Re: [patch 2.6.9 0/11] Add MODULE_VERSION to several network drivers

On Thu, Oct 21, 2004 at 11:17:49AM +0200, Arjan van de Ven wrote:

On Wed, 2004-10-20 at 20:11, John W. Linville wrote:

quoted

Patches to add MODULE_VERSION lines to several network drivers...

Here is the list:

have you checked if the version of these drivers is actually useful? (eg
updated when the driver changes) If it's not I'd say adding a
MODULE_VERSION to it makes no sense whatsoever.

Why do I feel like I'm being baited...? :-)

I would have to suspect that if a version string exists, that it has at
least some meaning to the primary developers/maintainters.  It certainly
is beyond my control to force the maintainers to give meaning to their
version strings.

Is this a political statement against the MODULE_VERSION macro and/or
its purpose?  I'm not overly interested in debating that one...

John
-- 
John W. Linville
linville@tuxdriver.com

On Thu, 2004-10-21 at 14:22, John W. Linville wrote:

quoted

have you checked if the version of these drivers is actually useful? (eg
updated when the driver changes) If it's not I'd say adding a
MODULE_VERSION to it makes no sense whatsoever.

Why do I feel like I'm being baited...? :-)

I would have to suspect that if a version string exists, that it has at
least some meaning to the primary developers/maintainters.  It certainly
is beyond my control to force the maintainers to give meaning to their
version strings.

Since the skeleton driver includes a define for that, I suspect your
assumption is a bit overly optimistic. 

Is this a political statement against the MODULE_VERSION macro and/or
its purpose?  I'm not overly interested in debating that one...

Not really. I have absolutely no problem with a MODULE_VERSION macro
*IF* the version it advertises means something. However if the version
it advertises has no meaning whatsoever (eg the version number never
gets updated) then imo it's better to NOT advertise anything so that
other tools (like dkms) don't make assumptions and decisions based on
nothing-meaning data.