On Sat, Aug 13, 2011 at 3:37 PM, Nicholas Clark <nick@ccl4.org> wrote: > ..\doio.c(1773) : warning C4244: '=' : conversion from 'IV' to 'I32', possible loss of data > I'm suspicious about all the conversion warnings. I'm assuming that UV and IV > are 64 bit types, and __int64 will definitely be a 64 bit type. Surely they > shouldn't be there if everything is working correctly? Aren't those warnings about truncations that will happen on any platform where IV and UV are 64 bits? For example, that last one in doio.c is about: while (++mark <= sp) { I32 proc; SvGETMAGIC(*mark); if (!(SvIOK(*mark) || SvNOK(*mark) || looks_like_number(*mark))) Perl_croak(aTHX_ "Can't kill a non-numeric process ID"); proc = SvIV_nomg(*mark); APPLY_TAINT_PROPER(); if (PerlProc_kill(proc, val)) tot--; } Since proc is an I32 and SvIV_nomg returns an IV (which may be either 32 or 64 bits), the warning seems to me a fair complaint. But not likely to be part of the build problems reported here.Thread Previous