Re: [Qemu-devel] [PATCH] curses: build with -std=gnu99

[Markus Armbruster]

Samuel Thibault <address@hidden> writes:

> Markus Armbruster, on Fri 28 Oct 2016 08:51:20 +0200, wrote:
>> Samuel Thibault <address@hidden> writes:
>> 
>> > Peter Maydell, on Thu 27 Oct 2016 17:52:17 +0100, wrote:
>> >> http://www.thecodingforums.com/threads/wchar_t-is-useless.806149/#post-4398211
>> >
>> > UURrgll... So we can't use L'\u23bd' on such systems, it would just not
>> > work either, we have to use iconv to get these right...
>> 
>> I guess we can if they actually bother to conform to C99.  6.4.3
>> Universal character names:
>> 
>>     The universal character name \Unnnnnnnn designates the character
>>     whose eight-digit short identifier (as specified by ISO/IEC 10646)
>>     is nnnnnnnn.  Similarly, the universal character name \unnnn
>>     designates the character whose four-digit short identifier is nnnn
>>     (and whose eight-digit short identifier is 0000nnnn).
>> 
>> Note that it doesn't say here that L'\u23bd' should be equal to 0x23bd.
>
> I know, but as discussed in the thread, that could only work if gcc
> embeds something like calls to nl_langinfo and iconv to properly convert
> from the unicode value to the proper wchar_t according to the locale.  I
> wouldn't be surprised that they don't.

Before we uglify our code to work around (egregious!) defects in certain
compilers, we should find out whether and where exactly these defects
exist, and whether it's worth targeting the defective compilers.