Kane York wrote:

After you like a post, the server sends a message to everyone viewing the topic to refresh the post. This includes you, and sometimes when you reload the post, your like isn't there. So the like button reappears. If you click the like button again, the server rejects it (you already liked the post), so the like button reappears because the request failed.

Reloading the page gets to the correct state.

Posts: 2

Participants: 2

Read full topic

Sam Saffron wrote:

One very cool feature Discourse has is "oneboxing". If you include a link to your favourite site, it will try to create a usable snippet for you automatically.

For example: paste the link http://en.wikipedia.org/wiki/Spite_house on a line by itself and you will see a nice onebox from Wikipedia

---

Some samples:

http://meta.discourse.org/t/how-do-you-spoiler-an-image-or-onebox/55

http://meta.discourse.org/t/how-do-you-spoiler-an-image-or-onebox/55/2

en.wikipedia.org

Spite house

A spite house is a building constructed or modified to irritate neighbors or other parties with land stakes. Spite houses often serve as obstructions, blocking out light or access to neighboring buildings, or as flamboyant symbols of defiance. Because long-term occupation is at best a secondary consideration, spite houses frequently sport strange and impractical structures. Spite houses are much rarer than spite fences. This is partially attributable to the fact that modern building codes often p...

http://en.wikipedia.org/wiki/Spite_house

amazon.com

Walter the Farting Dog Farts Again

Glenn Murray, Audrey Colman (Illustrator) William Kotzwinkle

When Walter the Farting Dog is sold for a measly ten pounds, Billy and Betty are distraught. But Walter doesn't realise he's been caught up in a dreadful plan involving a fake clown, litres of his ...

http://www.amazon.com/Walter-Farting-Farts-Picture-Puffins/dp/0141382074

https://play.google.com/store/apps/details?id=com.machineworksnorthwest.duke3d

https://play.google.com/store/apps/details?id=com.machineworksnorthwest.duke3d

itunes.apple.com

Duke Nukem 3D

***************************************************V1.4 is here! Now supports "Drag to look" controls as well as dual stick!THE ORIGINAL KING OF ACTION RETURNS!With dozens of Game of the Year awards under his belt, Duke remains the King of Action. Larger than life action is what Duke is all about, mixed with humor and innovative explosive action in this 20+ hour game. Los Angeles. Under attack by aliens intent on stealing our chicks for breeding. Duke Nukem springs into action dealing damage lik

https://itunes.apple.com/us/app/duke-nukem-3d/id325995561?mt=8

gist.github.com

Node.js message passing test

gistfile1.js

var cluster = require('cluster');

var m = 10000000;

function bounce(msg, out) {
  if (msg < m) {
    out.send(msg + 1);
    return null;
  } else {
    console.log("Finished with", msg);
    out.disconnect();
    return msg;
  }
}

if(cluster.isMaster) {
  var worker = cluster.fork();
  worker.on('message', function(msg) {
    bounce(msg, worker);
  });
  worker.send(1);
} else {
  process.on('message', function(msg) {
    bounce(msg, process);
  });
}

https://gist.github.com/3200829

github.com

https://github.com/rack/rack/blob/master/lib/rack/etag.rb

require 'digest/md5'

module Rack
  # Automatically sets the ETag header on all String bodies.
  #
  # The ETag header is skipped if ETag or Last-Modified headers are sent or if
  # a sendfile body (body.responds_to :to_path) is given (since such cases
  # should be handled by apache/nginx).
  #
  # On initialization, you can pass two parameters: a Cache-Control directive
  # used when Etag is absent and a directive when it is present. The first
  # defaults to nil, while the second defaults to "max-age=0, private, must-revalidate"
  class ETag
    DEFAULT_CACHE_CONTROL = "max-age=0, private, must-revalidate".freeze

    def initialize(app, no_cache_control = nil, cache_control = DEFAULT_CACHE_CONTROL)
      @app = app
      @cache_control = cache_control
      @no_cache_control = no_cache_control
    end

    def call(env)
      status, headers, body = @app.call(env)

      if etag_status?(status) && etag_body?(body) && !skip_caching?(headers)
        original_body = body
        digest, new_body = digest_body(body)
        body = Rack::BodyProxy.new(new_body) do
          original_body.close if original_body.respond_to?(:close)
        end
        headers['ETag'] = %("#{digest}") if digest
      end

      unless headers['Cache-Control']
        if digest
          headers['Cache-Control'] = @cache_control if @cache_control
        else
          headers['Cache-Control'] = @no_cache_control if @no_cache_control
        end
      end

