Clean up analyze_commits printout

[thislooksfun]

Running analyze_commits with more than one commit results in all but the first commit on the next line from the version number. Additionally, they all end in the | separator:

[11:20:19]: -----------------------------
[11:20:19]: --- Step: analyze_commits ---
[11:20:19]: -----------------------------
[11:20:19]: Found a tag v1.9.1 associated with version 1.9.1
[11:20:19]: Found 3 commits since last release
[11:20:19]: 1.9.1: chore(gitignore): ignore bundle-generated directories|
[11:20:19]: 1.10.0: 
feat: add option to hide version + commit printout|
[11:20:19]: 1.11.0: 
feat: clean up version path printout|
[11:20:19]: Next version (1.11.0) is higher than last version (1.9.1). This version should be released.

This change makes it so only the stripped subject is printed out:

[11:21:47]: -----------------------------
[11:21:47]: --- Step: analyze_commits ---
[11:21:47]: -----------------------------
[11:21:47]: Found a tag v1.9.1 associated with version 1.9.1
[11:21:47]: Found 3 commits since last release
[11:21:47]: 1.9.1: chore(gitignore): ignore bundle-generated directories
[11:21:47]: 1.10.0: feat: add option to hide version + commit printout
[11:21:47]: 1.11.0: feat: clean up version path printout
[11:21:47]: Next version (1.11.0) is higher than last version (1.9.1). This version should be released.

Additionally this adds a new option, show_version_path, which, if set to false (true by default) will disable the printout entirely:

[11:24:35]: -----------------------------
[11:24:35]: --- Step: analyze_commits ---
[11:24:35]: -----------------------------
[11:24:35]: Found a tag v1.9.1 associated with version 1.9.1
[11:24:36]: Found 3 commits since last release
[11:24:36]: Next version (1.11.0) is higher than last version (1.9.1). This version should be released.

---

Notes:

• The new option show_version_path, since it defaults to true, is fully backwards-compatible.
• Changing the printout format is a potentially breaking change, but only if people are scraping the command's stdout, which I don't think should count, as that is not supported behavior anyway.