Emoji request - :whoosh:
-
[Uploading...]()
-
This is some impressive bikeshedding.
-
we learned from the master
-
Which looks better:
(the one that's currently being used, on the left, isn't square and the animation is kind of jerky.)
-
Hmm, your fixed version is slightly better.
-
Basically all I did was slightly change the positions of all the bottom layers (up to the point where the nose of the plane starts coming up).
If you look at them both full-size you'll be able to better see it.
-
It's a good thing we scared Jeff away, otherwise we would now have gotten a rant from him about how animated gifs for emoji are and all that.
And he'd tell his minion @sam that they should restrict emoji to PNG only which should be enforced server side.
-
And he'd tell his minion @sam that they should restrict emoji to PNG only which should be enforced server side.
Which should be hosted on a CDN.
-
Which should be hosted on a CDN.
Makes you wonder why they aren't using the emojione CDN ...
-
Probably because they didn't create it, so they can't be sure that it's sufficiently WTFy for Discourse.
-
Probably because they didn't create it, so they can't be sure that it's sufficiently WTFy for Discourse.
What they should do is set up emojicdn.discourse.fuckyou.
A request to that does a database lookup for the matching DiscourseEmjoy to Emojione CDN.
It then launches a server-side web request to fetch the matching emoji from Emojione, and downloads it to the DiscourseEmojiCDN
Then it renames the file to be emoji-userid-gitversion-datetime.png so that it's remotely cached per user.
It then does a date-time comparison against all other emjois for the same user and same gitverison, but different dates. If the one it just downloaded is more newer than the old ones, it then does another web request to get a fresh copy, and sets that as the stack-topped emoji.
Then it runs it through ImageMagik to resize it to an optimized size (increasing the size by 400kb, of course).
Then it serves it.
-
@Lorne_Kates said:
A request to that does a database lookup for the matching DiscourseEmjoy to Emojione CDN.
It then launches a server-side web request to fetch the matching emoji from Emojione, and downloads it to the DiscourseEmojiCDN
Then it renames the file to be emoji-userid-gitversion-datetime.png so that it's remotely cached per user.
It then does a date-time comparison against all other emjois for the same user and same gitverison, but different dates. If the one it just downloaded is more newer than the old ones, it then does another web request to get a fresh copy, and sets that as the stack-topped emoji.
Then it runs it through ImageMagik to resize it to an optimized size (increasing the size by 400kb, of course).
Then it serves it.
This is all sounding disturbingly familiar, though it didn't save the version number in the table....
-
@Lorne_Kates said:
Then it runs it through ImageMagik to resize it to an optimized size (increasing the size by 400kb, of course).
It also needs to fuck up greyscale images for that authentic Disco-inept touch.
-
@Lorne_Kates said:
Then it runs it through ImageMagik to resize it to an optimized size (increasing the size by 400kb, of course).
It also needs to fuck up greyscale images for that authentic Disco-inept touch.
And then one day, a Discodev adds a new Emoji. But doesn't add it right. And it permanently and irrevocably fucks up all Emoji references that have ever been used up to that point.
-
Don't forget the part where you can't render static PNG images without 50MB of JavaScript.