X-Authentication-Warning: delorie.com: mail set sender to djgpp-workers-bounces using -f
X-Recipient: djgpp-workers AT delorie DOT com
Message-ID: <52854404.6080802@gmx.de>
Date: Thu, 14 Nov 2013 22:43:32 +0100
From: Juan Manuel Guerrero <juan DOT guerrero AT gmx DOT de>
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:16.0) Gecko/20121025 Thunderbird/16.0.2
MIME-Version: 1.0
To: djgpp-workers AT delorie DOT com
Subject: Re: DJGPP tests update
References: <526E3D35 DOT 7010600 AT iki DOT fi> <527640BE DOT 30500 AT gmx DOT de> <5276A3FD DOT 4090005 AT iki DOT fi>
In-Reply-To: <5276A3FD.4090005@iki.fi>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-Provags-ID:  V03:K0:Sy8V6U+6AL50G5eXGKz5RzajzB5aefngeMReoqtPo9lncgTIlKl
 cGoYPij9g6V17zGRA4iTXQWwkj06XskcnoYJMtil1+wDkAZzoTNsv+qLl7VwCb1lUqztn3E
 x4lAaXQtrGk7UGqJOkr1yUdk3e0ihmQzKDu25MvF7S6/QYmNxfvOT5wPhjD0dGTCOsZIfx0
 S2LGpu5lRNJahe8TiOByQ==
Reply-To: djgpp-workers AT delorie DOT com

Am 03.11.2013 20:29, schrieb Andris Pavenis:
> On 11/03/2013 02:25 PM, Juan Manuel Guerrero wrote:
>> Am 28.10.2013 11:32, schrieb Andris Pavenis:
>> [snip]
>>> cygnus tests shows problems with strtod*: the returned
>>> tail part (second parameter) does not match the expected
>>> value for part of tests.
>>
>> Sorry, may be I am missing something but can you show
>> me the place where it happens so I can try to fix it.
>>
> From unchanged test sources I'm getting something like:
>
> strtod:13 inacurate answer: (00000001 should be 00000000)
>
> and many following in the same way for both strtod and strtodf.
>
> Adding debugging output to strtod test shows that
> the return value is correct, but the returned pointer to
> tail (second parameter) is not was is expected
>
> Tested for some given strings under Linux and got the expected
> values of tail for those which I tested.
>
> This is however not too serious as the problem is for such strings
> that could be considered as abnormal.
>
> Andris
>
> PS. Earlier tests was with 1st build of gcc-4.8.2 for DJGPP. Now tested
> with gcc-4.9-20131101 and build went OK and the test results were
> the same. So it looks we will not need to patch DJGPP sources for
> upcoming gcc-4.9
>

OFYI: I have fixed the issue in stro[d|f|ld] functions.  The cygnus checks now
pass.  I have also added some checks for the newly supported hex format.  But
more hex format specific checks have been implemented in some other files.  I
have also added a check for strtof.  Because the cygnus check suite does not
support checks for long double I have not added any strtold specific checks
to the cygnus check suite.

Regards,
Juan M. Guerrero


2013-11-09  Juan Manuel Guerrero  <juan DOT guerrero AT gmx DOT de>


     * djgpp/src/libc/ansi/stdlib/strtod.c: Set tail pointer to string beginning if the string contains no valid numeric data.

     * djgpp/src/libc/ansi/stdlib/strtold.c: Set tail pointer to string beginning if the string contains no valid numeric data.

     * djgpp/src/libc/c99/stdlib/strtof.c: Set tail pointer to string beginning if the string contains no valid numeric data.

     * djgpp/tests/cygnus/conv_vec.c: Checks for hex string conversion added.

     * djgpp/tests/cygnus/convert.c: Check of strtodf replaced with strtof replaced.

     * djgpp/tests/cygnus/test.c: Support functions biggerf, mag_of_errorf and test_mokf added.

     * djgpp/tests/cygnus/test.h: Declaration for test_mokf and float_type and sprint_float_type structures added.


Patch omitted due to its size.