Hello Brian, Tristan, All,

< Supporting block comments for example would be easy. But IMO block
comments are broken; the standard copied the C way when a much more
robust, less bug-prone style, well proven in Modula-2, would have been
trivial. I wonder if it's worth raising this on P.1076.

Can you pinpoint a specific set of VHDL-2008 features you need?
/>

I feel it is worth raising this on P1076 if there is a much better way to
do block commenting, however there will most probably be some resistance
(e.g. from EDA vendors) which we need to wade through. It is good if the
solution we propose is backward-compatible with the VHDL-2008's existing
solution. I admit I haven't taken time to look into Modula-2's block
commenting, so I won't be able to comment on how feasible it is for a
change to be implemented at this time.

I try as much as possible to use many of VHDL-2008's new features, most of
them are summarised in Jim's paper:
http://www.synthworks.com/papers/VHDL_2008_end_of_verbosity_2013.pdf
Among the features taken from Jim's list, here's a list of the ones I more
frequently use (plus 1 or 2 additional):
1. Simplified process sensitivity list.
2. Simplified conditionals.
3. Reading out ports.
4. Enhanced bit string literals.
5. Expressions in port maps.
6. Block commenting.
7. Matching relational operators, matching case statement.

Others that I also frequently use are some of the major enhancements made
in VHDL-2008, such as type generics, subprogram generics, and package
generics.


Hello Tristan,
< as GHW is a binary format, there is not grammar for it.  You have to
look at the sources to understand the format.  I am not sure that
a grammar is the best way to describe this format (for the same
reason, I suppose no one has written a grammar for mp3).
/>

I have just begun to sift through some of GHDL's sources. I believe
ghwlib.c is the place to start?
Perhaps you are right that since GHW is a binary format, describing it in
(E)BNF would not be possible. However, do you think you could come up with
something like a frame format describing where the data boundaries are
(e.g. preamble, header, payload, footer, or something of this sort)?

< GHW supports all the vhdl types, including floating point types.
/>

This is what I like about GHW, since the format is VHDL-friendly.


Hello Nikolaos, Adam,
<
> It would be nice to maybe extend FST or GHW (which are more
> VHDL-oriented) with real data types (along either additions
> like enumerations). A proper solution would involve the entire
> VHDL type system.
>

For the last couple of months I've been toying with the notion
of an [HDF5][1] based signal file for GHDL.
/>

I'm not familiar with the FST and HDF formats. Who / which tool uses these
currently? A quick check on Google regarding FST did not seem to return
meaningful results.

regards, daniel
_______________________________________________
Ghdl-discuss mailing list
[email protected]
https://mail.gna.org/listinfo/ghdl-discuss

Reply via email to