Re: [PATCH] Draft for xattr support

[James Youngman]

On 4/4/07, address@hidden <address@hidden> wrote:
> > <attr/attributes.h> maynot be available on all platforms, so configure
> > should check for it.
>
> Yeah, that's a TODO for me.

I just committed a change to the polzer branch in the git repository
(http://repo.or.cz/w/findutils.git) which does this.   More generally
the change restore the ability to compile the tool on systems lacking
the xattr headers or library.

> > fnmatch_sanitycheck(); -- this only needs to be called once, ever.
>
> Alright, but currently each predicate supporting globbing seems to be
> responsible of calling it.

Hrm, they probably should not.   I guess it's not too bad to do it
once, at parse time, though.

> Maybe think of a way of centralizing that decision?

Yes. I will.  But separately.  Meanwhile, leave it in or not, as you see fit.



> Why certainly, hacking find's well-structured code is a breeze
> compared to other projects! :)
>
> Are you responsible for this, or did you find it when you got to
> findutils?

Mostly, it was already like that.   The data structures for parse
entries and the predicate table already existed.  I modified it a bit
to eliminate duplication of information (e.g. by adding all those
macros like PARSE_OPTION).

My main contributions have been in the areas of security, bugfixes +
tests, the ability to search deeper hierarchies (by switching to the
use of Jim Myering's fts implementation), POSIX conformance, new
predicates like -readable, -delete and -execdir.   The only major
refactoring was in the case of locate, in collaboration with Bas van
Gompel.

... that reminds me to remind you: don't forget the tests and documentation!

James.