Small Shift in Color Values

[Bastien Dejean]

Marc Lehmann a écrit :

> On Sun, Nov 27, 2011 at 02:33:56PM +0100, Bastien Dejean <nihilhill at gmail.com> wrote:
> > > >     URxvt.foreground: #9E9E9E
> > > 
> > > Your hardware is likely only capable of displaying colours #9d9d9d9d9d9d
> > > and #9e9e9e9e9e9e - urxvt picks the nearest colour to the one you specify
> > > (#9e009e009e00).
> > 
> > If I issue:
> >     printf '\033[38;5;247m%b\033[0m\n' '█'
> > The color of the corresponding block is seen as '#9E9E9E' by gpick.
> 
> Boy that must be some broken program - or maybe gpick samples an area
> or something and you are not aware of that? Doesn't seem to be the case
> though.

No, I set *oversample* to 0.

I've looked at grabc[1] (no other dependencies except X), the relevant
(bogus?) piece of code seems to be:

    r=(color.red >> 8);
    g=(color.green >> 8);
    b=(color.blue >> 8);

Where `color` is the *real* hardware color.

> In any case, I just installed gpick 0.2.4-1, and when using the picker

I use the latest mercurial version.

> thing in the lower right, I get "#9c9c9c" for both normal text and the
> block your command prints.

It should[2] be '#9E9E9E'.

I tried the following:

    URxvt.foreground: #FFFFFF

xmag sees it as (fefe, fefe, fefe).
(Why?)

[1] http://www.muquit.com/muquit/software/grabc/grabc.html
[2] http://www.calmar.ws/vim/256-xterm-24bit-rgb-color-chart.html

Greetings,
-- 
Bastien