Re: [rfc] Colorized output for GNU make?

[Edward Welbourne]

>> The other thing I wonder about is the hardcoding of ASCII colorized
>> strings and the start/stop character strings (\033[...).  Are there
>> other methods of colorizing?
>
> Yes.

Indeed - most obviously, anyone running a build-'bot that reports its
build logs via the web (it's a common solution for continuous
integration setups, for example) will be interested in colouring by use
of HTML/CSS.  Of course, they can probably do that by streaming your
ASCII colorized stuff through a suitable perl script, but they might be
happier if they were in a position to do it directly.

The other reason to not hardcore ASCII colorizing is simply to leave the
user at liberty to chose which colours to use for which semantic types.
If you hard-code it, I'm guessing you'll also be hard-coding your
choices of how to colour which parts, where a system that permits user
customization may be preferable.  For example, you only have so many
colours available to you; and I'm guessing that a truly systematic study
of what different types of text are worthy of colouring would leave you
with more types than colours.  You'll thus end up conflating some types
of text, making them use the same colour, simply to fit the possible
types into the available colours.  A user with different interests than
yours might well want to separate types of text that you opt to
conflate, and be willing to conflate some you separated in order to free
up the colours to permit that.

Of course, how far you want to go towards flexibility may be limited by
the amount of palava and upheaval you'd have to inflict on the code to
attain it,

        Eddy.