Date: Wed, 28 Nov 2007 10:03:22 +0000 From: Alex Zbyslaw <xfb52@dial.pipex.com> To: Chuck Robey <chuckr@chuckr.org> Cc: User Questions <freebsd-questions@freebsd.org> Subject: Re: how to compile and install a new driver Message-ID: <474D3CEA.2020505@dial.pipex.com> In-Reply-To: <474CA85D.6080501@chuckr.org> References: <539c60b90711271248o6eb0b9fw757e40fbcdfdb3c2@mail.gmail.com> <20071127151259.H27150@wonkity.com> <474CA85D.6080501@chuckr.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Chuck Robey wrote: > Nice description, but you'd better include enough info so that they > could make FreeBSD-stype diffs: diff has the unfortunate default of > making an output that is compatible with ed(1). This supplies > extremely little information to use, in case the file you're trying to > patch with that diff has changed, and is also damned hard for mere > humans to understand. There are two other options you can give to > diff that change the format: -c gives "context" diffs, and -u gives > "unified" diffs, and the -u is the option that is standard with FreeBSD. It's not FreeBSD's default, it's POSIX's and has been the default behaviour of diff since forever (i.e. before -c or -u had even been invented); it is also the default behaviour of diff on every version of Linux I've ever used as well and probably every unix-like box you can get your hands on. Backwards compatibility *is* a wonderful thing. -u is a newbie to the diff world of options and is the strongly *preferred* way of providing patches, and probably not just for FreeBSD. It's generally the easiest for humans to read and the best for general use, though -c isn't that hard either and can sometimes be easier to comprehend. --Alex
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?474D3CEA.2020505>