[PicForth] Re: Feature Requests

[Samuel Tardieu]

>>>>> "Alex" == Alex Holden <alex at linuxhacker.org> writes:

Alex> 1. It would be great if any words which are defined but never
Alex> used weren't compiled into the binary (I realise this would
Alex> require two passes to figure out how many times each word is
Alex> used). In a similar vein, it would be nice to use the same
Alex> information to automatically inline any words that are only used
Alex> once.

Mmm, I really don't want to rearchitecture everything in PicForth to
do that. However, maybe someone could implement that in the statistics
you suggest below and report words used only once and containing a
simple return statement at the end as well as unused words.

Alex> 2. It would be extremely helpful for beginners if the
Alex> documentation included a glossary of every picforth word with a
Alex> brief description of what it does.

Be my guest :-) I can help, but I don't have the resources to do
that.

Alex> 3. The picforth source tree would be a little easier to navigate
Alex> if the example source files and the library files were both in
Alex> subdirectories of their own.

You may be right.

Alex> 4. It would be handy if there was some way for the 'dis' and
Alex> 'map' commands to output directly to a file. At the moment I run
Alex> picforth twice every time I compile- the first run dumps the hex
Alex> file and pipes the map through sort to a file and the second run
Alex> pips the disassembly to a file.

Right again.

Alex> 5. It would be nice if there was some kind of 'statistics'
Alex> command which after a compilation run tells us stuff like how
Alex> many instructions and how many RAM bytes were used, how many
Alex> words there were, the maximum return stack depth, etc.

Be my guest again :)

  Sam
-- 
Samuel Tardieu -- sam at rfc1149.net -- http://www.rfc1149.net/sam