Dan Sugalski:
# It's impossible to have clone in a threaded environment guarantee that
it
# won't deadlock.
If I recall correctly, all shared variables are owned by a single shared
interpreter. What if, the first time the traversal encounters a shared
PMC, it gets a lock on the shared interpreter? If a second traversal in
another thread encountered another shared PMC, it would block until the
first traversal was finished. (And if there isn't a shared interpreter,
a global mutex would probably work just as well.)
Pseudocode:
bool seen_shared=false;
...
if(PMC_is_shared(pmc)) {
if(! seen_shared) {
Interp_lock(shared_interpreter);
seen_shared=true;
}
PMC_lock(pmc);
}
...
if(seen_shared) {
Interp_unlock(shared_interpreter);
}
--Brent Dax <[EMAIL PROTECTED]>
Perl and Parrot hacker
"Yeah, and my underwear is flame-retardant--that doesn't mean I'm gonna
set myself on fire to prove it."
