Great! Let me know if I can help.

--Christopher

On Tue, Mar 29, 2016 at 4:34 PM, Ben Albrecht <[email protected]> wrote:
> Hi Christopher,
>
> This is a pretty interesting case, and I'm surprised it is not already
> tracked in our test suite. We are (now) aware of what it causing the error
> and are working towards a solution. A future should be submitted within a
> fews days, unless the bug is fixed in master by then, in which it will be
> filed as a test instead. The delay is due to the fact that we're just about
> to publish a new Chapel release, version 1.13.
>
> Thanks for the bug report!
>
> Best,
> Ben
>
> From: Christopher Swenson <[email protected]>
> Date: Monday, March 28, 2016 at 2:51 PM
> To: "[email protected]" <[email protected]>
> Subject: [Chapel-bugs] Tuple destrcutor compiler crash
>
> I ran into an internal error.
>
> Minimal reproducer is
>
> proc f() : (int, int) {
>   return f();
> }
> proc main() {
>   f();
> }
>
> output:
> repro.chpl:2: internal error: number of actuals does not match number of
> formals [callDestructors.cpp:1120]
>
> (Also at https://gist.github.com/swenson/34490cd907cc8330614e)
>
>
> I did some debugging, and the function where the asserting is reads:
>
> ...
> void insertReferenceTemps(CallExpr* call) {
>   for_formals_actuals(formal, actual, call)
> ...
>
> It appears as though the tuple is created to contain the return value, and
> when the destructor is invoked, it has to pass the implicit self (_ref),
> which is what I think insertReferenceTemps is doing. However,
> for_formals_actuals only works if the implicit self has already been
> inserted?
>
> I have a workaround that patches for_formals_actuals, but I'd be afraid it
> introduces more bugs than it cures, so I'm trying to figure out why this is
> only occurring here, and not for every destructor call. Any pointers would
> be appreciated. :)
>
> --Christopher
>

------------------------------------------------------------------------------
Transform Data into Opportunity.
Accelerate data analysis in your applications with
Intel Data Analytics Acceleration Library.
Click to learn more.
http://pubads.g.doubleclick.net/gampad/clk?id=278785471&iu=/4140
_______________________________________________
Chapel-bugs mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/chapel-bugs

Reply via email to