http://d.puremagic.com/issues/show_bug.cgi?id=7883

           Summary: Compiler crashes with double inheritance and function
                    contract
           Product: D
           Version: D2
          Platform: x86
        OS/Version: Windows
            Status: NEW
          Severity: normal
          Priority: P2
         Component: DMD
        AssignedTo: nobody@puremagic.com
        ReportedBy: dlang@grotmol.net


--- Comment #0 from Øyvind Grotmol <dlang@grotmol.net> 2012-04-10 03:14:43 PDT ---
Created an attachment (id=1087)
Test case crashing the compiler

Class C inherits from B which inherits from A. All three classes define the same function. A and B specifies an out contract on the function (C may or may not specify an out contract, it doesn't matter). In this case, dmd crashes without error message, and doesn't generate the .exe file.

Here is the verbose output:
$ dmd -v e
binary    C:\D\dmd2\windows\bin\dmd.exe
version   v2.058
config    C:\D\dmd2\windows\bin\sc.ini
parse     e
importall e
import    object
(C:\D\dmd2\windows\bin\..\..\src\druntime\import\object.di)
semantic  e
semantic2 e
semantic3 e

At this point it just stops. The rest of the output should have been as
follows, if I remove one of the two contracts:
code      e
function  e.A.foo
function  e.A.foo.__ensure
function  e.B.foo
function  e.C.foo
function  D main
C:\D\dmd2\windows\bin\link.exe e,,nul,user32+kernel32/noi;

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------

http://d.puremagic.com/issues/show_bug.cgi?id=7883



--- Comment #1 from Øyvind Grotmol <dlang@grotmol.net> 2012-04-10 03:18:09 PDT ---
Note the attached file has commented out one of the contracts and thus compiles; remove the "//" and the crash arises.

Here is the Windows 7 crash information:

Problem signature:
  Problem Event Name:    APPCRASH
  Application Name:    dmd.exe
  Application Version:    0.0.0.0
  Application Timestamp:    00000000
  Fault Module Name:    dmd.exe
  Fault Module Version:    0.0.0.0
  Fault Module Timestamp:    00000000
  Exception Code:    c0000005
  Exception Offset:    0004d784
  OS Version:    6.1.7601.2.1.0.256.48
  Locale ID:    1044
  Additional Information 1:    c0d2
  Additional Information 2:    c0d2fd7c47233db608dcc0a8dfa7e295
  Additional Information 3:    77a4
  Additional Information 4:    77a4546d10565e53c3340bd335ee63b4

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------

http://d.puremagic.com/issues/show_bug.cgi?id=7883



--- Comment #2 from Øyvind Grotmol <dlang@grotmol.net> 2012-04-10 03:19:46 PDT ---
By the way, the test case compiles fine with gdc

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------

http://d.puremagic.com/issues/show_bug.cgi?id=7883



--- Comment #3 from Øyvind Grotmol <dlang@grotmol.net> 2012-04-10 07:06:35 PDT ---
(From update of attachment 1087)
This file compiles as it stands; remove the "//" to make it crash dmd.

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------

http://d.puremagic.com/issues/show_bug.cgi?id=7883



--- Comment #5 from Øyvind Grotmol <dlang@grotmol.net> 2012-04-21 11:52:23 PDT ---
(In reply to comment #4)
> It compiles and runs on DMD 2.059 Win32

Hi,

Could you check this one also please?
By mistake the file I uploaded had commented out one of the contracts, and then
the file does indeed work. It is only when both contracts are in place, that
dmd crashes.

class A {
    int foo(int i)
    out (result) { }
    body { return 1; }
}

class B : A {
    int foo(int i)
    out (result) { }
    body { return 1; }

}

class C : B {
    int foo(int i) { return 1; }
}

void main() {
}

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------

http://d.puremagic.com/issues/show_bug.cgi?id=7883



--- Comment #6 from SomeDude <lovelydear@mailmetrash.com> 2012-04-21 12:13:38 PDT ---
(In reply to comment #5)
> (In reply to comment #4)
> > It compiles and runs on DMD 2.059 Win32
> 
> Hi,
> 
> Could you check this one also please?
> By mistake the file I uploaded had commented out one of the contracts, and then
> the file does indeed work. It is only when both contracts are in place, that
> dmd crashes.
> 


Yes, it crashes DMD Win32 here too.

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------

http://d.puremagic.com/issues/show_bug.cgi?id=7883


Don <clugdbug@yahoo.com.au> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |ice
                 CC|                            |clugdbug@yahoo.com.au
            Version|D2                          |D1 & D2
            Summary|Compiler crashes with       |Compiler segfaults with
                   |double inheritance and      |double inheritance and
                   |function contract           |function contract
         OS/Version|Windows                     |All


--- Comment #7 from Don <clugdbug@yahoo.com.au> 2012-05-14 03:16:38 PDT ---
Also crashes on D1 Linux.

It's in func.c(1987), in FuncDeclaration::mergeFensure(Statement * sf)

fensure is NULL (In fact, it's called with fensure = mergeFensure(fensure)

            if (sf)
            {
                sf = new CompoundStatement(fensure->loc, s2, sf);
            }

Changing this to sf->loc prevents the segfault. But, there's something else wrong -- after this change, it prints

bug.d(21): Error: undefined identifier result

whereas it should compile without error.

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------

http://d.puremagic.com/issues/show_bug.cgi?id=7883


Henning Pohl <henning@still-hidden.de> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |pull
                 CC|                            |henning@still-hidden.de


--- Comment #8 from Henning Pohl <henning@still-hidden.de> 2013-07-24 12:29:54 PDT ---
https://github.com/D-Programming-Language/dmd/pull/2379

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------

http://d.puremagic.com/issues/show_bug.cgi?id=7883



--- Comment #9 from github-bugzilla@puremagic.com 2013-07-25 20:01:03 PDT ---
Commits pushed to master at https://github.com/D-Programming-Language/dmd

https://github.com/D-Programming-Language/dmd/commit/ba2caff1a4819fecd36954a627f07019ec5ff99c
fix issue 7883 - Compiler segfaults with double inheritance and function
contract

https://github.com/D-Programming-Language/dmd/commit/ba1ed2c4aaf52edc90e8cd64e736896b31ecf012 Merge pull request #2379 from hpohl/7883

fix issue 7883 - Compiler segfaults with double inheritance and function contract

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------

http://d.puremagic.com/issues/show_bug.cgi?id=7883


Kenji Hara <k.hara.pg@gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Version|D1 & D2                     |D1
            Summary|Compiler segfaults with     |(D1 only) Compiler
                   |double inheritance and      |segfaults with double
                   |function contract           |inheritance and function
                   |                            |contract


--- Comment #10 from Kenji Hara <k.hara.pg@gmail.com> 2013-07-25 20:08:44 PDT ---
Fixed in D2.

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------