On Thu, Dec 5, 2024, at 10:45 PM, Nick Bowler wrote: > This implements support for NO_COLOR in test suites using AT_COLOR_TESTS > to enable colour by default. NO_COLOR is an informal standard[1] way > for users to express their preference to not use colours by default.
Thanks for the patch. > If NO_COLOR is set in the environment to anything other than the empty > string, the default should be "no colour" (but the command-line --color > option is expected to work as usual). I'm not sure I understand this. Could you please confirm whether the intended semantics are that the command-line --color option overrides NO_COLOR if both are used? And whether that is what your change actually does? Would you mind sending a revised patch that also includes documentation and a NEWS entry for this user-facing feature? Thanks, zw
