Comments vs. commit log

Comments vs. commit log

[Ludovic Courtès]

Hi Andy!

Just a meta-comment before I actually look into the code and fully
appreciate your work.  :-)

"Andy Wingo" <wingo@pobox.com> skribis:

> commit 02ebea537fa805c615df44c4228db6a44d74c4b3
> Author: Andy Wingo <wingo@pobox.com>
> Date:   Fri Sep 23 12:43:04 2011 +0200
>
>     peval: simpler and more precise treatment of mutability
>     
>     * module/language/tree-il/optimize.scm (peval): The old approach of
>       optimistically producing constants and then de-constifying them at
>       their uses was not only cumbersome but incorrect: it both failed to
>       preserve identity in some cases and failed to retain immutable
>       constant values.  Instead, now we only produce constants if they
>       really are constant and immutable.  The constant folder has to have a
>       few more algebraic cases to be as effective as it was, to destructure
>       (car (cons _ _)) appropriately.  On the plus side, now constructors
>       and deconstructors can handle impure cases more generally.

[...]

> +(define* (build-var-table exp #:optional (table vlist-null))
> +  (tree-il-fold
> +   (lambda (exp res)
> +     (match exp
> +       (($ <lexical-ref> src name gensym)
> +        (let ((var (vhash-assq gensym res)))
> +          (if var
> +              (begin
> +                (set-var-refcount! (cdr var) (1+ (var-refcount (cdr var))))
> +                res)
> +              (vhash-consq gensym (make-var name gensym 1 #f) res))))
> +       (_ res)))
> +   (lambda (exp res)
> +     (match exp
> +       (($ <lexical-set> src name gensym exp)
> +        (let ((var (vhash-assq gensym res)))
> +          (if var
> +              (begin
> +                (set-var-set?! (cdr var) #t)
> +                res)
> +              (vhash-consq gensym (make-var name gensym 0 #t) res))))
> +       (_ res)))
> +   (lambda (exp res) res)
> +   table exp))
> +
> +(define-record-type <counter>
> +  (%make-counter effort size continuation recursive? data prev)
> +  counter?
> +  (effort effort-counter)
> +  (size size-counter)
> +  (continuation counter-continuation)
> +  (recursive? counter-recursive?)
> +  (data counter-data)
> +  (prev counter-prev))

Would it be possible to move the explanations from the commit log to
comments in the code?

Looking at the above code, I have no idea what a var-table or a counter
is, what it’s used for, etc.

Thanks,
Ludo’.

Re: Comments vs. commit log

[Andy Wingo]

On Tue 27 Sep 2011 22:43, ludo@gnu.org (Ludovic Courtès) writes:

> Would it be possible to move the explanations from the commit log to
> comments in the code?
>
> Looking at the above code, I have no idea what a var-table or a counter
> is, what it’s used for, etc.

Sure.  I forget sometimes, because I've read that Waddell & Dybvig paper
so many times that I forget it's not written down in the source ;)

Andy
-- 
http://wingolog.org/

Re: Comments vs. commit log

[Ludovic Courtès]

Hi!

Andy Wingo <wingo@pobox.com> skribis:

> On Tue 27 Sep 2011 22:43, ludo@gnu.org (Ludovic Courtès) writes:
>
>> Would it be possible to move the explanations from the commit log to
>> comments in the code?
>>
>> Looking at the above code, I have no idea what a var-table or a counter
>> is, what it’s used for, etc.
>
> Sure.  I forget sometimes, because I've read that Waddell & Dybvig paper
> so many times that I forget it's not written down in the source ;)

I wonder what the GCC hacker community would be like if its source code
wasn’t commented.  :-)

Seriously though, I think that pointing to the paper would be nice, but
the paper is higher-level than the actual implementation, so comments
could fill in the gap.

Thanks,
Ludo’.