Re: [PATCH] About the :distant-foreground face attribute

[Jan D.]

Daniel Colascione skrev 2014-01-15 09:05:
> On 01/14/2014 10:33 PM, Jan Djärv wrote:
> > I have stated my justification as such: User and system
> > configurations should override Emacs settings and code in the default
> > case, i.e. when the user has made no explicit customization in
> > Emacs.
>
> You haven't stated any justification at all.  All you've done is
> re-assert your position.

And my position is what I say above.

> You clearly think it's okay for Emacs to
> override the foreground selection color in some circumstances, otherwise
> you'd support keeping the 24.3 behavior.

Yes I do, otherwise I wouldn't have fixed the bug report.

> You're justifying your
> particular strategy for handling contrast inherent in the new behavior
> by invoking a principle you *JUST VIOLATED* by creating the problematic
> new behavior.

No need to shout. There is no behaviour that is problematic AFAIK, you
are objecting on basis of design, not because the behaviour of letting
font lock faces shine through is problematic in some way.

I did not violate the principle by this code.  As I said, when there is
a choice between system/user defined color and generated colors, the
first should be preferred.  This does not in any way imply that
user/system colors shall always be used in all circumstances they might 
be used.  For example if we show font-lock foreground and
selection background we are not considering user/system color for the 
foreground because we are providing a feature (font-lock) where system
color is not applicable.

There exists a default font size for text in the system also.  We are
using that by default.  But info makes headers in a larger font size.
This does not mean that we have overridden the system font size, and
are violating the principle of using system font size.  It is a totally
separate feature where system font size is not applicable.  And so is
the case of marking font-locked text.

> We have to do something about contrast problems. Why do you think your
> solution produces better results than mine?

As I said before "better" is subjective, and from my point of view, it
has nothing to do with "better".

> If, to render something the
> user can actually read, we have to choose a foreground color other than
> what normal face logic would produce, why use the system selection
> foreground color instead of some other color we algorithmically create?
>
> Please don't invoke the "system settings should override Emacs" crap
> again: that idea clearly doesn't hold in the case we're talking about.

Yes it does.

        Jan D.