Factor out baseline calculation code from textreporter, for use in pretty and chart

[jdmarshall]

I like your chart reports but humans can sometimes struggle a bit even with bar charts. So I wanted to append the faster/slower logic from textReporter.

Then I noticed some redundancy with pretty, and ended up here.

Before:

 ⇒ prom-client@latest                         | ██████████████████████████████ | 15,686,573 ops/sec | 11 samples
 ⇒ prom-client@trunk                          | █████████████████████████----- | 12,871,014 ops/sec | 12 samples
 ⇒ prom-client@current                        | ██████████████████████████---- | 13,426,771 ops/sec | 10 samples

After:

 ⇒ prom-client@latest                         | ██████████████████████████████ | 14,303,655 ops/sec | 12 samples (baseline)
 ⇒ prom-client@trunk                          | ██████████████████████████████ | 14,280,376 ops/sec | 10 samples (1.00x slower)
 ⇒ prom-client@current                        | ██████████████████████████████ | 14,343,253 ops/sec |  9 samples (1.00x faster)

[RafaelGSS]

I liked the change on chart.js, but it seems the text.js is breaking. See full diff comparison on: https://gist.github.com/RafaelGSS/9b2e7d8d48d0a866a2cc283667ff9322

• text reporter shouldn't have a \n between each one of them.
• It seems you created the baseline analyzer, but you aren't using it on pretty reporter

Could you provide the screenshot when you change something? Otherwise it takes some time for me to checkout your PR, test the changes, and compare the result with main.

[jdmarshall]

I wonder what the difference in settings is between prom-client and bench-node, because Webstorm lights up with every formatting issue with the former and I'm getting no warnings locally from the latter.

[jdmarshall]

Finally got past the linter. Woof.

[RafaelGSS]

@jdmarshall, would you mind providing the graphic difference between the main and this PR? I mean, what changes in the screen (for pretty, text, and chart report)? This would help people when looking at the CHANGELOG and ending up on this PR.

[jdmarshall]

In the PR or in the commit message?