#218: can't sort a PIR subclass of an RPA.
-------------------------+--------------------------------------------------
Reporter: coke | Owner: whiteknight
Type: bug | Status: reopened
Priority: normal | Milestone: 1.1
Component: core | Version:
Severity: high | Resolution:
Keywords: tcl blocker | Lang:
Patch: | Platform: all
-------------------------+--------------------------------------------------
Comment(by whiteknight):
Okay, I've looked back into a few of these issues, and I think that the
tests added in r38234 are not good tests.
ResizablePMCArray and FixedPMCArray do not supply get_pointer VTABLEs, and
fall back to Default.get_pointer(). This, of course, simply returns SELF
which is the behavior that these tests are probing for. However, returning
a pointer to the PMC itself is far from the most common use of
VTABLE_get_pointer. Hash PMCs for instance use get_pointer to return a
pointer to the underlying Hash* structure. Sub PMCs use get_pointer to
return the opcode_t* address of the PIR subroutine. So, it doesn't make
sense to force the *Array PMC get_pointer VTABLEs to return the address of
the PMC itself and not a pointer to it's underlying data structure, when
other PMCs aren't held to this rule (and since it makes more sense to use
get_pointer here in a different way).
Now, that brings me to the issue of the get_addr opcode, which has
problems. get_addr is used in two distinct ways from PIR (it is used in a
variety of other ways internally):
1) To retrieve opcode_t* addresses in things like Sub and Continuation
PMCs
2) To retrieve the actual location in memory of array PMCs (which is how
pmichaud's tests use it)
I would suggest that use #2 is not the right use of get_addr, and if
people are looking to get the actual memory location of a PMC in a
consistent way a new opcode for that purpose should be created.
I suggest the creation of a get_memaddr opcode to return the actual
location in memory of the PMC structure, and let get_addr be a wrapper
around the get_pointer VTABLE (with the knowledge that it will return
inconsistent results based on PMC type). Thoughts?
--
Ticket URL: <https://trac.parrot.org/parrot/ticket/218#comment:15>
Parrot <https://trac.parrot.org/parrot/>
Parrot Development
_______________________________________________
parrot-tickets mailing list
[email protected]
http://lists.parrot.org/mailman/listinfo/parrot-tickets
