Roberto_Pezzali wrote:
In my discourse installation I notice the lounge label with a small icon: how can I add icon to other categories?
Posts: 1
Participants: 1
Roberto_Pezzali wrote:
In my discourse installation I notice the lounge label with a small icon: how can I add icon to other categories?
Posts: 1
Participants: 1
Abhishek Gupta wrote:
Continuing the discussion from So, you want to help out with Discourse:
Okay, Done.
Already send a pull request.I would appreciate if someone try it on production environment, as i am concerned with some performance issues, i.e the links are converted to
goo.gl
links at the time of rendering the post menu, so if a page containsHUGE
no of posts it may take some time for such large amount ofPOST
requests. Because i don't have a SSD yet, i can't really tell if the time lag is because of code or my hardware .And Yay, my first contribution to discourse
Posts: 12
Participants: 6
Rikki Tooley wrote:
On Windows Phone (and Android I think?), pressing the back button closes overlays and popups in native apps. To me, Discourse on mobile feels like an app (I have my site pinned to my start screen and everything), meaning I press the back button to close overlays by habit... I think it should behave like a native one in this respect. Currently, pressing the back button when e.g. the post editor is open will navigate to the last page.
I'm not sure whether this should apply to desktop as well, but I can't think of a downside - even if the user expected that the last page would be shown (instead of closing the popup), all they need do is press back again.
If the post isn't clear, I'll find a link to the Windows Phone UX guidelines when I'm back at my desktop... that may help.
Posts: 1
Participants: 1
probus wrote:
Continuing the discussion from How do I clear the little number in the upper right corner of page?:
I don't know how often this comes up, but since this behaviour is different from how most other sites (like facebook) handle notifications and I had to scan the attached picture a couple of times to spot the correct message even when I knew what to look for it seems to me there is some room for improvement here.
I can offer a few i suggestion here:
- pin unread messages on top of the notifications list,
- use a different icon for opened messages (though there seems to be no opened envelope icon in font awesome) and
- use a different shade of green that is easier to notice for the highlight.
Posts: 2
Participants: 2
Hunter wrote:
I'm implementing the timeouts, and will probably implement the byte limits. I have it working internally with hardcoded values, but was wondering if anyone has any advice on what the default value should be and how/if users should be able to change it. Also, was I correct to put this in feature rather than ux?
Posts: 6
Participants: 5
Johan wrote:
The current grey background in the user profile page looks like a placeholder for something that could contain an image.
What about adding an option to link/upload a background image Twitter/GPlus style.
There could be an option to tile the background image like Twitter has.
PS. My email is public on my website, don't worry bout it. =)
Posts: 4
Participants: 3
Kane York wrote:
Clicking the like button results in a 404 to
/post_actions
and it reverting in the UI.
Posts: 4
Participants: 2
Jonathan Allard wrote:
I just noticed my node on Docker wasn't sending out mail.
I put the required settings in
app.yml
:DISCOURSE_SMTP_ADDRESS: smtp.sendgrid.net DISCOURSE_SMTP_PORT: 587 DISCOURSE_SMTP_USER_NAME: xyz DISCOURSE_SMTP_PASSWORD: abc
I took a look to the other posts
But none really apply here. Besides, ain't those settings supposed to be set correctly by Docker?
Two things are weird to me:
root@dockerdiscourse-app:~# ls /var/www/discourse/log development.log production_errors.log production.log unicorn.stderr.log unicorn.stdout.log
No mail or sidekiq log.
Also,
Delivery Method sendmail
which seems weird considering it should be SMTP.
I don't know where to check anymore.
Posts: 4
Participants: 2
Roberto_Pezzali wrote:
Hello, another annoying but in my Discourse installation (meta.discurse.com is fine).
With my iPhone with iOS 7 when I try lo load Discourse I have a version without CSS on Safari Mobile while with Chrome everything is perfect.
Any idea?
Posts: 2
Participants: 2
Kenneth Lowrey wrote:
I trying to add about 350 users to a group, however I can't seem to add more than 201 members. I looked for a setting but didn't find one. Is there a setting that limits the number of users allowed in a group?
Is there a bug that is limiting the number of users that I can add to a group?
Is the universe just messing with me?
Posts: 6
Participants: 2
Kenneth Lowrey wrote:
Hi everyone
So we have been running Discourse for over two months and I'm in the middle of a project to improve the experience our customers are getting.
The issue we are having is that our customers are getting lost in the muck of the hundreds of conversations that are going on. We want to organize the site making the experience more tight.
One thing that we want is an official notifications panel. this would be a static panel (or simply section of the main page) where official announcements are displayed.
This is something that battle.net has (http://us.battle.net/sc2/en/forum/).
It seems to me that this feature would be something that could be accomplished via plugin. Am I correct?
Also is it possible (via plugins or other means) to modify the look and feel of the main discourse page?
Could anyone give me any insight into what approach we could take to make this enhancement to our discourse install?
Thanks in Advance
Posts: 2
Participants: 2
Briancguy wrote:
Hi,
Its unclear to me how to clear the little number 1 in the upper right corner of the page. I've checked all the postings listed (see below) when I click on the speach bubble - but the number does not go away. This is very confusing.
Posts: 7
Participants: 6
Corey Robertson wrote:
So any plugin client settings get jammed into a big object amongst all of the other site settings. Every plugin will have to use unique setting names if this is to be the case.
