Bug? Or does chrome now do weird things?

[justinbmeyer]

I'm trying to use this to build a web-locks polyfill. However, I'm seeing that two different browser tabs can acquire a lock at the same time. It appears they will both:

• Set X
• See Y is null
• Set Y as themselves
• And read X as themselves.

Obviously, that last set shouldn't be possible.

Any idea what's going on? I'm seeing this in chrome. Thanks!

[justinbmeyer]

Putting another setTimeout here:

setTimeout(()=>{

   // if x was changed, another client is contending for an inner lock
   let lsX = this.getItem(x);
   if (lsX !== this.clientId) {

seems to fix things. It seems to let chrome actually see the other tab's write.

[justinbmeyer]

Hello!

After playing with FastMutex for a while, I wasn't able to get it to work. I took some ideas from it and did a rewrite where that uses "only one setTimeout" at a time.

This seems to have made it work.

You can see the result here: https://github.com/bitovi/web-locks-polyfill.

I added a reference in the main function.

Thanks for building fast-mutex. It saved me at least 2 days of work.

[happy15]

any updates on this issue? @chieffancypants

[nieyuyao]

I'm trying to use this to build a web-locks polyfill. However, I'm seeing that two different browser tabs can acquire a lock at the same time. It appears they will both:

• Set X
• See Y is null
• Set Y as themselves
• And read X as themselves.

Obviously, that last set shouldn't be possible.

Any idea what's going on? I'm seeing this in chrome. Thanks!

Are these two tabs opened at the same time? @justinbmeyer

[happy15]

Putting another setTimeout here:

setTimeout(()=>{

   // if x was changed, another client is contending for an inner lock
   let lsX = this.getItem(x);
   if (lsX !== this.clientId) {

seems to fix things. It seems to let chrome actually see the other tab's write.

is this because that localStorage needs time to sync data?