The way blockers are called in v.5's replace/push broke our app

[BingWangEd]

We upgraded to History 5 and we found history.replace was not able to update the history. I dug into the source code and this is what I found.

In v.5, when history.replace is called here, it would check allowTx. If my understanding is correct, allowTx would always return false if there’s any functions push to blockers.

This is a behavioral change from v.4, in which the replace function here would still update history.enries, if the result from blocker function here is not false.

Please correct me if I got anything wrong, but when I commented out any callback we push to blockers via history.block, the app indeed loads as expected. Is this an intentional change? I don't think I found it in documentation or in the undocumented one.

[StringEpsilon]

Hi, the gist and github issue you linked are not the official documentation. Just something I compiled.

The change in blocking behavior is alluded to in the 5.0 release notes here

Better history.block API, with support for retrying transitions

and the version 5 documentation here. I've linked that latter one in my gist under the "Blocking transitions" headline, stating that

The entire blocking system has been redesigned from scratch and is therefore completely incompatible.

[BingWangEd]

I see. Thanks for the quick response! Closing this issue now