On 7/19/12 7:02 PM, Greg Banks wrote:
On Fri, Jul 20, 2012, at 12:13 AM, Дилян Палаузов wrote:
The non-portable trickery does not harm. It is supported by GCC and
Clang (according to
http://clang-developers.42468.n3.nabble.com/Does-clang-support-attribute-visibility-quot-default-quot-td3944043.html
).
There are still some other compilers left in the world. The Sun, sorry
Oracle, compiler, for example, will parse and ignore both
__attribute__(visibility()) in the code and -fvisibility=hidden on the
commandline, without failing.
Adding gcc-ism's without testing for gcc in configure is a terrible
idea. Please don't do it. The entire universe doesn't use gcc's
non-standard compiler gunk.
The semi-portable way to do this is with a linker map (the way ELF
versioned symbols are done), but even that is difficult to write in a
way that works on a majority of platforms without post-processing the
map file in configure to handle the syntax idiosyncrasies.
Isn't there enough _real_ work to do on the code and build process
without wasting time on this sort of thing?
--
Carson
