable: Total TPS: 144.27 Avg. Response time: 0.034 sec. Max Response time: 0.675 sec Total Requests:1951 Total Errors: 0 So I believe my problem was the session middleware. At present I don't require the session middleware (I will move the admin). However in the near future the forum will change to requiring user authentication, any thoughts about the best way for me to enable caching with the sessions enabled? If need be what would help my problems is if the session caching time can't be improved then if sessions are enabled only for certain views? As an example, in one of my sites I have no session requirements, except when generating a captcha and then verifying the captcha when the comment is submitted. Thanks for your help! On 11/23/05, Eugene Lazutkin <[EMAIL PROTECTED]> wrote: > > For comparison I ran openload against my blog page > (http://lazutkin.com/blog/). It is 38012 bytes at the moment. I don't know > how many db transactions required to build it, but with cache it's a moot > point anyway. Incidentally I use "file:" cache. The site is hosted with > DreamHost --- cheap community hosting company. > > Over Internet using my home connection I got: > > Total TPS: 12.54 > Avg. Response time: 0.781 sec. > Max Response time: 1.641 sec > Total Requests: 503 > Total Errors: 0 > > With average ping ~40ms, it correspondes to ~400kbps, which looks about > right order of magnitude. > > Judging by your times you probably used "localhost" for load testing, or you > had some intermediate caching, or you have really fat connection to get 24k > in 20ms. ;-) > > I don't know what the exact problem is, but I would suggest to check basic > stuff first: check footprint (maybe swapping is an issue); check database > speed (e.g., using Django's database API with some rudimentary timing > program); if you run db and apache/django on the same server, try to > separate them (sometimes there is an interference between programs); to get > realistic test results, try to run openload from separate computer (or > several copies of openload from separate computers for even more realism). > If everything works as expected, you probably found a bug in cache subsystem > and we need to find a way to reproduce it. > > BTW, could you cut'n'paste your MIDDLEWARE_CLASSES? The order is important. > > Thanks, > > Eugene
Re: init fails
On Wednesday 23 Nov 2005 2:46 pm, Kenneth Gonsalves wrote: > hi > i have a working django app. i did an svn checkout of the app > into another machine. When i try to run init, i am getting this > error: > > Error: The database couldn't be initialized. > name 'true' is not defined > this is with postgres - any clues? ok - solved, that was a syntax error in my model -- regards kg http://www.livejournal.com/users/lawgon tally ho! http://avsap.org.in ಇಂಡ್ಲಿನಕ್ಸ வாழ்க!