process: improve hrtime() error message

[Trott]

Change error message from the form this format:

The "time" array must have a length of 2. Received length 0

...to this format:

The array "time" (length 0) must be length 2.

Checklist

• make -j4 test (UNIX), or vcbuild test (Windows) passes
• commit message follows commit guidelines

Affected core subsystem(s)

process errors test

[Trott]

Labeled semver-major because the CTC hasn't yet made a decision about whether these sorts of message changes are considered semver-major or not. (Decision expected in the next few days.)

HOWEVER, it does not really matter for this PR either way because the error message that this changes is not in Node.js 8.0 and is itself semver-major. In other words, whether we land this PR or not, the message is going to change in a semver-major fashion in Node.js 9.0

[lpinca]

Is there a particular reason apart from reducing the error message length? The original seems ok to me.

[Fishrock123]

Could be more concise, aren't we semver-minor for these now that we have the error codes thing?

Imo not worth the major change for this though, pretty much just the same thing?

[Trott]

Is there a particular reason apart from reducing the error message length? The original seems ok to me.

@lpinca In addition to reducing the error message length, I'm of the opinion that the current message is awkward.

I was originally going to fix that by simply changing The "time" array must have... to The array "time" must have..., but then I figure that if I was in there making a semver-major change anyway, I might as well improve it as much as I can.

[Trott]

Could be more concise, aren't we semver-minor for these now that we have the error codes thing?

@Fishrock123#13937 especially #13937 (comment)

There are at least 4 CTC members who believe these changes should still be semver-major at this time, and there are at least 2 who believe they should be semver-patch. So there is not consensus. Thus, I labeled this semver-major based on the "if there's any doubt, then it's major" principle.

[Trott]

Imo not worth the major change for this though, pretty much just the same thing?

@Fishrock123 As I wrote above, the semver major that introduces this error message in the first place has not landed. So the "not worth the major change" is not the right way to think about this, as the semver cost is zero. I'm making a (tiny) semver-major change to another semver-major change and they're both going to land in 9.0.

[Trott]

Needs one more @nodejs/ctc approval, so ping!

[Trott]

CI: https://ci.nodejs.org/job/node-test-pull-request/9240/

[refack]

Shouldn't it be must have length or must be of length?

[gibfahn]

+1 to must be of length

[Trott]

OK, updated with that text.

[mcollina]

LGTM

[evanlucas]

LGTM

[Trott]

New CI: https://ci.nodejs.org/job/node-test-pull-request/9257/

[Trott]

Landed in 43e105f