[mercury-users] map and c_pointer

[Tyson Dowd]

On 11-Apr-2002, Simon Taylor <stayl at cs.mu.OZ.AU> wrote:
> On 11-Apr-2002, Michael Day <mikeday at bigpond.net.au> wrote:
> > If I have a map(string, foo), where foo ---> foo(c_pointer), should I
> > expect a segfault when I attempt to insert a value into the map?
> 
> Why? Unification and comparison of c_pointers is handled by address
> comparison, so map(foo, string) would also work. The Mercury code
> won't even care whether the pointers are valid. You have to really
> be trying to get a segfault in Mercury.

Just as a side note, address comparison doesn't work for backends with
garbage collectors which move pointers.  In general it is not a good idea
to do things this way -- you should use a stable name for this kind
of code.  We don't have a general stable name system yet.

See:

@inproceedings{ jones99stretching,
    author = "Simon L. Peyton Jones and Simon Marlow and Conal Elliott",
    title = "Stretching the Storage Manager: Weak Pointers and Stable Names in Haskell",
    booktitle = "Implementation of Functional Languages",
    pages = "37-58",
    year = "1999",
    url = "citeseer.nj.nec.com/peytonjones99stretching.html"
}

for how this can be done and why.

There was code that did this on startup of the Mercury runtime system
(sorted a map on c_pointers) which causes quite a lot of trouble with both
the accurate GC and .NET backend.


Any code that interfaces to C and passes C pointers around is not pure
Mercury code anyway, so you don't have to try that *hard* to get a
segfault.

-- 
       Tyson Dowd           # 
                            #  Surreal humour isn't everyone's cup of fur.
     trd at cs.mu.oz.au        # 
http://www.cs.mu.oz.au/~trd #
--------------------------------------------------------------------------
mercury-users mailing list
post:  mercury-users at cs.mu.oz.au
administrative address: owner-mercury-users at cs.mu.oz.au
unsubscribe: Address: mercury-users-request at cs.mu.oz.au Message: unsubscribe
subscribe:   Address: mercury-users-request at cs.mu.oz.au Message: subscribe
--------------------------------------------------------------------------