Instead, what if we move plugin settings out of site settings completely, and instead register PluginSettings? Add a Plugin top level nav item in the admin settings that will be similar to Site Settings, but only list all of the plugins there. Bonus is that the site settings Left Nav won't get any longer as you add plugins.
Then, instead of asking for your plugin settings like
Discourse.SiteSettings.mySettingsyou could do
Discourse.PluginSettings[:myPlugin].mySetting.Any thoughts on this? I'd love to work on it if it's a direction we want to go.
Posts: 3
Participants: 3
Sam Saffron wrote:
I just added a couple of classes that allow you to override the way we lookup the current user. This is particularly important to people looking to integrate using cookie auth from a top level domain.
Essentially you can elect a provider for the current user functionality, to do so:
Implement a class that inherits off:
class Auth::CurrentUserProvider # do all current user initialization here def initialize(env) raise NotImplementedError end # our current user, return nil if none is found def current_user raise NotImplementedError end # log on a user and set cookies and session etc. def log_on_user(user,session,cookies) raise NotImplementedError end # api has special rights return true if api was detected def is_api? raise NotImplementedError end # we may need to know very early on in the middleware if an auth token # exists, to optimise caching def has_auth_cookie? raise NotImplementedError end def log_off_user(session, cookies) raise NotImplementedError end end
If you just feel like extending the current behavior, inherit off
Auth::DefaultCurrentUserProvider
, this allow you to run special code when a user is logged on or off and so on.To wire your own provider, run
Discourse.current_user_provider = MyCurrentUserProvider
This is way cleaner and more predictable than monkey patching.
Posts: 11
Participants: 8
Victor Ng wrote:
I'm constantly getting errors in my sidekiq "Failed Jobs" list where the error is:
Excon::Errors::NotFound: Expected(200) <=> Actual(404 Not Found) response => #"NoSuchBucketThe specified bucket does not exist"
Is there anything I can do to remove those clean up jobs from the queue?
This error is tracked in https://github.com/discourse/discourse/issues/1899
Posts: 4
Participants: 2
Kane York wrote:
On clicking like:
Sometimes this happens instead:On submitting a reply or new topic:
( http://i.imgur.com/XNwu0CB.jpg )Deleting a post silently fails, but the UI still updates
Posts: 3
Participants: 3
Nition wrote:
There may be a good reason for this (escape characters?) but if someone types \{ into a post, it needs two slashes to show one slash (I've used two here), otherwise you just get {. I haven't come across any other cases where backslashes don't show up.
Tends to show up when someone writes a file path like:
C:\{your username}\filesAlso, if a topic title contains a backslash as the last character it doesn't show up at all. However, if the title is then edited, the backslash shows up in the edit box, and if it's saved again (or the backslash is added in an edit), it then shows up.
Posts: 3
Participants: 2
Sam Saffron wrote:
- Go to first post on a 100 post topic
- Reply to it
- Reply is indeed at the end, however posts 20 -> 100 are missing above your post.
cc @eviltrout
Posts: 2
Participants: 2
Tattoo wrote:
It is very useful when we need to ask users to vote on something
Posts: 7
Participants: 5
thepractice wrote:
Continuing the discussion from Problem during initialization:
OK, here's the background. I wanted to change some words in the client.en.yml file. So I edited that file and the changes didn't appear on my site. Then I read that rebooting the server (VPS) would help show changes so I rebooted my Digital Ocean VPS with "sudo reboot". In retrospect this was very dumb.
After the server reboot, I reloaded nginx and reloaded bluepill per the instructions in the official ubuntu install guide. This has worked for me in the past when I took a snapshot of my Digital Ocean droplet, which involves shutting down server. However, it did not work for me this time. That is, my 502 Bad Gateway error page remained even after reloaded nginx and typing the commands to reboot bluepill.
I did some investigating. My nginx error log showed "no live upstreams while connecting to upstream". My tmp/sockets directory was empty; I think it should have had some files like thin.0.sock.
So I decided to fix my problems by doing the update process described in the Ubuntu install guide. I checkout the master branch, stash my local changes (client.en.yml, discourse.conf, database.yml), git pull, then "git stash pop" to get my local changes back. Then I enter this command:
RUBY_GC_MALLOC_LIMIT=90000000 RAILS_ENV=production bundle exec rake db:migrate
And I get the "fe_sendauth: no password supplied" error message.
"No problem," I'm thinking because I got this message upon this command before when upgrading from 9.7.9. Back then, I fixed the problem by resetting psql 'discourse' user password and editing config/discourse.conf to look like:
# username accessing database db_username = discourse # password used to access the db db_password = my_password
Also I had changed etc/postgresql/9.1/main/pg_hba.conf so that "all" users use md5 method instead of peer, like so:
local all all md5
I had known this approach was working because the error message from running rake changed to
FATAL: password authentication failed for user "discourse"
.Back to my present problem. In the updating process, when I run the rake:db migrate command, I always get the "fe_sendauth: no password supplied" error message, no matter what I do. My database.yml is stock/default. My config/discourse.conf has the db_username and db_password lines uncommented and properly filled out, as shown above. I tried with and without quotation marks around the password. I tried editing the pg_hba.conf file so that "all" users have "trust" method. I even tried editing production.rb to override global default user_name and password settings.
Nothing works, and I'm stuck with "fe_sendauth: no password supplied" error message upon trying to
rake:db migrate
.Any help would be greatly appreciated
Posts: 21
Participants: 5