Replace dependency on create-hash and create-hmac with bundled hash implementations

[jeskew]

This PR is a follow-up to #1857. create-hash and create-hmac bring in a transitive dependency that some customers have reported as causing build problems (e.g., #1857 (comment)). This passes the browser and react-native tests locally using karma + phantomjs, but it's worth trying in a few different browsers.

[codecov-io]

Codecov Report

Merging #1880 into master will increase coverage by 0.05%.
The diff coverage is 96.98%.

@@            Coverage Diff             @@
##           master    #1880      +/-   ##
==========================================
+ Coverage   95.91%   95.97%   +0.05%     
==========================================
  Files         214      219       +5     
  Lines        6933     7298     +365     
  Branches     1364     1398      +34     
==========================================
+ Hits         6650     7004     +354     
- Misses        283      294      +11

Impacted Files	Coverage Δ
lib/api_loader.js	100% <ø> (ø)	⬆️
lib/dynamodb/set.js	100% <ø> (ø)	⬆️
lib/browserHashUtils.js	72.22% <72.22%> (ø)
lib/browserHmac.js	94.28% <94.28%> (ø)
lib/browserSha256.js	97.7% <97.7%> (ø)
lib/browserSha1.js	98.92% <98.92%> (ø)
lib/browserMd5.js	99.24% <99.24%> (ø)
... and 2 more

---

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 8d63025...03c1019. Read the comment docs.

[chrisradek]

Changes look good. I tested this branch with react-native on iOS and verified that my application could load, the S3/V4 signers work, and tested several operations.

[jeskew]

The integration tests run without issue in IE, Edge, Chrome, and Firefox on my computer, though there is an unrelated failure running the unit tests in IE and Edge. I saw the same error with master, with this PR, and with version 2.177.0 (the last SDK release to include a full version of crypto-browserify), so this is not a new regression but something we should look into. I opened #1881 and assigned it to myself.

[lock]

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs and link to relevant comments in this thread.