Few ideas...

Few ideas...

[Krzysztof Halasa]

Hi,

There was a thread about MODULE_MAINTAINER but how about a tag which
a) would be present in source files, not necessarily one per module
c) would be machine-parseable
d) would have _no_ effect on object code (i.e., would be ignored by
   the compiler completely)?

I guess it would be more useful than current MAINTAINERS file and we
would at last know immediately what to put in "To:" line.



Another idea? Why don't we move all the kernel-related lists to
vger.kernel.org, making them "open" in the process?



Want more? :-)
-- 
Krzysztof Halasa

Re: Few ideas...

[Jan Engelhardt]

On Feb 29 2008 23:48, Krzysztof Halasa wrote:
>Hi,
>
>There was a thread about MODULE_MAINTAINER but how about a tag which
>a) would be present in source files, not necessarily one per module
>c) would be machine-parseable
>d) would have _no_ effect on object code (i.e., would be ignored by
>   the compiler completely)?

d: I think it would not bad if it were included in the resulting
object file like MODULE_AUTHOR is.
If anything, MODULE_AUTHOR could be removed, because the original
author(s) are usually listed at the top of the .c file and not
always the ones to talk to when there is a bug (=> the maintainer
is).

Re: Few ideas...

[Krzysztof Halasa]

Jan Engelhardt <jengelh@computergmbh.de> writes:

> d: I think it would not bad if it were included in the resulting
> object file like MODULE_AUTHOR is.

I specifically don't want it in the binary - maintainers change, it's
not a point of contact for end users. It would be for source code ops
only.

> If anything, MODULE_AUTHOR could be removed, because the original
> author(s) are usually listed at the top of the .c file and not
> always the ones to talk to when there is a bug (=> the maintainer
> is).

Perhaps it's there because of copyright.
-- 
Krzysztof Halasa

Re: Few ideas...

[Jan Engelhardt]

On Mar 1 2008 00:17, Krzysztof Halasa wrote:
>
>> d: I think it would not bad if it were included in the resulting
>> object file like MODULE_AUTHOR is.
>
>I specifically don't want it in the binary - maintainers change, it's
>not a point of contact for end users. It would be for source code ops
>only.

So what apart from the parsability of a MODULE_MAINTAINER() tag,
what is different from authors engraving their name into a
comment at the start of the .c file?

>> If anything, MODULE_AUTHOR could be removed, because the original
>> author(s) are usually listed at the top of the .c file and not
>> always the ones to talk to when there is a bug (=> the maintainer
>> is).
>
>Perhaps it's there because of copyright.

A number of .c files (even those that are not just built-in, but can
also be built as =m) do not have a MODULE_AUTHOR(); and copyright is
even valid when there is no apparent author name.

Re: Few ideas...

[Krzysztof Halasa]

Jan Engelhardt <jengelh@computergmbh.de> writes:

> So what apart from the parsability of a MODULE_MAINTAINER() tag,
> what is different from authors engraving their name into a
> comment at the start of the .c file?

Authors != maintainers.

A script could parse the tags and fill Cc: appropriately.

> A number of .c files (even those that are not just built-in, but can
> also be built as =m) do not have a MODULE_AUTHOR(); and copyright is
> even valid when there is no apparent author name.

Sure, though it's a bit harder to find who is it exactly.
-- 
Krzysztof Halasa