svn commit: r216119 - head

[Alexander Leidinger]

Quoting Ulrich Spoerlein <uqs at FreeBSD.org> (from Thu, 2 Dec 2010  
21:43:49 +0100):

> On Thu, 02.12.2010 at 15:45:22 +0100, Alexander Leidinger wrote:
>> Quoting Ulrich Spoerlein <uqs at FreeBSD.org> (from Thu, 2 Dec 2010
>> 10:46:05 +0000 (UTC)):
>>
>> > Author: uqs
>> > Date: Thu Dec  2 10:46:05 2010
>> > New Revision: 216119
>> > URL: http://svn.freebsd.org/changeset/base/216119
>> >
>> > Log:
>> >   Remove more duplicates.
>> >
>> >   Just so that I can lookup the command for the next time:
>>
>> Can you please add it to OldFiles.inc instead, with a comment what it does?
>>
>> >   for t in `make -V TARGETS universe`; do
>> >     __MAKE_CONF=/dev/null make -f Makefile.inc1 -m $PWD/share/mk  
>> TARGET=$t \
>> >     -V OLD_FILES -V OLD_LIBS -V OLD_DIRS delete-old | \
>> >     xargs -n1|sort|uniq -c|egrep -v "^ *1 ";
>> >   done
>>
>> While it does not matter for "-V OLD*", the spirit of the command is
>> more like the check-old target (which is including shared libs) than
>> the delete-old target (which does not delete shared libs but only
>> static libs). Using check-old would also remove the (wrong, but not
>> far fetched) fear of actually deleting something with this command.
>
> Something that bugs me more, is all the TARGET_ARCH conditionals in that
> file, that serve practically no purpose. If a file only existed on i386,
> and it has been removed on i386, there's no point in wrapping this under
> TARGET_ARCH=i386. The wrapping of all lib32 files for amd64 is a great
> example :/

My initial conditionals where only for things which where available  
everywhere, but got removed only on some platforms. Examples are the  
change in the threading libs, which where changed on most arches  
except for one or two, and the disk labeling things (IIRC the MBR  
stuff was removed from non-x86ish platforms). If this changed, it is a  
bug when you still have the initial design in mind.

> The wrapping is only ever required, if some file has been removed from a
> set of arches, but must be retained on others.

Correct.

> Anyway, please see attached.

Does not look bad...

Bye,
Alexander.

-- 
If at first you don't succeed, try, try again.
Then quit. No use being a damn fool about it.
		-- W. C. Fields

[Also attributed to Roy Mengot.  Ed.]

http://www.Leidinger.net    Alexander @ Leidinger.net: PGP ID = B0063FE7
http://www.FreeBSD.org       netchild @ FreeBSD.org  : PGP ID = 72077137