      [status, headers, body]
    end

    private

      def etag_status?(status)
        status == 200 || status == 201
      end

      def etag_body?(body)
        !body.respond_to?(:to_path)
      end

      def skip_caching?(headers)
        (headers['Cache-Control'] && headers['Cache-Control'].include?('no-cache')) ||
          headers.key?('ETag') || headers.key?('Last-Modified')
      end

      def digest_body(body)
        parts = []
        digest = nil

        body.each do |part|
          parts << part
          (digest ||= Digest::MD5.new) << part unless part.empty?
        end

        [digest && digest.hexdigest, parts]
      end
  end
end

https://github.com/rack/rack/blob/master/lib/rack/etag.rb

http://www.flickr.com/photos/eho/149282456/

http://imgur.com/gallery/1PGTI

thenextweb.com

Kim Dotcom’s plan to give New Zealanders free internet could just work

Kim Dotcom's story is something out of Hollywood. The self-proclaimed king of the internet was the founder and operator of file sharing giant Megaupload -- a haven for pirates and ...

http://thenextweb.com/au/2012/11/18/kim-dotcoms-plan-to-give-new-zealanders-free-internet-could-just-work/?fromcat=au

http://www.youtube.com/watch?v=9bZkp7q19f0

Between Two Ferns with Zach Galifianakis: Richard Branson from Between Two Ferns

http://www.funnyordie.com/videos/05c8ec50ed/between-two-ferns-with-zach-galifianakis-richard-branson

<img src='/uploads/default/2428/d63ecff95d125d8e.jpg'>

Posts: 19

Participants: 13

Read full topic

Erlend Sogge Heggen wrote:

Have you noticed this?

Screen Shot 2014-03-18 at 18.22.32.png2560x784 240 KB

I've seen smileys popping up in a lot of unintentional places on this forum today.

Posts: 3

Participants: 2

Read full topic

Tommy Mancino wrote:

I am not sure if this is intended behavior or not.

I have implemented SSO, which works great. However, I am unable to make API calls when I make the forum private by checking login_required. If I uncheck it, everything works fine. I have tried it with the master API key, as well as a private API admin key for the signed in user.

I see that I am being redirected to login when trying to access the json via the api, using the discourse-api gem.

ActionView::Template::Error (757: unexpected token at 'You are being redirected.'):

Is this working as intended? Thanks....

Note: I found this discussion and will look at the source https://github.com/discourse/discourse_api/pull/20

Resolution: This was being caused by the discourse_api gem being out of date. I updated and submitted pull request.

Posts: 1

Participants: 1

Read full topic

Ony A wrote:

I've tried searching for this with no luck, so hopefully this hasn't been posted and I'm posting it in the right place.

Alright so I've been browsing the site a few days and I thought I got a hang of how permalinks to individual posts work: If you're looking at post 30, then the url will have 30 appended to the name, the up-down button at the bottom of the page will also say 30.

However, I was on my phone (IE 11, Windows Phone), and I noticed two things

1. The 'jump' button sits at the bottom of the page.. is super wonky. If you're in one place, it's fine, but once you scroll the button kind of bounces up and down till you stop.
2. More importantly, on long posts, the location on the button is different from where you are.

For instance, this post:

If I scroll down to post #3, the jump button will tell me I'm at #5. The longer I scroll the less in sync the numbers are. I tried forcing the navigation to post 50 by adding 50 to the url on my phone and I ended up on post 36. The url had redirected to 44, despite what I had typed. I tried it on my browser (Chrome) and it jumped to 46 and the browser said 45.

I'm not sure if this is a bug, or expected behaviour (I hope it's the former), because navigating long discussions is going to get really frustrating really fast when you can't even follow a permalink.

Posts: 5

Participants: 2

Read full topic

Hugo Almeida wrote:

I would like to see if someone can reproduce this,

Its funky on Firefox and Chrome, so lets say you have a link
google 1

if i middle click once on it will open a new tab, if i middle click it again the cursor will get stuck on drag navigation. If you let left click in place to go away you'll end up clicking the link again.

On Firefox its worse, because if you close the new tab by accident it won't let you open with middle click and just be stuck at drag navigation.
On Chrome you'll be able to middle click and it will open in new tab but be stuck at drag navigation.

I know its kind of weird use case, but is even more weird and frustrating to be stuck in drag navigation.

---

Another thing I've noticed, but not sure if bug or browser feature.

In Chrome, middle click will open new tab but keep focus on current page, while in Firefox it will open new tab and switch focus.

I've also tried going to google with Firefox and in this case middle click would open new tab but keep focus on current page.

