Re: percent codes for analyses

Geoffrey.Coram wrote:
> Kevin Cameron wrote:
>   
>> Actually there are lots more than 26 since you could use uppercase and
>> they don't have to be single character or just a-z.
>>     
>
> Tradition is that case is not considered; see Table 10-4 in 
> the AMS LRM 2.2 (%h or %H, %d or %D, ... %s or %S).  The only
> differences are %e/%E which determine whether it's 1e-3 or 1E-3.
>   
I think the %e/%E case sets the precedent that you can have the case 
interpreted differently on new codes.
> Does Verilog allow multiple-character percent codes?  C has
> some (%ld for long int), but I don't think I've seen them in 1364.
>   
No reason you can't do them going forward. If a simulator doesn't 
understand a new code (e.g. %a) then it's unlikely it's going to 
understand the rest of the context, but I cant think of any useful 
multi-character sequences so maybe that's a moot point anyway.

Given that there aren't many codes left I'd vote for making new ones 
case sensitive, alternatively you could use "%#" for the run number.

Kev.
>   
>> Maybe we could add (say) %A to indicated the run number e.g.
>> "results_%a.%A" would be "
>>
>>   "results_dc.0","results_dc.1",...
>>     
>
> It seems reasonable to use %A and %a to be some sort of
> variants on the analysis name, since the basic data is
> the same (a la %e/%E: same number, different format).
>
> -Geoffrey
>
>   

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.