test: add test coverage for fs.truncate

[christian-bromann]

Add test to check:

• for null as len parameter
• if error is propagated into callback if file doesn't exist
• if an error is actually thrown if len is not a number

I think it is safe to say that we could remove:

  } else if (len === undefined) {
    len = 0;
  }

because there is an validateInteger(len, 'len'); right after. Another option would be to just set len to 0 for all cases it is not a Number. What do you think?

Checklist

• make -j4 test (UNIX), or vcbuild test (Windows) passes
• tests and/or benchmarks are included
• commit message follows commit guidelines

[refack]

Hello @christian-bromann and thank you for the contribution 🥇

[refack]

the following:

node/lib/fs.js

Lines 617 to 622 in bcbb937

	if (typeof len === 'function') {
	callback = len;
	len = 0;
	} else if (len === undefined) {
	len = 0;
	}

Is a sort of JS idiom to handle optional function arguments. IMHO it could stay as as.

[trivikr]

CI (pending): https://ci.nodejs.org/job/node-test-pull-request/1​77​96/

[thefourtheye]

If we are expecting an error, its bettor to explicitly assert the error. But, I think you meant to write assert.ifError here.

[refack]

Since this passes CI, I'm assuming it fails with code ENOENT.
Assertion should have been near

node/test/parallel/test-fs-error-messages.js

Lines 521 to 545 in 9f7e3a4

	// ftruncate
	{
	const validateError = (err) => {
	assert.strictEqual(err.syscall, 'ftruncate');
	// Could be EBADF or EINVAL, depending on the platform
	if (err.code === 'EBADF') {
	assert.strictEqual(err.message, 'EBADF: bad file descriptor, ftruncate');
	assert.strictEqual(err.errno, UV_EBADF);
	} else {
	assert.strictEqual(err.message, 'EINVAL: invalid argument, ftruncate');
	assert.strictEqual(err.errno, UV_EINVAL);
	assert.strictEqual(err.code, 'EINVAL');
	}
	return true;
	};
	common.runWithInvalidFD((fd) => {
	fs.ftruncate(fd, 4, common.mustCall(validateError));
	assert.throws(
	() => fs.ftruncateSync(fd, 4),
	validateError
	);
	});
	}

But since it's not, I agree it should be here.

[christian-bromann]

@thefourtheye added explicit check for error

[refack]

CI: https://ci.nodejs.org/job/node-test-pull-request/17871/

[lundibundi]

Nit: maybe use some more descriptive filename that will indicate that this file is indeed non-existent and make sure that nobody creates it accidentally (via other test), perhaps: 'non-existent-truncate-file.txt'?

[thefourtheye]

Nit: this line should be aligned to the '

[christian-bromann]

It says Missing subsystem. which is odd because I started the commit with test:...

[addaleax]

/cc @Trott

@christian-bromann It’s not the first time that that happened, but I think we thought we had fixed it. I don’t think you need to do anything here, though.

CI: https://ci.nodejs.org/job/node-test-pull-request/17877/

[christian-bromann]

Maybe to add context to this situation: I used git commit --amend for all changes that happen after I made the initial PR

[refack]

Landed in 74f854e 🎉
Congratulations @christian-bromann on GitHub promoting you form

to

Hope we see you contributing more in the future 😉

[christian-bromann]

Hope we see you contributing more in the future 😉

Will try my best. Thanks for the support. I really enjoyed the code and learn session at Node+JS Interactive.