The use case here is about a post with 2 links that i want to open while i keep reading the post.
google 1 and google 2

Posts: 11

Participants: 4

Read full topic

William Karavites wrote:

Here is a screenshot of my version numbers

error1.PNG1285x1408 146 KB

And here's a link to the output for my logs.

http://pastie.org/private/pg79veeiz6lyogahhlptpg

my site is located at http://www.tintonfallscommunity.org

From the logs, it looks like that there is a syntax error? If you need more info please let me know!

Posts: 3

Participants: 2

Read full topic

Kane York wrote:

I am 70% sure that something I did in the Ember code is horrific. Not sure what though. Care to point it out?

Posts: 2

Participants: 2

Read full topic

Ioann wrote:

Watching on Discourse development for a few months, and it's still looks impressive and cool. But want to know a lil bit about developers plans about initial release of this amazing project. I understand that it's really hard to talk about time when you are developing, but seems like @sam know something, but don't want to share.

Like this topic if you want to dive into the future too!

P.S. Polls are awesome.

Posts: 4

Participants: 3

Read full topic

Tommy Mancino wrote:

Queries done through the API key are scoped to the user that created them, which is perfectly reasonable and expected. For instance, calling client.categories will return an array of hashed categories available to that user.

What is the best practice for generating and accessing per user API keys? Are you storing them (encrypted) in a separate database? Is their a way to 'impersonate' a user via the API?

Posts: 1

Participants: 1

Read full topic

David McClure wrote:

We are excited to try out discourse to communicate with our relatively small community of users, (usually employees of other businesses, professional freelancers) and partners. Think hundreds of users, not tens of thousands. Much of what we intend to do with the forum is to replace and supplement our existing channels for support, which is often not easy to fit into a 'ticketing system' because a lot of the support is discussion about what the best way to do something might be and which could benefit a wider audience in the future.

We are getting ready to say goodbye to Zendesk, and steer people away from emailing support@ and folks personal emails for supporting our professional movie and television post-production products going forward.

While I'm tempted to spin up a digital ocean box because it might be fun, I'm not sure I want the responsibility and our own development team and engineers already have enough on their plates with the details of our business. It took us far too long to get rid of our own mail server, and no one is looking forward to going back to those days.

So I'm likely to get convinced (or perhaps I'll convince myself) that DIY isn't the right solution for the company, but I'm still open to considering it at the moment.

If going to a hosted solution, I'm on the fence with heading to discoursehosting.com vs waiting to see where pricing lands for the service the team here will be offering though.

Curious to hear peoples opinions below or via PM.

Posts: 4

Participants: 4

Read full topic

resure wrote:

What is correct way to connect web-only and data Discourse docker containers?

There is note about DISCOURSE_HOST_IP in readme, but how should we specify that variable in container config yml file? Can we just link that images together?

Posts: 3

Participants: 3

Read full topic

Kane York wrote:

After flagging a topic, an unnecessary modal pops telling me that I have "successfully flagged the topic". Yeah, thanks, I already knew that.

Posts: 8

Participants: 3

Read full topic

Bill Ayakatubby wrote:

Continuing the discussion from Windows Phone IE scroll position quirks:

I'm able to reproduce the above bug report in Chrome 33 on Win 8 desktop. Post 56 is the one displayed at the top of the screen, though the top of it is scrolled off just ever so slightly. If I middle-click to open it in a new tab, the URL bar shows 56, 57 if opened in the same tab. There seems to be no difference in scroll position on same-tab vs new-tab. The first time I tried, the progress bar said 51. Now it consistently says 52.

Posts: 7

Participants: 4

Read full topic

James Potter wrote:

First I couldn't get SSL working, so I've been trying to SSH into my container for some debugging. But that's not working either despite messing around with a few different methods.

When I run ./launcher ssh app it does nothing and times out after a few minutes. Same if I SSH in manually with ssh -p 2222 root@localhost -v (it gets to the key exchange and sits there, then eventually times out).

I'm starting to suspect that there's some root cause of all this. If I run ./launcher logs app I get this, ad infinitum:

./run: line 4: exec: rsyslogd: not found
./run: line 4: exec: rsyslogd: not found
./run: line 4: exec: rsyslogd: not found
./run: line 4: exec: rsyslogd: not found
./run: line 4: exec: rsyslogd: not found

Anyone know what's up? I followed the Docker install guide to the letter and I'm running in a Digital Ocean droplet.

Posts: 6

Participants: 2

Read full topic

Julien Dumétier wrote:

At the request from @Maomao, as a translator we've got sometimes issues translating some messages, this thread will try to be the missing link between the dev team and the translators to help them provide the best translation possible.
Please @sam, @codinghorror, @zogstrip and the others... help us

Posts: 7

Participants: 4

Read full topic

Neil wrote:

How do you setup multiple forums running on the same server in the same rails instance? I'll tell you! In this example, I'll create a new site dinotalk.example.com operated by T Rex to talk about dinosaurs.

