speed up logic for NodeViewModel.isViewable() logic by using a map instead of doing array.include()
throttling ko observable change triggers.
removed unnecessary observables.
cleaned up how refs are calculated for NodeViewModel. NodeViewModel.ref() is no longer depends on branchesAndLocalTags() and remoteTags().
introduced graph.isBigRepo() to disable some of the features
branches.js's branch sorting
animations
hover graphice ovrlay
graph.dropareaGraphActions to be computed
Can someone test on their big repos?? It is an improvements but I'm not sure if it is enough of an improvement.
Node count matters less because we incrementally load nodes. But the over all git ref count from show-ref -d command was the biggest reason for the slow down, and indirectly create nodes.
This PR was tested with a repo with 14498 refs and before ungit was unusable but now it usable without any big issues now.
closes: #1091
NodeViewModel.isViewable()
logic by using a map instead of doingarray.include()
NodeViewModel.ref()
is no longer depends onbranchesAndLocalTags()
andremoteTags()
.graph.isBigRepo()
to disable some of the featuresbranches.js
's branch sortinggraph.dropareaGraphActions
to be computedCan someone test on their big repos?? It is an improvements but I'm not sure if it is enough of an improvement.
Node count matters less because we incrementally load nodes. But the over all git ref count from
show-ref -d
command was the biggest reason for the slow down, and indirectly create nodes.This PR was tested with a repo with 14498 refs and before ungit was unusable but now it usable without any big issues now.