[libre-riscv-dev] [Bug 316] bperm TODO

[bugzilla-daemon at libre-soc.org]

https://bugs.libre-soc.org/show_bug.cgi?id=316

--- Comment #11 from Luke Kenneth Casson Leighton <lkcl at lkcl.net> ---
(In reply to Cole Poirier from comment #10)
> (In reply to Luke Kenneth Casson Leighton from comment #9)
> > (In reply to Cole Poirier from comment #8)
> > added you.
> 
> Awesome! It worked :)

exxxcellent. muhahaha.

> 
> > think in terms of multi-way collaboration: you need to do work, other people
> > need to do work, the code under development needs to work.
> > 
> > the guidelines (based on 20+ years of experience of working with software
> > libre) balance all those three things.
> 
> I'm with you fully Luke. I try to follow your guidelines as best as I can.

i am inviting you to think for yourself, to derive them from first principles
and thus be able, if there are any... unusual circumstances, to not be either
frozen solid or disruptive.


> > and if you *really* want to put your computer into meltdown, run "synth"
> > followed by "show top".
> 
> I just might try it.

be prepared to have to run "killall xdot".

>  
> > just from a few lines of python, eh?  shows how careful you have to be
> > when working in H(igh) D(efinition) L(anguage)s.
> 
> It really is incredible. Looking forward to improving the level of care I am
> capable of taking with regard to HDL.

if you at least run "show top" and also inspect submodules visually ("show
presstabkey presstabkey") and if you can relate what you see back to the actual
code and still follow the graph, the design is "clean".

if however you cannot see the graph even at fullscreen because boxes are too
small even on a 1920x1080 screen, then something is wrong, you need to break it
down into smaller submodules.

if possible.

or in this particular case, redesign it.

but not right now.

write unit test first.

research examples on how to do that ok?

-- 
You are receiving this mail because:
You are on the CC list for the bug.