John H. Robinson, IV wrote:
> Christopher Smith wrote:
>>>
>> Okay, so again... why is all this going to break if you have a function
>> which accesses this string?
>
> Nothing is going to break. The end result is a lot of noise in the
> output of that function, when called. You also have the added expense of
> Yet Another Function.
>
> In the past:
>
> static char *RCSid="$Id$";
I have occasionally seen some criticism of such tags as undesirable
"garbage" or "clutter" .. mostly from SCM people. I suspect it does
cause some headaches for replicated or distributed systems.
Maybe this is one of those one-size-cannot-fit-all situations. I do like
the idea of always providing a user-accessible identifier. I also
appreciate the concern for identifying components. With an SCM system
that records system-wide ("checkpoint" -- is that the right term?)
identifiers, then embedding that number seems like it _should_ enable
tracing every contributing component. You do see more "build" number
information these days, eh?
The kernel practice of delivering configuration info with the kernel is
nice. Apache has some nice features for discovering configuration and
component information, does it not?
>
> Now:
>
> static char *RCSid="$Id$";
>
> char *get_module_version() {
> return RCSIdl
> }
>
> Perhaps it is not a lot, but it is certainly more than there used to be.
> I have always liked to embed the RCS $Id$ into my source code, and you
> will almost always see a $Id$ near the top of my code, including shell
> programs.
>
> Oh, make sure no two modules define the same get_module_version()
> function. That could be bad.
Regards,
..jim
--
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg
