[PATCH] color: add default colors for draft and secret changes

Matt Mackall mpm at selenic.com
Tue May 26 17:23:21 CDT 2015

On Tue, 2015-05-26 at 14:51 -0700, Kyle Lippincott wrote:
> On Tue, May 26, 2015 at 2:10 PM, Jordi GutiƩrrez Hermoso <jordigh at octave.org
> > wrote:
> > On Tue, 2015-05-26 at 14:04 -0700, Kyle Lippincott wrote:
> > > Not everyone uses a black background (I don't understand why not,
> > > but..); yellow is essentially unreadable on white background. The
> > > standard blue is too dark on black backgrounds, the standard yellow
> > > is too light on white backgrounds. The 'safe' colors in the 16
> > > normally representable that I've encountered are:
> >
> > This is a "theme" problem. All ANSI colours are meant to be readable.
> > Whether the "yellow" ANSI code is dark or light depends on the
> > terminal emulator. We must assume that all 8 ANSI colours have enough
> > contrast. It's really only with the 256 extended colours that contrast
> > has to be taken into consideration.
> >
> >
> >
> AFAICT ANSI has never stated what the individual color values need to be.

Nope.. because it was implicit in the design of the dirt-simple displays
that the scheme was created for: direct-driven 3- or 4-bit RGB systems
based on something like a Moto 6845 CRT controller like CGA. Between the
constraints of white balance and uncalibrated CRT gamma curves, the
colors were more or less defined. Historically, the only variables were
"how bright is the intensity bit (if present)" and "is it additive or
multiplicative". Everything in the wikipedia chart assumes palette
registers and DACs.. which postdates ANSI color by about a decade.

Mathematics is the supreme nostalgia of our time.

More information about the Mercurial-devel mailing list