Improve performance #5835

[math-GH]

Closes #5835

Changes proposed in this pull request:

• 1s => 0s waiting

How to test the feature manually:

1. mark an article as read
2. see that it is faster now

Pull request checklist:

• clear commit messages
• code manually tested

[Frenzie]

A timeout of 0 is meaningless. ;-) That's equivalent to removing the entire asBatch boolean and related functionality. The batching is for performance reasons. See #2199.

[math-GH]

0.1s would be better than 1s or 0s?

[Alkarex]

When the mark_read() is called, check the asBatch parameter. The idea is not to flood the server with requests but pool them (especially in the case of marking as read during scroll).
Manually clicking on the envelope to mark an article as read is not using the batch logic, so the code you have changed is not even called.
How do you mark an article as read? And please check in the Web browser debugger which code path is triggered.

[Alkarex]

Ah right, now I see that it was to address #5835
But the proposal was to make the logic smarter, not to remove it :-)

[Frenzie]

Manually clicking on the envelope to mark an article as read is not using the batch logic, so the code you have changed is not even called.

That sounds a bit suspect. That means that clicking 10 icons in a row isn't batched while pressing j or k 10 times in a row is. (Which is indeed the case.) It would seem that either the former should be too or the latter shouldn't either.