Setup a new database for the new site:

Create a new PostgreSQL user (trex in this example):

sudo su - postgres
createuser --createdb --superuser --pwprompt -Upostgres trex

Create a new database:

psql -c "create database dinotalk owner trex encoding 'UTF8' TEMPLATE template0"

Setup multisite.yml

Create a new file in the application directory: configs/multisite.yml

Add an entry for your new database. Be sure to give it a unique db_id value! Here's an example:

dinotalk:
  adapter: postgresql
  database: dinotalk
  username: trex
  password: smallarms
  host: /var/run/postgresql
  pool: 5
  timeout: 5000
  db_id: 2
  host_names:
    - dinotalk.example.com

If you're doing this in development mode, add an entry for dinotalk.example.com to your hosts file:

127.0.0.1 dinotalk.example.com

Now it's time to create the database objects.

Migrate the database

Run database migrations against the new database (and all multisite databases):

bundle exec rake multisite:migrate

Next, load some seed data into the database:

RAILS_DB=dinotalk bundle exec rake db:seed_fu

Note that we use the RAILS_DB env variable to choose which site to run commands against.

Make an admin user

Every discourse forum needs at least one admin user, and one of them needs to be used as the sender of system messages. It's a good idea to set a real person as the system user, instead of an unfriendly account called something like "admin" or "noreply".

u = User.new(email: 'trex@dinotalk.example.com', username: 'trex', name: 'T-Rex')
u.admin = true
u.save
EmailToken.confirm( u.email_tokens.first.token )
SiteSetting.system_username = 'trex'

This user doesn't have a password yet, so start up the server and use the "forgot password" link to set it. In development mode, just do bundle exec rails server as usual. No need for a RAILS_DB env variable. Then go to the host you set in your hosts file at the correct port (4000 if you're using vagrant). http://dinotalk.example.com:4000

That's it! Now you can have fun setting up your forum. Here are some tasks you'll need to do:

• Create categories
• Write category descriptions
• Configure some site settings (/admin/site_settings):
  • Facebook, Twitter, etc. credentials (facebook_app_id, twitter_consumer_key, etc.)
  • Use your own logo (logo_url, logo_small_url)
  • Set your company name so the terms of service uses it (company_full_name, etc.)
  • Add your Google Analytics tracking code (ga_tracking_code)
  • Configure Amazon S3 or Imgur for uploads if you want to use them (enable_s3_uploads, etc.)

See the Admin How To Guide for more detailed instructions.

Rails commands

You'll want to use RAILS_DB=dinotalk whenever you want to use one of the sites defined in multisite.yml. For example, to open a rails console and connect to dinotalk:

RAILS_DB=dinotalk bundle exec rails console

Without RAILS_DB, you'll be connecting to the database defined in database.yml. When launching a server process, you don't need to set this variable. The rails server command will handle all sites in the same process.

Posts: 18

Participants: 9

Read full topic

Grigori Vdmo wrote:

Hey Guys>>

Would love to see if anyone could offer some help or assistance. I recently got discourse going on digital ocean.
everything was running smoothly.

Then i got suggestion to upgrade from Your version: 0.9.8.11 to New version: 0.9.9.

I had followed the steps and then server started throwing 502 Bad Gateway/ nginx/1.4.6 errors.

Has anyone experienced this or could make some valuable suggestions what to do to fix this?

Much love in advance

Posts: 3

Participants: 2

Read full topic

Magnus Skog wrote:

I have a forum running in the default docker configuration on 12.04 as recommended. When I upgraded this morning to the latest version it seems to have stopped working. Nginx just gives me a 502 with nothing in the logs except the 502 error.

The server has a constant load of 2 or so and all the ruby processes like the unicorn launcher and bundle consume a lot of cpu.I tried destroying the docker container and rebuild it from scratch using the bootstrap command: same problem.

I have tried waiting but it has been like that for an hour now. A constant server load of about 2.

I can't find anything relevant in the logs either. Is there anything I can try before I try scrapping the host server as well?

I am running on Docker 0.9.1, Ubuntu 12.04.

Posts: 7

Participants: 4

Read full topic

Evaryont wrote:

As a staff member, I should be able to edit a user's post regardless of the state of their too-many-edits flag, no?

A user in my forum had split up a post into 31 posts, and then tried to edit each one. The forum has the default rate limit of 30 per 24 hour period. He reached out to me to edit the last one post, but I couldn't edit it as the "too many edits" alert popped up. Shouldn't my admin power trump that user's flag? Or not even trigger it, since I'm the one editing the post?

Posts: 5

Participants: 4

Read full topic