Re: [myhdl-list] List Of Constants

[Josy B. <jos...@gm...>]

> <snip>
> >
> > Any workaround is a good as any other workaround, but still is a
> > workaround. I just added the 4 (in my actual module, not the 
example)
> > constants in the call.
> 
> Workaround was a poor choice or words in this
> case, it is an alternative method.  Either
> way I don't think it was what you were
> looking for.
> 
> > None of the workarounds can be classified as _pythonic_, I'd say. 
What
> > if I come up with a 8 * 8 constant matrix (like in JPEG DCT)? And it 
is
> > well supported in VHDL ...
> 
> Either of those approaches, I would argue, are
> pythonic but they don't solve the problem you
> are trying to solve.
> 
> A 1D list of constants should be convertible a
> 2D I believe is not.  This (2D) is something that
> needs to be added.
> <snip>

Chris,

Replace my wording 'not pythonic' by 'less beautiful'. And 'pythonic' by 
'idiomatic Python' as coined by Jeff Knupp in his book (see  
http://www.jeffknupp.com)
You may want to take a look at the example: 
https://gist.github.com/josyb/af6b7809d976b2559ae4 anyhow.
The conversion passes if I send a list of Signals, but the VHDL signals 
are not initialised. My (final or main) question was whether I am wrong 
to expect to have those initialised in the VHDL code, as I have assigned 
in the Python code (and were used by the simulation). I can live with 
workaround but do believe in improving MyHDL. But I do understand that 
the current conversion process/code has its limits. Hopefully we can 
find a good candidate for the GSOC 'Refactor Conversion Code' project. 
To me the Python (annotated) AST to V* code in one step is perhaps 
better cut in two phases where we on a line-by-line basis introduce an 
intermediary representation which we finally translate into text output. 
The code in _toVHDL.py has seen a few flags introduced to steer the 
casting (I too did so to 'solve' issues #14 and #15). Maintenance just 
gets more and more difficult by each added flag?

A side-question, how do we label an issue on Github as Enhancement or 
Question? I drew up this question in the issue page, but didn't post it 
as I realised it might be interpreted as a 'bug report' which certainly 
is not my intention. 


A 1D list of constants only converts in  the 'special' case of inferring 
memory. As I explained to Henry (Gomersall) my code has a 'fixed' index 
and doesn't get recognised as such. The 2D array of constants was just 
an (unfortunate?) example to indicate how cumbersome and error-prone 
typing the 'alternatives' would become.

Best regards,

Josy