Re: [PATCH v2 1/2] Introduce config variable "diff.defaultOptions"

Peff,

I'm replying to http://permalink.gmane.org/gmane.comp.version-control.git/108158

On Tue, 3 Feb 2009, Jeff King wrote:

All in all, this was a lot more complicated than I was expecting. Why isn't 
the behavior of "diff.primer" simply "pretend as if the options in diff.primer 
were prepended to the command line"? That is easy to explain, and easy to 
implement (the only trick is that you have to do an extra pass to find 
--[no-]primer). Is there some drawback to such a simple scheme that I am 
missing?

In order to answer your questions as convincingly as possible, I wrote up a 
one-page PDF document, downloadable here:
                http://preview.tinyurl.com/c769dd

You will see I clarified my arguments, and I found very compelling reasons for 
my design.  Also, BTW, v3 supports all diff options under the sun, instead of a 
limited subset.  That addresses your primary complaint WRT functionality.  
Please take a look at the PDF and I hope you agree.

                                 -- Keith