Re: Fix build failure on GNU Hurd (issue 319400043 by address@hidden)

[David Kastrup]

Benkő Pál <address@hidden> writes:

> 2017-03-04 19:03 GMT+01:00  <address@hidden>:
>> On 2017/02/25 15:19:47, felix.janda_posteo.de wrote:
>>>
>>> The fix probably breaks compilation on everything not using glibc:
>
> Is lilypond supposed to be built on anything outside GNU/Linux?

More or less yes.  Heeding reports and warnings outside of GNU/Linux
make us more robust against relying on features and quirks specific to
current versions of GCC (which has a tendency to retire everything not
making it into C++ standards) and possibly certain platforms.

It's not a priority, but we tend to try keeping up with the other world.
Recently there were reports about overloading problems with Clang++ that
I decided _not_ working around since

a) it was a clear compiler error of Clang++ in heeding the standard, so
there was reasonable expectation that Clang++ would eventually get
around to fixing it.

b) workarounds would have definitely impacted the maintainability of the
code base.

But in general, we do try to "play nice" where it doesn't cost us
significantly.

-- 
David Kastrup