Summary:
There was a frustrating issue previously where the UI would jump everytime a new frame came in with bloks data. This would occur constantly as the client was generating new frames every 10 seconds. The issue was because the previous appraoch to speed up the augmentation was to display the raw client raw and then asynchronously process a new frame with the added derived components.
The issue was the raw client frame has less nodes so we would breifly display this, and because the reduciton trace was cache on the next tick we would show this causing a jump.
Now we have a new solution:
1. We use an event emitter to push frames / metadata into the interceptor
2. The interceptor should always produce 1 frame, this can be the same or augmented, for open source its just the same
3. We use react query to check if we already have the reduction trace cached, if so then we are able to quickly process the whole thing, if not its split into 2 parts.
React query is also nice since we can add a cache time for the reduciton traces which are very large and not useful one your screen is gone. This should help keep memory usage under control a bit.
one other benefit to this approach is i can remove the error handling stuff from the main plugin code, since the blok augmentation is not on the critical path and it can retry internally and push out a new frame whenever.
Reviewed By: antonk52
Differential Revision: D49272152
fbshipit-source-id: e5539d36231a32754e8612014195449b9faafdb1
Summary:
UINode has never been a good name, we have 3 versions of a node.
ClientNode Previously UINode (the raw data from the client)
NestedNode (for the visualiser)
TreeNode (extends ClientNode and adds stuff specific to the tree like indentation and expanded states)
Arguablely we dont need nested node but that is another story
Reviewed By: elboman
Differential Revision: D47547529
fbshipit-source-id: 9a3b119d1230ea7b6734e7a3270c28287b04faf1
Summary: This file was getting overwelhming, and this seemed like a reasonable way to split it
Reviewed By: elboman
Differential Revision: D47547532
fbshipit-source-id: ab2bfa22daabbed13ec1445da0cf8ba88bda12d7
Summary:
There were a small number of styles / attributes that were missing, this addresses this.
Similar to the bloks minification map there are some special hardcoded reserved ids, if that doesnt work then you can look into a special place for each style
I also refactored the transform metadata method as it was getting very nested and hard to understand.
Reviewed By: aigoncharov
Differential Revision: D46799392
fbshipit-source-id: 8de8a2740f45ddb23a1bc9abb7bb36f969cf1fbb
Summary:
In order to support this the stream inceptor transform nodes now is able to return a list of new meta enties, since these are new attributes we try to mimic what would have happened if they were generated on the client. This makes the rest of the logic downstream simpler
Introduced metadata register, same idea as on the clients.
The attributes available are a mixed bag, will work with blok server to imporove
Reviewed By: antonk52
Differential Revision: D45177781
fbshipit-source-id: 9d761b2f682e7e0fd4710f5b2e9d9d6ff26741fc
Summary:
Added stream interecptor which gets a chance to augment the messages off the wire. Stream interceptor transformations are async and can fail due to network errors so added error state with a retry button. The retry button will just call the function again.
I am also handling errors better generally when this method fails unexpectedly, logging more clearly what went wrong and communicating it to the user
Did some refactoring of subtree update event to support this
Reviewed By: lblasa
Differential Revision: D44415260
fbshipit-source-id: a5a5542b318775b641d53941808399a8fa4634d3
