@LuisMendo Ok I ACTUALLY found the issue this time. What was happening was that the table was trying to be redrawn every time you entered a character. On Desktops this isn't a big deal because the browser can render quickly but this is an issue on mobile browsers. So what I've done now is that it doesn't render until 250ms after you stop typing. That way the typing should be relatively fluid and rendering occurs when it think that you're done. Let me know if that delay isn't long enough

@Suever Yes! Now I notice the speedup. I can also notice how it doesn't try to update until I have finished typing, which is good. Great job! As for the exact value of the delay, I don't know. Current value seems to be fine

Also, the other issue ("Run" button not responding) seems to have improved. Now only occasionally does it not respond, and when it doesn't it usually suffices to press the button again, without having to refresh the page (I have been trying it lately with quite a few animated ASCII art challenges)

@LuisMendo Glad to hear that we finally got to the bottom of the docs table issue. I'll go ahead and close the issue on Github

\o/

Yea I have it automatically reconnect without a page load. I'm monitoring "failures" on the backend though and they are still happening but definitely less frequently

And yea, hitting the run button again should make it work in most cases

Did you see? 6667 challenges currently :-)

I had noticed that you had used it, animated ASCII challenges are great stress tests for it

6667 challenges?

answered in MATL?

Yes. I even noticed how shortly after my answer it worked more slowly, probably because of more people trying it

@Suever Yup

@Suever Haha no, in total

That's awesome!

In MATL there were 800-something, last time I checked with your script

Ah ok

yea I have been working (slowly) on the analytics piece to the site

The script is great. You should make it public. Currently it's in a gist I think

I should re-write it in MATL ;)

Heh. For performance :-D

Pushed:

In Za the third input can now also be a logical scalar, and then it's interpreted like the second input