Guilherme Jorge wrote: > I just got this message error while compiling float_pkg compatible > version in Quartus II 5.0 > > Error: VHDL syntax error at float_pkg_c.vhd(1026) near text > "function"; expecting a sequenial statement, > Error: VHDL error at float_pkg_c.vhd(1026): type of identifier > "test_boundary" does not agree with its usage as void type > > And so on... ;P > > Fixed_pkg works just fine... im compiling them as library work. The problem here is the "test_boundary" funciton. In this function I check to see the REAL number given is valid for the floating point format selected. To do this, I need to use the math_real function "**". Altera just doesn't like this funciton. 1) Call this into Altera as a bug. 2) Work around (pretending that denormal, and infinite numbers don't exist): function test_boundary ( arg : REAL; -- Input, converted to real constant fraction_width : NATURAL; -- length of FP output fraction constant exponent_width : NATURAL; -- length of FP exponent constant denormalize : BOOLEAN := true) -- Use IEEE extended FP return boundary_type is begin -- function test_boundary if arg = 0.0 then return zero; else return normal; end if; end function test_boundary; -- David W. Bishop dbishop@vhdl.org All standard disclaimers apply.Received on Wed Aug 24 06:41:35 2005
This archive was generated by hypermail 2.1.8 : Wed Aug 24 2005 - 06:42:29 PDT