X-Authentication-Warning: delorie.com: mail set sender to djgpp-workers-bounces using -f
X-Recipient: djgpp-workers AT delorie DOT com
Message-ID: <504F6729.5000505@iki.fi>
Date: Tue, 11 Sep 2012 19:30:33 +0300
From: Andris Pavenis <andris DOT pavenis AT iki DOT fi>
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:15.0) Gecko/20120828 Thunderbird/15.0
MIME-Version: 1.0
To: djgpp-workers AT delorie DOT com
Subject: Re: Invalid tests for __STDC_VERSION__ in DJGPP header files
References: <201209111230 DOT q8BCU0rl001296 AT dexter DOT ludd DOT ltu DOT se> <504F3374 DOT 3080400 AT iki DOT fi>
In-Reply-To: <504F3374.3080400@iki.fi>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 8bit
X-MIME-Autoconverted: from Quoted-Printable to 8bit by delorie.com id q8BGUfS3009476
Reply-To: djgpp-workers AT delorie DOT com

On 09/11/2012 03:49 PM, Andris Pavenis wrote:
> On 09/11/2012 03:30 PM, Martin Strömberg wrote:
>> According to Andris Pavenis:
>>> -_STDC_VERSION__ is not defined in many cases any recent GCC version. It is only seems to be
>>> defined in several cases depending on the command line parameter -std and for C language only
>> ...
>>> As the result tests used in header files like
>>>
>>> #if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) \
>>>     || !defined(__STRICT_ANSI__)
>>>
>>> are invalid.
>> Those tests aren't invalid. They are checking for __STDC_VERSION__ >=
>> 199901L and not __STRICT_ANSI__.
>>
>> It's possible that old gcc was lying and claiming conformance where it
>> wasn't.
>>
>> ...
>>
>>> Use of the broken test for __STDC_VERSION__ prevents libstdc++-v3 (GCC-4.8.0 development versions
>>> from some last weeks) building OK due to actual contents of stdint.h being compiled out. Variant 1
>> That seems to say that libstdc++-v3 requires standard comliance in the
>> compiler. How about adding -std=c89/c99/whatever to libstdc++-v3's
>> compile options?
>>
>>
>> On the other hand I'm not sure I understand the problem properly so
>> perhaps I'm totally wrong.
>>
>>
> According to C standard drafts I have predefined macro __STDC_VERSION__ seems to mandatory.
> It is however not defined in many cases by GCC as far as I have tested. I do not know what would
> be best way how to handle that for C
>
> At least for C++ in last standard draft I have (I have not bought released version document) in
> 16.8 is written
>
> 2 The following macro names are conditionally defined by the implementation:
> _ _ STDC _ _
> Whether _ _ STDC _ _ is predefined and if so, what its value is, are implementation-defined.
> _ _ STDC_MB_MIGHT_NEQ_WC _ _
> The integer constant 1, intended to indicate that, in the encoding for wchar_t, a member of the basic
> character set need not have a code value equal to its value when used as the lone character in an
> ordinary character literal.
> _ _ STDC_VERSION _ _
> Whether _ _ STDC_VERSION _ _ is predefined and if so, what its value is, are implementation-defined.
> _ _ STDC_ISO_10646 _ _
> An integer constant of the form yyyymmL (for example, 199712L). If this symbol is defined, then every
> character in the Unicode required set, when stored in an object of type wchar_t, has the same value
> as the short identifier of that character. The Unicode required set consists of all the characters
> that
> are defined by ISO/IEC 10646, along with all amendments and technical corrigenda as of the specified
> year and month.
>
> That means that __STDC_VERSION__ is not required to be predefined for C++ (if the draft is OK).
> As the result I think we should include  '|| defined(__cplusplus)' in the condition
>
> I'll check whether it fixes libstdc++ build problem in gcc-4.8.0 development versions later
>

Tried

--- stdint.h.orig       2012-09-09 19:12:59.000000000 +0300
+++ stdint.h    2012-09-11 19:19:56.936112793 +0300
@@ -9,7 +9,7 @@
  #endif

  #if (defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L) \
-  || !defined(__STRICT_ANSI__)
+  || !defined(__STRICT_ANSI__) || defined(__cplusplus)

  typedef signed char int_least8_t;
  typedef unsigned char uint_least8_t;

Fixes libstdc++ build problems for gcc-4.8.0 20120909 (again cross-native build as
after removing libstdc++-v3 build directory 'make -j12' on Intel 3930K rebuilds it
under Linux in less than 30 seconds)

The same should be done for other header files which has similar checks.
Does it look OK?

Andris