From: Junio C Hamano <junkio@cox.net>
Date: Tue, 19 Apr 2005 22:45:02 GMT
Subject: Re: [GIT PATCH] I2C and W1 bugfixes for 2.6.12-rc2
Message-ID: <7vpswqpes1.fsf@assigned-by-dhcp.cox.net>
In-Reply-To: <20050419222609.GE9305@pasky.ji.cz>

>>>>> "PB" == Petr Baudis <pasky@ucw.cz> writes:

PB> I'm wondering if doing

PB> if [ "$(show-diff)" ]; then
PB> 	git diff | git apply
PB> else
PB> 	checkout-cache -f -a
PB> fi

PB> would actually buy us some time; or, how common is it for people to have
PB> no local changes whatsoever, and whether relative slowdown of additional
PB> show-diff to git diff would actually matter.

"show-diff -s" perhaps.  Also wouldn't it be faster to pipe
show-diff output (not git diff output) to patch (not git apply)?