Hi Paul, > > Why? Around 1992-1997 I learned the lesson that one cannot really rely > > on the values of __STDC_VERSION__ and the like. Too many compilers lie > > about it. > > Although that's certainly true in general
OK, thanks for the confirmation. > It has been used that > way for something similar in Gnulib's lib/vla.h, without reported > problems. Well, lib/vla.h is more about function declarations, rather than about VLAs inside functions. What would be useful is to have the test results of m4/vararrays.m4 with various compilers. More precisely, with the _minimum_ version of the Solaris C compiler that advertises __STDC_VERSION__ == 201112L (because usually they don't regress on language features, but they often advertise a full standard as being supported although some language features are not yet implemented). And likewise for HP-UX C, MSVC, Intel C, and so on. But I don't want to spent 3 hours on this. The benefit is too small. Bruno
