Hi Chris,
On 2025-01-26 23:19:31+0100, Chris Hofstaedtler wrote:
> it appears commit 4bdf22803f9a41329694e0c3c82189635dccdcd0 breaks
> disabling colors using the documented mechanism of touching
> /etc/terminal-colors.d/dmesg.disable. See below for a user report.
Thanks for the report.
> On Sun, Jan 26, 2025 at 07:19:09PM +0100, Axel wrote:
> > Package: util-linux
> > Version: 2.40.4-1
> >
> > Running dmesg after upgrading to trixie started showing color. The manpage
> > mentions creating /etc/terminal-colors.d/disable which should disable colors
> > for all. Creating the also mentioned /etc/terminal-colors.d/dmesg.disable
> > doesn't disable color output also. I had /etc/terminal-colors.d/disable
> > configured which did disable colors on -stable.
>
> From what I can tell, the new rc = -ENOENT in colors_readdir
> introduced in 4bdf22803f9a41329694e0c3c82189635dccdcd0 bubbles up to
> colors_init:
>
> int rc = colors_read_configuration(cc); // calls colors_readdir, now
> returns -2
> if (rc)
> cc->mode = UL_COLORMODE_DEFAULT;
> else {
> Evaluating color scores is skipped, effectively ignoring the
> "disable" file.
>
> Could you please take a look?
Could you test https://github.com/util-linux/util-linux/pull/3397 ?
Thomas
