First time here? Check out the FAQ!

sysko's profile - activity

2018-05-21 19:31:59 -0500 received badge Famous Question (source)
2015-06-09 05:15:24 -0500 received badge Popular Question (source)
2013-12-05 15:03:03 -0500 received badge Famous Question (source)
2013-10-22 18:56:20 -0500 received badge Nice Answer ( source )
2013-09-01 16:37:20 -0500 received badge Famous Question (source)
2012-12-04 21:45:44 -0500 received badge Scholar ( source )
2012-12-04 21:45:42 -0500 received badge Supporter ( source )
2012-12-03 08:37:01 -0500 asked a question Deactivate ability to post without logging in

As it seems to be misleading, at least for my students (I'm using askbot to provide them a platform to ask me questions about the class), could it be possible to have a switch to replace the possibility to ask before login. and instead being redirected on the page to log in / create account?

2012-12-03 08:34:17 -0500 asked a question view counter is no more updated (celery task not executed)

I've set up my own askbot server

jisuanjiwenti

it's based on the master's branch, since some days (without git pull between) my number of view are not updating

And to be honest, being not very confident with django code, I don't know where to beging to debug the code and know from where the problem is coming. If someone can give me even some little hint that would be greatly appreciated.

Edit: I've debugged the code and I've arrived to the conclusion that for some reason in the file askbot/views/readers.py

    tasks.record_question_visit.delay(
        question_post = question_post,
        user_id = request.user.id,
        update_view_count = update_view_count
    )

the method record_question_visit is never called (I've put some debug around and inside this method, so this line is executed, but it seems the task is delayed and never executed

because if I "hack" the code like this

    #TODO MYHACK
    question_post.thread.increase_view_count()
    tasks.record_question_visit.delay(
        question_post = question_post,
        user_id = request.user.id,
        update_view_count = update_view_count
    )

The views are correctly updated

2012-11-02 09:24:41 -0500 received badge Self-Learner ( source )
2012-11-01 21:25:23 -0500 asked a question Have the form to create a new account on the same page as sign-in

I've read somewhere that you lost 10% of your users at each steps you add Moreover in some case people may not want or are not able to use an external service to provide login (i.e in China because all the provider listed are either blocked or unknown) and as the link to create a new account is not that obvious it may discourage people to register

2012-11-01 21:20:16 -0500 asked a question Chinese translation has signin and signup at the same url

both are put to "登录" and by so it's impossible to create a local account when using the chinese version

I've taken a look at the .po files and that confirms that the problem come from there, if i replace one of the translation by something else it works

after I'm not a native Chinese and I'm a bit confused by sign up and register (which has certainly also confused the translator I think)

on most of the website you have

注册  to create a new account
登录  to login with an already existing account
2012-11-01 20:05:50 -0500 answered a question Possibility to not have dependency on googleapis.com (China problem)

it seems to me that a beinning of answer to my point 1 is to use

ASKBOT_USE_LOCAL_FONTS = True

but it's written nowhere, and as the template precise "USE_LOCAL_FONTS" it's misleading even if you're reading the code (and have very few knowledge of django). I think it must be included in the settings.py

moreover it does not cover everything (for example traditional.css and askbot/const/__init__.py )

2012-11-01 19:46:42 -0500 asked a question Possibility to not have dependency on googleapis.com (China problem)

As you may know in China some websites are not accessible, and the domain googleapis is one of them (note that it vary a lot from times and province so something which is not accessible today in Shanghai is maybe accessible in Beijing, and maybe tomorrow it will be the opposite)

The bad things is that sometimes the connection is not blocked directly but will instead never get a reply, making askbot to load for ages, waiting for googleapis.com and fonts.google.com to timeout and only then askbot's html will be visible

By using some "noscript" plugin on firefox to block this two domains, the website load fine

So I think it would be great if

  1. one can choose not to use it
  2. in the case one still decide to use, maybe by doing some tricks in the way these resources are loaded, to not make it block the load process

so that I dunno if that will "block" some features, but at least we'll be able to see the content, which is already better than nothing

2012-11-01 15:23:24 -0500 received badge Student (source)
2012-11-01 15:22:56 -0500 received badge Necromancer ( source )
2012-11-01 15:22:56 -0500 received badge Teacher ( source )
2012-11-01 11:53:21 -0500 answered a question OperationalError at /account/signin/

It certainly due to the fact that before MySQL was using by default MyISAM as it's storage engine and the version 5.5 now use InnoDB, it seems to create some uncompatibility on altering table containing foreign key constraint (i.e it works with myisam but not innodb) so the migrate fails

doing this solve the issue for me (I know it's a work around, but it may helps to create a more generic solution)

open a MySQL shell (or phpmyadmin whatsoever) and run

 show create table askbot_anonymousanswer;

you should see a line that looks like this

CONSTRAINT `question_post_id_refs_id_648edfb25e596e3` FOREIGN KEY (`question_post_id`) REFERENCES `askbot_post` (`id`)

note the name of this question_post_id_refs_id_648edfb25e596e3

and then run

alter table askbot_anonymousanswer drop foreign key   question_post_id_refs_id_648edfb25e596e3 ;
alter table askbot_anonymousanswer change question_post_id question_id int(11) ;
2012-11-01 10:32:03 -0500 commented answer STATIC_URL value gets mangled when specifying URL starting with 'http://'

though it does solve my problem i see two problems to this solution: 1 - it's no more like the django normal usage and should at least be precised in the comments of settings.py 2 - does it work well with https ?

2012-11-01 09:47:04 -0500 received badge Editor (source)
2012-11-01 09:45:51 -0500 asked a question STATIC_URL value gets mangled when specifying URL starting with 'http://'

If I specify

STATIC_URL  = "http://static.example.com/media"

in the html I will get

http:/static.example.com/media

(the double / of http:// is replaced by one)

After investigation it seems to me that the problems come from this line in ./askbot/skins/utils.py:

url = os.path.normpath(url).replace('\\', '/')

(line 163 on the current source code) , normpath may replace the // by /