Re: [PATCH] 256 color support for libvte (and so gnome-terminal)

[Daniel Brockman <daniel brockman se>]

Havoc Pennington <hp redhat com> writes:

> On Mon, 2006-02-27 at 07:57 +0100, Daniel Brockman wrote:
>> I wasn't sure where to send this, but this list seems
>> close enough.  Yell at me if it isn't.
>
> Should also file it in bugzilla probably. 

Damn it, I *knew* I was forgetting some important
communications channel!

> I don't know anything about 256 color mode, but a couple of generic
> style things:
>
>  - this isn't X, so you don't have to put every patch in an #ifdef 
>    ;-) (if there's some genuine reason to turn this off, it should 
>    probably be runtime configurable if anything)

Okay.  The reason I put it in ifdefs was a suspicion that
some people who don't care for the feature at all might not
want to waste a byte in a data structure that exists in one
copy for each character cell.  (Say you have a few terminals
with a few thousand lines of scrollback each.  Well, that's
maybe a megabyte of wasted space.)

There's not much point in turning the feature off at runtime
unless that also makes the data structure smaller.

>    (I think the #ifdef is just popular in X because it was 
>    consortium-run and the vendors didn't want to argue about everything 
>    so they just made a big #ifdef mess)

I'm actually not familiar with the X codebase. :-)

>  - "diff -u" - much easier to read, set it in your .cvsrc and you'll
>    be glad you did

Some people will say the exact opposite (assuming you can
consider context diffs the opposite of unified diffs).

-- 
Daniel Brockman <daniel brockman se>