refactor: just disable color output rather than tracking terminal width

[G-Rath]

This lets us drop some conditionals - weirdly it doesn't seem to work if we move the conditional text.DisableColors call into reporter.New even though that does work for the vertical reporter...

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 65.40%. Comparing base (556817b) to head (86bff5b).

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1087      +/-   ##
==========================================
- Coverage   65.40%   65.40%   -0.01%     
==========================================
  Files         150      150              
  Lines       12518    12521       +3     
==========================================
+ Hits         8188     8189       +1     
- Misses       3872     3873       +1     
- Partials      458      459       +1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[another-rex]

Thanks!

[another-rex]

This lets us drop some conditionals - weirdly it doesn't seem to work if we move the conditional text.DisableColors call into reporter.New even though that does work for the vertical reporter...

Maybe it's because there multiple reporter.New() calls every run, and the disable color function is a single global variable.

[G-Rath]

yeahhhhh but then why would it work for the vertical output...? something odd is def going on but I think it's probably best we deal with that once we've landed the vertical report since we plan to shuffle things around a bit re table anyway right

[G-Rath]

Weirdly too it seems like it's a very stable global? I've run the test suite a ton and it never seems to give different results like I'd expect if this is an unstable use of a global...?

(fwiw I vote we land this for now - we can easily revert if it turns out to be flakey)