On 2012.06.27 19:39, Ludovic Rousseau wrote:
> Maybe the solution is to use -std=c99 instead.

If we use c99, then we have to revert to setting the -std option for 
each individual toolchain, as we cannot use c99 on cygwin and MinGW for 
some time. We used to do that at some stage, but we moved away from it.

> I could also rebuild on Mac without any -std=..
> So maybe the best solution is to completely remove -std=

I don't think we can do that either, since we do need GNU99 extensions 
for problematic toolchains, and I wouldn't vouch for defaults without 
time consuming testing. The issue was fixed not so long ago, so I don't 
see it acceptable to risk people having to figure out the issue and 
upgrade their toolchain, when we can avoid wasting their time by keeping 
GNU99 for the time being.

> I could not find when and why -std=c99 was added.

It was added because we need to compile with MSVC, even if MVSC is far 
from implementing C99. The logic is that if we can steer away from GNU, 
we can avoid someone modifying core and using GNU extensions that are 
gonna break MSVC. The other option was to go with C89, since it is the 
actual lowest common denominator, but this was not deemed acceptable.

> libusb also has this compilation option.

What we have here comes from libusb. We haven't changed anything so far 
with regards to these options.

My take on this is that unless someone submits a patch, I'm not going to 
do anything. As long as Clang doesn't break, it is benign.

Regard,

/Pete

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
libusbx-devel mailing list
libusbx-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/libusbx-devel

Reply via email to