On 24/11/14 11:21, Gisle Vanem wrote: > Pádraig Brady wrote: > >>> I assume if some vendor have 'cosf' it should be undefined >>> before Gnulib is trying to override or replace it. Not vice >>> versa. So shouldn't this be: >>> # if @HAVE_COSF@ || @REPLACE_COSF@ >>> # undef cosf >> >> Well all the ...f() variants seem to be treated like this: >> acosf() coshf() asinf() atanf() atan2f() ceilf() expf() ... >> >> In the m4, the corresponding check is with: AC_CHECK_FUNCS([cosf]) >> I.E. the linker is used to find the function in libm. >> If it's only a macro and not in libm, then the code in math.h.in >> above makes sense. >> >> So are you saying that HAVE_COSF is 1 for you (i.e. it's in your libm), >> and you're also replacing it? > > Thank for your reply. Sorry for the late reply. > > I'm not sure if I'm replacing it or not. Since the .in.h-files seems > so contorted, it's not easy to see what the rationale is behind > those #ifdef. It seems to me Gnulib has an option to replace most > CRT functions with an 'rpl_*' version. But none of these math- > functions, right?
Well the ones being considered here are provisions for missing functions rather than replacements of problematic ones. > If so, I guess ./lib/test-cosf.c wont build unless cosf() is provided > by Gnulib. Right? The tests should work independent of replacement/provision. > Another thing that puzzles me with Gnulib's math-functions for MSVC. > MSVC v18 (VS Express 2013, Vista+) has more math functions than MSVC v16. > For example "exp2()". So if a Gnulib imp-lib built with MSVC v18 is linked > to a program an run on Win-XP, the program will not run; missing exp2() > in Win-XP's MSVCRT.dll. And vice-versa; possibly multiple symbols during link. > Does this mean a Gnulib DLL must be rebuilt for each Windows/MSVC version? I've no experience of windows TBH, but gnulib is primarily designed as a source inclusion mechanism for projects, rather than as a standalone library for distribution. thanks, Pádraig.
