Does the track method of Reaction have unnecessary batch operations?

[dSalieri]

I watched code base of Reaction class and some his dependencies and I've seen the next:

/// I omitted detailed code
track() {
  startBatch();
  trackDerivedFunction(this, fn, undefined);
  endBatch();
}

function trackDerivedFunction(derivation, f, context) {
  globalState.inBatch++;
  f.call(context);
  globalState.inBatch--;
}

So my question is the next: Why does the track method have batch operations, if we have operations inBatch++, inBatch-- inside trackDerivedFunction?

If startBatch and endBatch has important sense in track, please explain it, desirable with examples on real examples.

[mweststrate]

Sorry, most of it has been written 5+ years ago, so the details are not top of mind. (Why care?) probably has to do with updates not immediately starting new reactions but waiting until the current one is finished. If you want to know more, I recommend to disable the lines and run the unit tests, likely some will fail.

…

On Fri, 8 Sept 2023, 21:53 dSalieri, ***@***.***> wrote: I watched code base of Reaction class and some his dependencies and I've seen the next: /// I omitted detailed codetrack() { startBatch(); trackDerivedFunction(this, fn, undefined); endBatch();} function trackDerivedFunction(derivation, f, context) { globalState.inBatch++; f.call(context); globalState.inBatch--;} So my question is the next: Why does the track method have batch operations, if we have operations inBatch++, inBatch-- inside trackDerivedFunction? If startBatch and endBatch has important sense in track, please explain it, desirable with examples on real examples. — Reply to this email directly, view it on GitHub <#3754>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAN4NBCHSV3JBPR7R23ACOTXZNZSFANCNFSM6AAAAAA4Q2R6YQ> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[dSalieri]

@mweststrate Only one operation runs the track method (via onInvalidate_) and this is runReaction_ (it also has startBatch and endBatch). So I think the track method has useless batch operations (my opinion) and probably trackDerivedFunction also has unnesessary inBatch increment and decrement.

Tell me how I can run the tests you are talking about.

P.S I am investigating mobx architecture and I what to know how this works.

[mweststrate]

Check out the repo, yarn, yarn test

…

On Sat, 9 Sept 2023, 07:40 dSalieri, ***@***.***> wrote: @mweststrate <https://github.com/mweststrate> Only one operation runs the track method (via onInvalidate_) and this is runReaction_ (it also has startBatch and endBatch). So I thing the track method has useless batch operations (my opinion) and probably trackDerivedFunction also has unnesessary inBatch increment and decrement. *Tell me how I can run the tests you are talking about.* P.S I am investigating mobx architecture and I what to know how this works. — Reply to this email directly, view it on GitHub <#3754 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAN4NBC6BP5NKYGM2VQ2XBDXZP6OLANCNFSM6AAAAAA4Q2R6YQ> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[dSalieri]

Hmm, interesting... At first glance it seems that there are unnecessary batchings, but in fact everything is to the point.

It's sad that there aren't comments for some parts of code base of mobx

P.S Thanks that you responded me.