https://bugzilla.suse.com/show_bug.cgi?id=1218106 https://bugzilla.suse.com/show_bug.cgi?id=1218106#c13 Michal Koutný <mkoutny@suse.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Flags|needinfo?(systemd-maintaine | |rs@suse.de) | --- Comment #13 from Michal Koutný <mkoutny@suse.com> --- (In reply to Ancor Gonzalez Sosa from comment #8)

I did more experiments and SYSTEMD_LOG_COLOR seems to be irrelevant for this.

Right. (It's a gotcha since it applies to log-level coloring, not JSON coloring.)

which is not the case in YaST (we are executing the command and parsing the output).

I was previously focused on the env variables but the coloring based on tty/nontty output works for me. (Curiously, I don't see any recent changes in this area.) systemd-249.16-150400.8.35.5.x86_64 systemd-254.5-8.1.x86_64 (In reply to Stefan Hundhammer from comment #12)

The purpose of JSON is to be machine-readable, and escape sequences pretty much defeat that.

The reading machine parses a pipe or a saved file and those should be without color escapes. (The logic [1] should boil down to isatty(3).) Is it possible that something adds SYSTEMD_COLORS > 0 to the environment? (The variable seems to override tty detection.) [1] https://github.com/openSUSE/systemd/blob/071ac409a0564863657d8f8a5a35e6a4f91... -- You are receiving this mail because: You are on the CC list for the bug.