Issue 627

[CDNRae]

Fixes #627 by adding 3 tests to fs.write():

• Test that tries to use a non-existent fd
• Test that tries to open a file without the O_WRITE flag and then call fs.write
• Test that tries to write a buffer whose length is too small (buffer.length - offset < length)

[codecov-io]

Codecov Report

Merging #746 into master will increase coverage by 0.17%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master     #746      +/-   ##
==========================================
+ Coverage   86.71%   86.88%   +0.17%     
==========================================
  Files          16       16              
  Lines        1739     1739              
==========================================
+ Hits         1508     1511       +3     
+ Misses        231      228       -3

Impacted Files	Coverage Δ
src/filesystem/implementation.js	84.13% <0%> (+0.27%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 94d5ff8...01f3f1c. Read the comment docs.

[humphd]

This is great to get extra test coverage. I've included a few requests to make the checks more specific, and to clarify some things. Should be good to go after these fixes.

Thanks for working on these tests.

[CDNRae]

Made the changes -- hope these are good to go! Thanks for reviewing my work!

[humphd]

OK, this is looking good, thanks for making these fixes. I have one more I'd ask that you do, then I think it's ready.

[humphd]

fs.write takes an fd, so passing it a filepath as a string is really testing what we do when we don't pass an fd. If I do this in node, it throws. Probably returning EBADF is OK here, since it literally is a bad file descriptor.

That said, let's adjust the message on line 69 to be more clear: it('should fail if given a file path vs. an fd'

[humphd]

Looks great, thanks for working through these reviews.