16

Today I encountered an interesting problem with window.setInterval. When used with a sufficiently large delay (in this case the number of milliseconds in 30 days) it executes every second instead of every 30 days. Tested in latest Chrome and Firefox.

jsFiddle link

window.setInterval(function() {
    document.getElementById("first").innerHTML = new Date().toString();
}, 5000);
window.setInterval(function() {
    document.getElementById("second").innerHTML = new Date().toString();
}, 2592000000);

I couldn't find any authoritative documentation on the max value of a delay in setInterval, and the MDN documentation doesn't mention anything. Other sources online suggest that delay should be able to accommodate any signed 32-bit integer.

Does the delay parameter in window.setInterval have a maximum value and what is it?

tomfumb
  • 3,669
  • 3
  • 34
  • 50
  • A note: the standard *does not* restrict on type/ranges for the timeout argument https://www.w3.org/TR/2011/WD-html5-20110525/timers.html#timers (correct me if I'm wrong though) – zerkms Aug 17 '16 at 22:20
  • It does indeed seem to be at 2147483648 (the smallest positive integer that isn't a signed 32-bit integer) where it first occurs. – Paul Aug 17 '16 at 22:22
  • 1
    Explained in Aaron Dufour's answer here: http://stackoverflow.com/questions/12633405/what-is-the-maximum-delay-for-setinterval – Wake Aug 17 '16 at 22:25
  • Does not happen in Opera 12. – Bergi Aug 18 '16 at 20:11

1 Answers1

10

According to the setTimeout documentation on the public wiki MDN there is indeed a maximum, though it doesn't seem "official" - the limitation is a signed 32 bit integer.

Maximum delay value

Browsers including Internet Explorer, Chrome, Safari, and Firefox store the delay as a 32-bit signed integer internally. This causes an integer overflow when using delays larger than 2147483647, resulting in the timeout being executed immediately.

The value of 2592000000 is indeed larger than 2147483647 thus causing the overflow.

VLAZ
  • 26,331
  • 9
  • 49
  • 67
  • 4
    "According to the setTimeout documentation" --- it's not a "setTimeout documentation" but a public wiki – zerkms Aug 17 '16 at 22:21
  • 2
    @zerkms Since MDN is so much more readable than the official specification, it's generally considered a decent proxy. And the inclusion of implementation notes like this makes it more practical. – Barmar Aug 17 '16 at 22:24
  • 2
    @zerkms that would make it merely "not official documentation". However, it's a documentation nonetheless, because (with a risk of reaching semantic satiation) the page _documents_ the behaviour of `setTimeout`. It doesn't matter who wrote it, it still _is_ a writeup about the function. – VLAZ Aug 17 '16 at 22:29
  • 2
    @Vld I partially agree with you. For me it looks slightly misleading though. (when one speaks about standards - they need to be 200% accurate in terms) – zerkms Aug 17 '16 at 22:31
  • @zerkms fair enough - I've specified it's MDN. – VLAZ Aug 17 '16 at 22:32