Node: Add process.memoryUsage().toString for human-readable output

[Ginden]

Problem: result of process.memoryUsage() is hardly readable by humans, limiting it's usability for fast debugging.

Proposed solution: Add toString method to process.memoryUsage results.

Behaviour now: process.memoryUsage().toString() === "[object Object]"

Expected behavior:

/*
{ rss: 29233152,
  heapTotal: 7159808,
  heapUsed: 4383040,
  external: 8224 } */
process.memoryUsage().toString(); // "rss: 27.9M; heapTotal: 6.8M; heapUsed: 4.2M; external: 8K"

[mscdex]

I'm -1 on this kind of magical behavior. Besides, dividing the values by the appropriate amount to get the desired output is trivial enough in user land.

[cjihrig]

This would also be a breaking change for questionable upside. I'm also -1.

[Ginden]

This would also be a breaking change

Is there any code that relies on String(process.memoryUsage()) === "[object Object]"?

[mscdex]

Is there any code that relies on String(process.memoryUsage()) === "[object Object]"?

Quite possibly.

[addaleax]

Problem: result of process.memoryUsage() is hardly readable by humans, limiting it's usability for fast debugging.

To be clear – do you see the issue in the JSON format, or in the fact that the output uses “raw” decimal notation like 352994328, which is indeed a bit harder to read than 353M?

In either case, I’d see us more inclined to add a custom util.inspect overload rather than overriding toString()…

[jasnell]

A custom util.inspect overload is something I could :+1:

[apapirovski]

Given the feedback above, there's no way the original request will happen or make sense. I also really don't see the point of a custom util.inspect here.