Re: PTHR.hdr failure digest for Debian Sample Implementation

[Andrew Josey <ajosey@opengroup.org>]

Hi  Matt

Again this is one of  those differences between a general
certification tool and a regression tool. At the moment
the tool is  designed to be compiler unawares, hence this
would be a FIP. The normal way we recommend handling this
is for use of a script to front end the compiler to remove
any warnings (or copyright notices or other annoying messages),
something along the lines of:

#!/bin/sh
#
# ccc - front end to cc, hides the warning messages
#

PID=$$
cc "$@" 2>/tmp/tmp$PID
retcode=$?
grep -v 'insert your warning here' /tmp/tmp$PID
rm -rf /tmp/tmp$PID
exit $retcode

Again this is something we need to have documented.
regards
Andrew

On Aug 20,  2:29pm in "Re: PTHR.hdr failure", Matt Taggart wrote:
>
> > Whilst glibc may be putting this advice out to application
> > writers , these are required POSIX functions and thus required for
> > implementations seeking POSIX compliance.
>
> Ok, since I don't expect glibc is going to change it's output could the tests
> be changed to be able to deal with this particular kind of output?
>