[2481] in SIPB_Linux_Development
Re: Kernel interface changes (was Re: cdrecord problems on
daemon@ATHENA.MIT.EDU (Robert Thorncrantz)
Thu Feb 4 12:26:37 1999
Date: Thu, 4 Feb 1999 18:23:34 +0100
From: Robert Thorncrantz <rtz@pirx.df.lth.se>
To: Kev <klmitch@MIT.EDU>, yodaiken@chelm.cs.nmt.edu
Cc: Edwin Foo <efoo@MIT.EDU>, alan@lxorguk.ukuu.org.uk, warlord@MIT.EDU,
xiphmont@MIT.EDU, linux-kernel@vger.rutgers.edu, linux-dev@MIT.EDU,
jered@MIT.EDU, nemo@MIT.EDU, linux-scsi@vger.rutgers.edu,
cox@idecnet.com
In-Reply-To: <199902041411.JAA23283@x15-cruise-basselope>; from Kev on Thu, Feb 04, 1999 at 09:11:04AM -0500
On Thu, Feb 04, 1999 at 09:11:04AM -0500, Kev wrote:
> we're not talking about going from 2.0 to 2.2; we're talking about going
> from 2.0.33 to 2.0.34.
Perhaps you are, but other people have been complaining about things
changing between 2.0.x and 2.2.x, and between 2.1.x and 2.1.y.
> In the past, many changes have been made which
> have broken binary compatibility without warning and without good reason.
> *THIS MUST NOT CONTINUE HAPPENING* if Linux expects to get anywhere.
I'm a little curious, how many times have this happened actually? So
far all examples I've seen have are the same ioctl padding change
hashed over and over again. A change that looks more like a mistake to
me than a deliberate design change. (that is, the a new ioctl should
have been used, and the old one kept for compatibility at least for
the 2.0.x series, a method that has been used before, IIRC.) Mistakes
are bad enough, but it's not the same thing as doing arbitrary design
changes.
/Robert T.
--
Robert Thvrncrantz rtz@pirx.df.lth.se
Mundus Vult Decipi dat95rth@ludat.lth.se
dat95rth@student3.lu.se
