Handle EOF from multipart form file readers

[NathanBaulch]

I stumbled on a bug when trying to send a multipart file that happened to be a gzip reader downloaded from elsewhere. Looks like the multipart content type detection logic doesn't handle EOF correctly. This can also be easily reproduced by sending an empty file (bytes.NewReader(nil)).

Here's a self contained example that reproduces the issue with a gzipped string:

b := bytes.NewBuffer(nil)
w := gzip.NewWriter(b)
if _, err := w.Write([]byte("foo")); err != nil {
	panic(err)
}
if err := w.Close(); err != nil {
	panic(err)
}
r, err := gzip.NewReader(b)
if err != nil {
	panic(err)
}
_, err = resty.New().NewRequest().SetFileReader("file", "foo.bar", r).Post("http://foo.bar")
if err != nil {
	panic(err) // EOF!
}

---

Closes #431

[NathanBaulch]

Bump @jeevatkm

[jeevatkm]

@NathanBaulch Thank you for your PR appreciated, I'm sorry for the delayed response.

Travis CI build link was broken. I have migrated travis-ci.org to travis-ci.com

Can you please edit the PR to trigger the CI build?

[codecov]

Codecov Report

Merging #438 (716ebeb) into master (c1fa358) will increase coverage by 0.15%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master     #438      +/-   ##
==========================================
+ Coverage   96.66%   96.81%   +0.15%     
==========================================
  Files          10       10              
  Lines        1320     1320              
==========================================
+ Hits         1276     1278       +2     
+ Misses         25       24       -1     
+ Partials       19       18       -1     

Impacted Files	Coverage Δ
util.go	95.55% <100.00%> (+1.11%)	⬆️

---

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 c1fa358...716ebeb. Read the comment docs.

[jeevatkm]

@NathanBaulch It seems the test case is not covered for this PR update, can you please take a look at it?

https://app.codecov.io/gh/go-resty/resty/compare/438/diff#diff-dXRpbC5nbw==

[NathanBaulch]

The test case doesn't increase coverage because my fix is all about avoiding a false error. There is no new code path that the test could hit.

The test fails without the fix, passes with the fix and doesn't impact coverage, as expected. Hopefully that makes sense.

[jeevatkm]

@NathanBaulch Yes, it makes sense.
Actually, I was aiming to get coverage for the "error happened before reaching EOF" use case.

[NathanBaulch]

Not really relevant to my fix but sure, I've added a test that borrows brokenReadCloser from the request tests.

[jeevatkm]

@NathanBaulch Thanks for your PR.