Retimer bug

We have a case where retimer cause eclipse to crash. If I go into the VC
debugger it says
Unhandled exception in eclipse.exe (ECLIPSE.DLL) 0xC00000FD: Stack
Overflow
so I guess this is the C stack.

Here is the trace of eclipse just before crashing
  (2267958) 16 CALL  repair_filter_constraints   %> creep
  (2267959) 17 CALL  repair_filter_constraints(0)   %> spy
 +(2267959) 17 CALL  repair_filter_constraints(0)   %> leap
Information: At level: 0        No. of conflicts:277.  After filtering:
277
 +(2267959) 17 ELSE  repair_filter_constraints(0)   %> leap
Information:  LP Count: 2
 +(2441222) 18 CALL  repair_filter_constraints(1)   %> leap
Information: At level: 1        No. of conflicts:279.  After filtering:
275
 +(2441222) 18 ELSE  repair_filter_constraints(1)   %> leap
Information:  LP Count: 3
 +(2614195) 19 CALL  repair_filter_constraints(2)   %> leap
Information: At level: 2        No. of conflicts:280.  After filtering:
280
 +(2614195) 19 ELSE  repair_filter_constraints(2)   %> leap
Information:  LP Count: 4
 +(2790677) 20 CALL  repair_filter_constraints(3)   %> leap
Information: At level: 3        No. of conflicts:288.  After filtering:
280
 +(2790677) 20 ELSE  repair_filter_constraints(3)   %> creep
  (2817767) 21 CALL  conflict_constraints(fleet, Cs)   %> skip
  (2817767) 21 EXIT  conflict_constraints(fleet, [...])   %> skip
  (2817773) 21 CALL  [...] \= []   %> skip
  (2817773) 21 EXIT  [...] \= []   %> skip
  (2817774) 21 CALL  fleet_constraint_keys([...], KeyedCs)   %> skip
  (2817774) 21 EXIT  fleet_constraint_keys([...], [...])   %> skip
  (2851473) 21 CALL  sort(0, >=, [...], [...])   %> skip
(Eclipse dies here)
I guess the sort bip is causing the C stack to overflow.

Here is the assembler level view. I think what we have is the pushing of
some stuff at the begining of a C function. Unfortunately I dont get a
call stack

1000BD46   pop         ebp
1000BD47   pop         ebx
1000BD48   ret
1000BD49   nop
1000BD4A   nop
1000BD4B   nop
1000BD4C   nop
1000BD4D   nop
1000BD4E   nop
1000BD4F   nop
1000BD50   mov         eax,dword ptr [esp+8]
1000BD54   mov         ecx,dword ptr [esp+10h]
1000BD58   push        ebx
1000BD59   push        ebp
1000BD5A   push        esi
1000BD5B   mov         esi,dword ptr [esp+10h]
1000BD5F   push        edi 					(FAILS
HERE)
1000BD60   mov         edi,dword ptr [esp+1Ch]
1000BD64   test        al,al
1000BD66   jl          1000BE6E
1000BD6C   test        cl,cl
1000BD6E   jl          1000BEC4
1000BD74   movsx       ebx,cl
1000BD77   mov         ebp,ebx
1000BD79   movsx       edx,al
1000BD7C   shl         ebp,4
1000BD7F   add         ebp,ebx
1000BD81   mov         ebx,edx

Interestingly this error seems to happen at exactly the same point with
or without the eclipse debugger.
I have made this happen on v 5.1.2 Apr 5th as well as 5.2

I have included a tgz file with all the relevant code:
Unpack
cd Retimer/UnitTests/bug558
eclipse -g 400M -b ../jtop.pl
run_test.
(The code includes a writeln(before_sort)
statement to show where the bug is happening.

2nd Floor The Tower Building	EMAIL:
Stefano.Novello@parc-technologies.com
11 York Road                      TEL: +44 20 72614017
SE1 7NX LONDON                    FAX: +44 20 72614001

This e-mail message is for the sole use of the intended recipient(s) -
its contents are the property of Parc Technologies Limited (or its
licensors) and are confidential. Please do not copy, review, use (except
for the intended purposes), disclose or distribute the e-mail or its
contents or allow anyone else to do so without our prior permission. 
Parc Technologies Limited does not guarantee that this e-mail has not
been intercepted and amended nor that it is virus-free. You should carry
out your own virus checks before opening any attachment. Any opinions
expressed in this e-mail message are those of the author and not
necessarily Parc Technologies Limited.
Limited.