Improve mutation processing performance by eoghanmurray · Pull Request #1652 · rrweb-io/rrweb

Iterate over the added nodes in 'one pass' so that we don't need to backtrack; pushAdd requires that each new node has a parentId and a nextId

eoghanmurray

committed

Test changes, rearrangement of mutations

eoghanmurray

committed

Add some ids as I'm interested in tracing these nodes through pushAdd (console)

eoghanmurray

committed

Do away with the second pass as we can handle shadow DOM in the first pass

eoghanmurray

committed

Performance oriented refactor focusing on scenario where a large number of child nodes share the same parent

eoghanmurray

committed

Satisfy typescript which could be smarter here ... we can guarantee that if a possibly null node is in e.g. this.addedSet, then it is indeed not null. Similarly if `this.addedSet.size` is non zero,…

eoghanmurray

committed

Utilize `lastChild` to avoid possibly crawling through hundreds of nodes

eoghanmurray

committed

We've already got `nextSibling` here so can skip a step and avoid the confusing initialization to `IGNORED_NODE`

eoghanmurray

committed

Test rearrangements in the adds array due to new algorithm; should be equivalent but might have implications in terms of how replay should iterate over them

eoghanmurray

committed

We were calling `inDom` in all cases, so don't do the other ancestor checks if it goes bad early

eoghanmurray

committed

Don't think we're explicitly looking at the slimdom stuff in relation to mutation processing

eoghanmurray

committed

Add changeset

eoghanmurray

committed

Don't think `main` subfolder was ever used as an output target; this was preventing benchmark test from running

eoghanmurray

committed

Placate eslint (`while(true)` is a Pythonism rather than do..while) - could also use `checkLoops: allExceptWhileTrue` after eslint upgrade

eoghanmurray

committed

Forgot to add the mutation.html file - also add doctype

eoghanmurray

committed

Simplify the parentId check, doesn't need to ever by null

eoghanmurray

committed

Move mutation tests into their own files to demonstrate an idea which could be applied across all snapshots. Also, only mutation events are relevant here (reduce burden when manually expecting test…

eoghanmurray

committed

Apply formatting changes

eoghanmurray

committed

Some inconsequential tests to cover blocking scenarios

eoghanmurray

committed

Was trying to 'catch out' the mutation handling by having siblings present on mutation processing before they could be added, but couldn't (deliberately) break anything. Adding test anyway as it mi…

eoghanmurray

committed

fixup! Move mutation tests into their own files to demonstrate an idea which could be applied across all snapshots. Also, only mutation events are relevant here (reduce burden when manually expecti…

eoghanmurray

committed

I can't recreate a scenario for this case in testing, so add a warning and a speculative mitigation

eoghanmurray

committed

Put each snap file in it's own folder and shorten names

eoghanmurray

committed

Satisfy eslint

eoghanmurray

committed

Repeat the mutation tests but with the blocking/ignored nodes already in the DOM/mirror prior to mutation processing

eoghanmurray

committed

Indicate that replay no longer needs the queue, as added nodes should already be in the correct order for a single pass

eoghanmurray

committed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

Improve mutation processing performance #1652

Improve mutation processing performance #1652
eoghanmurray wants to merge 29 commits intorrweb-io:masterfrom
eoghanmurray:pushAddOrder

Commits on Mar 7, 2025

Iterate over the added nodes in 'one pass' so that we don't need to backtrack; pushAdd requires that each new node has a parentId and a nextId

Test changes, rearrangement of mutations

Add some ids as I'm interested in tracing these nodes through pushAdd (console)

Do away with the second pass as we can handle shadow DOM in the first pass

Performance oriented refactor focusing on scenario where a large number of child nodes share the same parent

Satisfy typescript which could be smarter here ... we can guarantee that if a possibly null node is in e.g. this.addedSet, then it is indeed not null. Similarly if `this.addedSet.size` is non zero,…

Utilize `lastChild` to avoid possibly crawling through hundreds of nodes

We've already got `nextSibling` here so can skip a step and avoid the confusing initialization to `IGNORED_NODE`

Test rearrangements in the adds array due to new algorithm; should be equivalent but might have implications in terms of how replay should iterate over them

We were calling `inDom` in all cases, so don't do the other ancestor checks if it goes bad early

Don't think we're explicitly looking at the slimdom stuff in relation to mutation processing

Add changeset

Don't think `main` subfolder was ever used as an output target; this was preventing benchmark test from running

Placate eslint (`while(true)` is a Pythonism rather than do..while) - could also use `checkLoops: allExceptWhileTrue` after eslint upgrade

Forgot to add the mutation.html file - also add doctype

Simplify the parentId check, doesn't need to ever by null

Move mutation tests into their own files to demonstrate an idea which could be applied across all snapshots. Also, only mutation events are relevant here (reduce burden when manually expecting test…

Apply formatting changes

Some inconsequential tests to cover blocking scenarios

Was trying to 'catch out' the mutation handling by having siblings present on mutation processing before they could be added, but couldn't (deliberately) break anything. Adding test anyway as it mi…

fixup! Move mutation tests into their own files to demonstrate an idea which could be applied across all snapshots. Also, only mutation events are relevant here (reduce burden when manually expecti…

I can't recreate a scenario for this case in testing, so add a warning and a speculative mitigation

Put each snap file in it's own folder and shorten names

Satisfy eslint

Repeat the mutation tests but with the blocking/ignored nodes already in the DOM/mirror prior to mutation processing

Indicate that replay no longer needs the queue, as added nodes should already be in the correct order for a single pass

Commits on Sep 5, 2025

Merge branch 'master' into pushAddOrder

Update packages/rrweb/src/replay/index.ts

Update packages/rrweb/src/record/mutation.ts

Uh oh!

Improve mutation processing performance #1652eoghanmurray wants to merge 29 commits intorrweb-io:masterrrweb-io/rrweb:masterfrom eoghanmurray:pushAddOrdereoghanmurray/rrweb:pushAddOrderCopy head branch name to clipboard

Commits

Commits on Mar 7, 2025

Commits on Sep 5, 2025

Improve mutation processing performance #1652
eoghanmurray wants to merge 29 commits intorrweb-io:masterfrom
eoghanmurray:pushAddOrder