Improve client's defaultsMiddleware performance

[tiagonapoli]

What is the purpose of this pull request?

Running tests on my computer I got around 50% improvement on times to get requests cached on memory.

To check the results you can:

• Clone https://github.com/vtex/dev-tools-experiments
• Run yarn on clients-cache-performance
• Follow the instructions on the readme and run for example node stressCLI.js @vtex/api 1000 100, which will show the time to do 100 gets using the lib specified in the @vtex/api entry (which is the master branch). Then run node stressCLI.js @vtex/api1 1000 100, which will show the time to do 100 gets using the lib specified in this entry.

You can generate flame graphs too to compare the two libs:

yarn global add 0x
0x stressCLI.js @vtex/api 1000 1000000

There are flamegraphs that shows how the ramda functions are "hot" (you'll have to clone the repository and access the following):
https://github.com/vtex/dev-tools-experiments/tree/master/clients-cache-performance/memoryCache/flamegraphs

Types of changes

• Performance improvement
• Bug fix (a non-breaking change which fixes an issue)
• New feature (a non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)
• Requires change to documentation, which has been updated accordingly.