[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: UTF8 problems with "find -ls" ?

From: Halim Issa
Subject: Re: UTF8 problems with "find -ls" ?
Date: Sat, 20 Dec 2008 23:41:45 +0100
User-agent: KMail/1.10.3 (Linux/; KDE/4.1.3; i686; ; )

On Saturday 20 December 2008 23:30:55 James Youngman wrote:
> On Sat, Dec 20, 2008 at 10:17 PM, Halim Issa <address@hidden> wrote:
> > On Thursday 18 December 2008 11:37:36 James Youngman wrote:
> >> The historic behaviour of -ls is to escape such characters.   See
> >> "Unusual Characters in File Names" in the info documentation for find.
> >
> > Thanks - It just didn't occur to me that "normal" unicode characters
> > would be "unusual" - was rather expecting that to cover escape sequences
> > and such...
> Sadly most changes to long-established behaviour are going to break
> things for _somebody_, and as far as I can tell, making test releases
> on alpha.gnu.org (like the current 4.3.x releases) doesn't seem to
> much prepare people for the change being introduced in the stable
> release series (currently 4.2.x).  People on this mailing list hear

Well - the main challenge from a complete non-coder point of view is that the 
current behaviour triggers find to be suspected as "non-UTF8-compliant" due to 

I guess it boils down to at which point unicode i so normal that it's the 
special handling (escaping) of unicode characters breaks / inhibits more 
scripts than it helps. The real challenge begins when most people need 
behaviour that doesn't differentiate between the plain old ascii characters 
and utf8 and start coding funny workarounds. That's when it really starts to 
fall apart when the problem is suddenly fixed...

Would it be an idea to include this in a new experimental release with a 
./configure option to revert back to the old behaviour?

Thanks again for quick and thorough response. I'm trying to track down all 
core applications with utf8 "challenges", and the response here tops the 
charts :-)

> about them and therefore know about the changes, but since most people
> who use findutils don't subscribe to this list, they're surprised by
> the changes when a new stable release series comes out.    It's
> possible that having shorter-lived development release series might
> help, but I'm not convinced.
> If we do make a change to the escaping, it would be useful to make
> things a little more consistent (for example, make -ls work more like
> -print, perhaps).
> James.

reply via email to

[Prev in Thread] Current Thread [Next in Thread]