test: buffer read functions tests for noAssert in one file

[larissayvette]

Checklist

• commit message follows commit guidelines

Affected core subsystem(s)

test

noAssert tested in buffer.read() functions

[mscdex]

-1 We already have tests that cover the buffer.read*()/buffer.write*() functions:

• test-readuint.js
• test-readint.js
• test-buffer-readuintle.js
• test-buffer-readuintbe.js
• test-readdouble.js
• test-readfloat.js
• test-writeuint.js
• test-writedouble.js
• test-writefloat.js

etc. ...

If anything, we should at least be combining these tests into two files: one for reads (e.g. 'test-buffer-read-numbers') and one for writes (e.g. 'test-buffer-write-numbers'). Perhaps we might even combine them all into one file (e.g. 'test-buffer-read-write-numbers').

[Trott]

-1 We already have tests that cover the buffer.read*()/buffer.write*() functions:

None of the existing tests cover noAssert functionality. This test covers it for all the functions.

I think we can use this as a starting point and combine the other read* tests into it. Either that or rename this test to add noAssert into the file name. I think either of those options are better than the alternatives that I can think of, which are:

• do nothing and leave code surface area uncovered
• add 12 separate files, one for each function, with all but one or two lines identical in each file
• add these tests to the existing somewhat-disorganized test files, copy/pasting nearly idential code in all of them
• etc....

[Trott]

(My vote would be add noAssert to the file name and leaving the other tests alone. Perhaps test-buffer-read-noassert.js or something like that.)

@larissayvette The commit message (and the PR description) don't contain any useful information. Can you please provide a description of the change? Perhaps something like this?: "This change tests the functionality of buffer.read*() functions when using the noAssert option."

[Trott]

LGTM with nits addressed (file name, commit message) and green CI

[Trott]

Nit: This would probably be better as something like: noAssert should not change return value for valid ranges

[larissayvette]

Alright doing those changes

[larissayvette]

@Trott I have updated the PR

[Trott]

@mscdex Are you still opposed to this given what I wrote above? If so, do you have an opinion as to a better way to introduce testing for the noAssert option for these functions?

[mscdex]

@Trott Additional code coverage is fine if that's really what this PR is doing, but at the same time I'm not keen on exacerbating the issue of test fragmentation for buffer.read*()/buffer.write*().

[Trott]

CI: https://ci.nodejs.org/job/node-test-pull-request/5802/

[Trott]

Landed in c8ed5f2. Thanks, @larissayvette! 🎉

Next steps:

• similar test for noAssert in buf.write*()
• refactor the buf.read*() and buf.write*() tests to at least have similar file names; might be good to combine some of them too