On Thu, Nov 08, 2007 at 03:40:11PM +0100, Abdelrazak Younes wrote: > Enrico Forestieri wrote: > > On Thu, Nov 08, 2007 at 02:45:36PM +0100, Enrico Forestieri wrote: > >> On Thu, Nov 08, 2007 at 02:25:29PM +0100, Abdelrazak Younes wrote: > >>> Enrico Forestieri wrote: > >> [...] > >>>> I don't have gcc 3.3 and could not test it. I'd rather not change it > >>>> unless someone reports problems. > >>> So I am wrong and USE_WCHAR_T is correctly detected and used. Then why > >>> did you put this in docstream.cpp? > >> That's Georg's doing, not mine. He is more scrupulous than me, seemingly, > >> and probably actually tested it with gcc 3.3. > > > > Sorry, I think I misread your question (I still had in mind your > > docstring.h observation). I think you meant asking why I put the > > insertion operator for chars in docstream.cpp. I had to put it > > somewhere and the natural place was there. > > OK, but this means that it is not inlined anymore. I guess this is not a > performance critical method anyway, is it?
Using forward declarations you cannot have it inlined without an actual definition, and this would defeat the purpose. As it is not critical, I don't think there's a problem here. -- Enrico
