Trim off 1.15 seconds of page load and 1/2 mb of data by removing autocomplete data
The auto complete data is very large and completely unnecessary. It loads 462kb of data, 99% of which you will never use. It loads 2215 emojis, 2724 issues, 163 merge requests, and all 56 members. For now. It gets larger as we get larger. This query takes. 1.5 seconds (on gitlab.com). This data is used for the auto complete data so you can use #
to reference an issue or !
to reference a merge request or @
to reference a user or :
to reference an emoji.
We have two choices as I see it. Load the auto complete data 1 by one if they use it. So if someone types a @
we load the username data only (once). If someone types a #
we load the issue data only etc.
The other option is to querying on the fly. which could be very quick too. If we query on the fly we do a search via ruby for @job
and it returns {username:@JobV,name:Job van der Voort}
.
If someone never does an autocomplete then the data shouldn't ever be loaded.
I think this issue is of highest importance.