From mboxrd@z Thu Jan  1 00:00:00 1970
From: Jonathan Larmour <jlarmour@redhat.com>
To: "Trenton D. Adams" <tadams@theone.dnsalias.com>
Cc: 'eCos Disuss' <ecos-discuss@sourceware.cygnus.com>
Subject: Re: [ECOS] CONFLICTS that SHOULD be!
Date: Thu, 16 Aug 2001 16:01:00 -0000
Message-id: <3B7C50B9.83E43B72@redhat.com>
References: <000701c12699$2dc520b0$090110ac@TRENT>
X-SW-Source: 2001-08/msg00569.html

"Trenton D. Adams" wrote:
> 
> The following two items were required by the File I/O package.  When I
> removed that package they were not disabled.  As a result I got an error
> saying that the file or directory didn't exist.  In particular,
> <cyg/posix/signal.h>.
> CYGBLD_ISO_SIGNAL_NUMBERS_HEADER
> CYGBLD_ISO_SIGNAL_IMPL_HEADER
> 
> The following two items were required by another package which I think
> was File I/O as well.  When I removed the package they were not
> disabled.  As a result I got an error saying that the file or directory
> didn't exist.  In particular, <cyg/posix/limits.h> and/or
> <cyg/fileio/limits.h>
> CYGBLD_ISO_POSIX_LIMITS_HEADER
> CYGBLD_ISO_OPEN_MAX_HEADER
> 
> To be or not to be, that is the question!

Hmmm... yes, it would do that. This is an old chestnut that has caught us
out even internally before: should values be reinferred every time
something changes. Or perhaps even just when the user explicitly wants.
Right now we do neither :-|.

I think we should at least allow the user to force a reinference. However
customers tend not to pay for host tools, so very little development gets
done except in "spare time", so I don't know when that would appear. Sorry.
Of course, you have the source :).

Jifl
-- 
Red Hat, Rustat House, Clifton Road, Cambridge, UK. Tel: +44 (1223) 271062
Maybe this world is another planet's Hell -Aldous Huxley || Opinions==mine