From mboxrd@z Thu Jan  1 00:00:00 1970
From: Jonathan Larmour <jlarmour@redhat.com>
To: Ivan Jakab <ivan@INTESIS.hr>
Cc: ecos-discuss@sources.redhat.com
Subject: Re: [ECOS] pow.exe test failure
Date: Mon, 21 May 2001 07:36:00 -0000
Message-id: <3B0927F9.7D228228@redhat.com>
References: <000601c0e1f7$96348550$8200a8c0@ivan>
X-SW-Source: 2001-05/msg00335.html

Ivan Jakab wrote:
> 
> Hi all,
> 
> I'm in the process of configuring eCos for a MIPS R3041 based target board
> and I reached the point where I managed to execute all the test programs
> supplied with the eCos CVS repository. All of them passed except the pow.exe
> one which gave me the following output:
> 
> Network stack using 65536 bytes for misc space
>                     65536 bytes for mbufs
>                     131072 bytes for mbuf clusters
> CS8900: Init
> CS8900: Loopback test passed
> CS8900: - type: 630E, rev: 800
> INFO:<Starting tests from testcase
> //f/Projects/ecoscvs/ecoscvs-working/ecos/pac
> kages/language/c/libm/current/tests/vectors/pow.c for Math library pow()
> functio
> n>
> INFO:<Still crunching, please wait...>
> Vector #1
> FAIL:<error not set correctly> Line: 304, File:
[snip]
> 
> BTW, I have POSIX compatibility selected in my configuration. I looked at
> the code and there are some automatically generated vectors in pow.h (the
> expected errno value is set also). How come I get the FAIL:<error not set
> correctly>? Does this just have to do something with the configuration
> settings? Did anyone experience something similar?

This is known to work fine on other MIPS boards, and POSIX compatibility is
the default mode for libm, so I'm afraid I think you'll have to debug it.
I'd recommend recompiling the test without optimization (-O0) otherwise the
inline function in vector_support.h is virtually impossible to debug.

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