Georg Baum a écrit :
Abdelrazak Younes wrote:
Georg Baum a écrit :
Jean-Marcs concern is still valid: If the include is not needed, remove
it. Simply replacing it by something else that is not needed either does
not make sense. And if the include is needed leave it like it is. The
c-prefixed variants are not available everywhere.
You guys thinks that LyX is still portable to some old Unix, I think you
should prove your affirmation.
No. You want to remove stuff, so you have to prove that it will not break
anything.
I think that every single macro should have a justification in comments.
Keeping macros because they might solve a problem somewhere is not a
good answer.
I would say that LyX is not portable to a
platform that don't support advanced C++.
That is true, but keep in mind that not every platform uses gcc, IIRC the
sun compiler was pretty early in C++ adoption, and there are several
others. And even if old systems are ruled out: *BSD is sometimes quite
different from linux or windows.
Then, which Macro is there to support *BSD, or Sun compiler, or... This
should be investigated and the only way to do that is actually to remove
the macro and see on which platform the compilation break. Then you will
know and you can put a comment justifying the macro.
I would go as far as to say
that LyX portability is set by the portability of gcc >= 3.3 and boost.
The minimum gcc version is 3.1 (at least it was like that two months ago, I
have that compiler at home and can check that), and boost supports an
astonishing large number of old compilers including VC++ 6 and gcc 2.95.
OK but that might change with exception and advanced template use which
as for as I remember are fixed in 3.3.
Abdel.
