Re: [ECOS] suspend/resume nesting

[Eric Donnat <e dot donnat at silicomp dot fr>]

On Wed, 21 May 2003 13:15:18 +0200
"Koeller, T." <Thomas.Koeller@baslerweb.com> wrote:

> Eric,
> 
> my intention was to make a proposal for an improvement
> in a particular field. At least I see it as an
> improvement, others may of course disagree (that's why
> I posted it here).
>

Agreed.
 
> I do not think that 'don't do that' is a valid response:
> this implies subdividing the API into a set of functions
> that may be used and others that may not. To a novice user
> this could be a bit confusing. If the user is not supposed
> to call a particular function, then the API must not expose
> it to him.

It's a bit of un-understanding. I was just arguing that for the
same purpose of the presented sample you were giving,
almost two or three other ways of using other APIs, (the
one that are eventually better documented) can succeed to
solve the sample.

I nicely wanted to show you that you were surely reinventing
the wheel.
 
> The scheme I described and used in my program is perfectly
> valid, it does not use anything that is not in the official
> API, and so it should just work. You are vigorously defending
> your point, but I still cannot see what benefits I get by
> disallowing negative suspend counts.

Of course, everything is valid. Even IT off - DO nothing is valid,
if this is just what you need. There are no benefits against
positive or negative suspend counts, until the day someone
would really need them.



-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss