Hi, For the time being, I use aliases in encapsulate function instead: https://bitbucket.org/hiyuh/kcpsm6-clone/changeset/3493ed99c60c
Thanks! 2011/6/13 Alex Huntley <[email protected]>: > The don't care value '-' in std_logic is an actual value - i.e. it doesn't > cause any value to match. > > As an example: > > case test is > when "111-" => > output <= '1'; > when others => > output => '0'; > end case; > > When examining the first clause "111-" a literal match is required - i.e. > "1110", "111Z", "111H" etc will NOT match and will not cause output to be > set. The only thing that will cause a match here is if the value really is > "111-". > > This means that you would effectively have to write out the case statement > long-hand, e.g. > > case test is > when "1111" => > output <= '1'; > when "1110" => > output <= '1'; > when "111H" => > output <= '1'; > when "111L" => > output <= '1'; > when "111Z" => > output <= '1'; > > -- etc > > when others => > output => '0'; > end case; > > VHDL 2008 introduced an alternative to this obvious pain in the backside. If > you write the "case" with a question mark then it will cause the '-' to be > evaluated as don't care rather than the actual value: > > ?case test is > when "111-" => > output <= '1'; > when others => > output => '0'; > end case; > > In this example (assuming that you've compiled the code as VHDL 2008 and > your compiler actually supports this) then "1110", "1111", "111H" etc will > all cause the output to be set. > > BUT, hopefully someone will correct me if I'm wrong but I don't think that > GHDL supports VHDL 2008 yet. Tristan obviously has a lot of work to do and > what he has produced to date is very admirable and generous, so please don't > read this as any criticism. I know ActiveHDL from Aldec only supports a few > features of VHDL 2008; I don't know about Modelsim but I wouldn't want to > touch that with a barge pole anyway. > > I guess that ties in with what you are seeing. Check the Doulos site for > correct VHDL 2008 syntax and hopefully someone will correct me if I'm out of > date on what GHDL can and can't do. > > As ever thanks to Tristan for his sterling efforts with GHDL. I think it > compares very favourably against the expensive proprietary tools and usually > sticks closer than any to the LRM. _______________________________________________ Ghdl-discuss mailing list [email protected] https://mail.gna.org/listinfo/ghdl-discuss
