From Tom.Garland@Ireland.Sun.COM  Fri Jun 20 11:43:15 1997
Received: from venus.Sun.COM (venus.Sun.COM [192.9.25.5]) by dkuug.dk (8.6.12/8.6.12) with ESMTP id LAA04706 for <i18n@dkuug.dk>; Fri, 20 Jun 1997 11:43:13 +0200
Received: from sunmail1.Sun.COM ([129.145.1.2]) by venus.Sun.COM (SMI-8.6/mail.byaddr) with SMTP id CAA09102 for <i18n@dkuug.dk>; Fri, 20 Jun 1997 02:42:40 -0700
Received: from Ireland.Sun.COM by sunmail1.Sun.COM (SMI-8.6/SMI-4.1)
	id CAA04721; Fri, 20 Jun 1997 02:42:36 -0700
Received: from ireserver by Ireland.Sun.COM (SMI-8.6/SMI-SVR4-sd.fkk108)
	id KAA03725; Fri, 20 Jun 1997 10:43:04 +0100
Received: from elc4.ireland.sun.com by ireserver (SMI-8.6/SMI-SVR4-su.fkk201)
	id KAA08658; Fri, 20 Jun 1997 10:43:04 +0100
Received: from elc4 by elc4.ireland.sun.com (SMI-8.6/SMI-SVR4)
	id KAA12467; Fri, 20 Jun 1997 10:43:02 +0100
Message-Id: <199706200943.KAA12467@elc4.ireland.sun.com>
Date: Fri, 20 Jun 1997 10:43:02 +0100 (BST)
From: Tom Garland SMI European Software Centre <Tom.Garland@Ireland.Sun.COM>
Reply-To: Tom Garland SMI European Software Centre <Tom.Garland@Ireland.Sun.COM>
Subject: localedef hebrew problem
To: i18n@dkuug.dk
MIME-Version: 1.0
Content-Type: TEXT/plain; charset=us-ascii
Content-MD5: ARyO43e6IhZHJUnNNguJBw==
X-Mailer: dtmail 1.2.0 CDE Version 1.2_30 SunOS 5.6 sun4m sparc 

Hi,
can anyone help with this problem?

Problem:

Even though I defined hebrew characters in the LC_CTYPE/alpha section
of the hebrew locale definition file, a call to 'isalpha(c)' with an
Hebrew character return false. The same problem occures with 'graph' and
'print' sub sections as well.

Why does it happen?:

1. /usr/include/ctype.h header.
   This header has an #ifdef which forces a primitive macro, instead
   accessing the libc is*class*() routines. I had to modify a local copy of
   ctype.h to overcome this.
   
2. localedef ignores alpha,graph,print sections.
   I suspect localedef does not take into acount the definitions
   in the alpha/graph/print sections. The contents of he.chrtbl.c file,
   created by localedef, seem to support this suspicion.
   
Is this standard localedef behaviour?

tom


