mirror of
https://github.com/discourse/discourse.git
synced 2025-01-18 11:42:53 +08:00
6493ddce17
We were doing get on Redis two times for each emoji while building the custom/standard/all lists which where resulting in ~3710 Redis calls. Given the emoji DB file is loaded in memory while we build/cache the emojis list this is unnecessary and slow. As a simplification in pseudo code here is an explanation of what we were doing: ```ruby emojis.each |emoji_name| aliases = get_aliases_from_redis_cache(emoji_name) is_tonable = get_is_tonable_from_redis_cache(emoji_name) build_emoji(emoji_name, aliases, is_tonable) end ``` The two redis calls are now simplified to a simple hash access: `@db[emoji_name]` |
||
|---|---|---|
| .. | ||
| assets | ||
| controllers | ||
| helpers | ||
| jobs | ||
| mailers | ||
| models | ||
| serializers | ||
| services | ||
| views | ||