This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Proto-patch for "%I64" on Microsoft

On Fri, 26 Nov 2004, Mark Mitchell wrote:

> I don't know.  Windows does not support "%ll" at all, so I'm not quite clear
> on what you're asking.  Windows has no equivalent of /usr/include, so there
> are probably multiple inttypes.h implementations for Windows. There's the one
> provided by MinGW, at least.  Microsoft Visual Studio may provide one too, but
> I'm not sure.  The MinGW version does not use "%ll", but does use "%I64".

My point is that there are two separate questions: do we support user code 
using "%I64" directly, and do we support user code using inttypes.h?  If 
inttypes.h uses %I64 and Windows doesn't support %ll at all, there's no 
choice of doing just one, but it would be possible for some system to 
support both %ll and %I64, and use %I64 in inttypes.h, in which case on 
that system we'd have the opportunity to fix inttypes.h to work with 
-std=c99 -pedantic -Wformat.

Actually fixing/replacing printf to support %ll is clearly too far to go 
in fixing a defective C library, though we do provide the mem* functions 
in libgcc on a few platforms that formerly didn't define 
TARGET_MEM_FUNCTIONS and there's been discussion of doing much the same 
with the float versions of <math.h> functions to avoid libstdc++-v3 and 
libgfortran both needing to work around pre-C99 libraries which lack those 

Joseph S. Myers      (personal mail) (CodeSourcery mail) (Bugzilla assignments and CCs)

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]