[tex-k] FAIL: tests/ttf.test

Jonathan Kew jfkthame at googlemail.com
Wed Nov 7 15:10:40 CET 2012

On 7/11/12 13:37, Norbert Preining wrote:
> Hi Jonathan,
> On Mi, 07 Nov 2012, Jonathan Kew wrote:
>> A workaround, in that case, might be to disable C++11 support with an
>> appropriate -std=... argument to the compiler.
> E.g.
> 	-std=c90
> would that suffice?

Perhaps; or maybe -std=c++98 would be more appropriate for C++ sources. 
I haven't looked into this in any detail, I'm just aware that a number 
of projects have recently reported issues that arise because ICU's UChar 
typedef changed from uint16_t to char16_t when it's built with a C++11 
compiler. This can cause problems for code that has its own Unicode 
character type based on unsigned short or uint16_t, and expects it to be 
compatible with ICU's UChar. The types themselves interoperate fine, but 
pointers to them are not considered compatible, AIUI.

Another solution (assuming this is the problem people are having) would 
be a sprinkling of explicit (UChar*) typecasts wherever such ICU APIs 
are used with external (unsigned short *) data types.


> Anyway, trying it now ... will report later
> But my first impression from
> 	export TL_COMPILER_GFLAGS="-g -std=c90"
> 	./Build
> does not show that it is actually used ... pain ..
> Best wishes
> Norbert
> ------------------------------------------------------------------------
> Norbert Preining            preining@{jaist.ac.jp, logic.at, debian.org}
> JAIST, Japan                                 TeX Live & Debian Developer
> DSA: 0x09C5B094   fp: 14DF 2E6C 0307 BE6D AD76  A9C0 D2BF 4AA3 09C5 B094
> ------------------------------------------------------------------------
> NEEN SOLLARS (pl.n.)
> Any ensemble of especially unflattering and particular garments worn
> by a woman which tell you that she is right at the forefront of
> fashion.
> 			--- Douglas Adams, The Meaning of Liff

More information about the tex-k mailing list