Re: About task-safeness

["Vinzent Hoefler" <0439279208b62c95f1880bf0f8776eeb@t-domaingrabbing.de>]

Maciej Sobczak wrote:

> On Feb 2, 10:14 pm, mockturtle <framefri...@gmail.com> wrote:
>
>> > Well, ARM 05, A(3/2) says:
>>
>> > |The implementation shall ensure that each language-defined subprogram is reentrant in
>> > |the sense that concurrent calls on the same subprogram perform as specified, so long
>> > |as all parameters that could be passed by reference denote nonoverlapping objects.
>>
>> > Apart from that, you probably have to trust the programmer - or some tool.
>>
>> Thank you, that is what I needed.
>
> Note that this is actually quite weak guarantee. It says that any call
> that refers to the same object need not be task-safe, even if a common
> sense would suggest otherwise. It was already discussed here, but two
> notable examples that are worth to keep in mind are:
>
> - two tasks printing something on stdout are *not* safe
>
> - two tasks "only reading" values from the same container are *not*
> safe

Of course not, because if you think a bit harder about it, the result is that
both calls use a "parameter that could be passed by reference" denoting the
same "non-overlapping object".


Vinzent.

-- 
You know, we're sitting on four million pounds of fuel, one nuclear weapon,
and a thing that has 270,000 moving parts built by the lowest bidder.
Makes you feel good, doesn't it?
   --  Rockhound, "Armageddon"