www.digitalmars.com         C & C++   DMDScript  

digitalmars.D.bugs - [Issue 6306] New: [CTFE] Strange behavior of indirect recursive call in CTFE

reply d-bugmail puremagic.com writes:
http://d.puremagic.com/issues/show_bug.cgi?id=6306

           Summary: [CTFE] Strange behavior of indirect recursive call in
                    CTFE
           Product: D
           Version: D2
          Platform: x86
        OS/Version: Windows
            Status: NEW
          Severity: critical
          Priority: P2
         Component: DMD
        AssignedTo: nobody puremagic.com
        ReportedBy: youxkei gmail.com


--- Comment #0 from Hisayuki Mima <youxkei gmail.com> 2011-07-13 02:06:02 PDT
---
void main(){}

struct Result{bool match; string value; string rest;}

Result paren(string input){
    if(input[0] == '('){
        input = input[1..$];
    }else{
        return Result(false, "", input);
    }
    string[] strs;
    while(true){
        Result r = function(string input){
            Result r = paren(input);
            if(r.match){
                return r;
            }
            if(input[0] == ')'){
                return Result(false, "", input);
            }else{
                return Result(true, input[0..1], input[1..$]);
            }
        }(input);
        if(r.match){
            strs = strs ~ r.value;
            input = r.rest;
        }else{
            break;
        }
    }
    if(input[0] == ')'){
        input = input[1..$];
    }else{
        return Result(false, "", input);
    }

    string value = "(";
    foreach(str; strs){
        value = value ~ str;
    }
    value = value ~ ")";
    return Result(true, value, input);
}

bool test(){//unittest
    auto r = paren("((a))");
    assert(r.match);
    assert(r.rest == "");
    assert(r.value == "((a))"); //assertion1
    return true;
}

unittest{
    static assert(test()); //line1
    test();
}

This code cannot be compiled by dmd v2.054 because the assertion1, which is the
compile-time assertion, fails.
(Incidentally, r.value is "(a(a))" at assertion1.)
However, dmd v2.052 can compile this code well.
When I  commented out the line 1, the code became able to be compiled dmd
v2.054 and the run-time unittest succeeded.

The function paren is a function which parses string in balanced-parentheses
and returns parsed string.
It has indirect recursive call or it calls a function which calls it.
When I rewrote the code as a code which has direct recursive call, the code the
code became able to be compiled dmd v2.054 and the run-time unittest succeeded,
too.
This is why I think indirect recursive call has strange behavior and it causes
the assertion1 to fail.

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
Jul 13 2011
next sibling parent d-bugmail puremagic.com writes:
http://d.puremagic.com/issues/show_bug.cgi?id=6306


Don <clugdbug yahoo.com.au> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |wrong-code
                 CC|                            |clugdbug yahoo.com.au


--- Comment #1 from Don <clugdbug yahoo.com.au> 2011-07-27 00:25:42 PDT ---
Reduced test case shows this is very nasty bug in failing to restore local
variable values after an indirect recursive call.
-------------
void recurse6306() {
    bug6306(false);
}

bool bug6306(bool b) {
    int x = 0;
    if (b)
        recurse6306();
    assert(x == 0); // fails!!!!!
    x = 1;
    return true;
}

static assert( bug6306(true) );

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
Jul 27 2011
prev sibling next sibling parent d-bugmail puremagic.com writes:
http://d.puremagic.com/issues/show_bug.cgi?id=6306


Don <clugdbug yahoo.com.au> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[CTFE] Strange behavior of  |Regression(2.054): [CTFE]
                   |indirect recursive call in  |Strange behavior of
                   |CTFE                        |indirect recursive call in
                   |                            |CTFE
           Severity|critical                    |regression


-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
Jul 27 2011
prev sibling parent d-bugmail puremagic.com writes:
http://d.puremagic.com/issues/show_bug.cgi?id=6306


Walter Bright <bugzilla digitalmars.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |bugzilla digitalmars.com
         Resolution|                            |FIXED


--- Comment #2 from Walter Bright <bugzilla digitalmars.com> 2011-07-31
12:08:56 PDT ---
https://github.com/D-Programming-Language/dmd/commit/7ef3b2bb9e740df39108957ae5e3b2aa8253d351

https://github.com/D-Programming-Language/dmd/pull/284

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
Jul 31 2011