www.digitalmars.com         C & C++   DMDScript  

digitalmars.D.bugs - [Issue 11853] New: Tuples fail "isAssignable"

reply d-bugmail puremagic.com writes:
https://d.puremagic.com/issues/show_bug.cgi?id=11853

           Summary: Tuples fail "isAssignable"
           Product: D
           Version: D2
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: regression
          Priority: P2
         Component: Phobos
        AssignedTo: nobody puremagic.com
        ReportedBy: monarchdodra gmail.com



Works in 2.064.2. Fails in master:

//----
import std.traits, std.typecons;

void main()
{
    alias T = Tuple!int;
    static assert(isAssignable!T);
}
//----
main.d(6): Error: static assert  (isAssignable!(Tuple!int)) is false
//----

-- 
Configure issuemail: https://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
Dec 31 2013
next sibling parent d-bugmail puremagic.com writes:
https://d.puremagic.com/issues/show_bug.cgi?id=11853




According to "dissect", the faulty commit is:
Fix swap for non-assignable:
https://github.com/D-Programming-Language/phobos/commit/97e1fb80d404899cc14d03483f61986791747e38#diff-ff74a46362b5953e8c88120e2490f839

Which added this static test in "swap":
static if (!isAssignable!T || hasElaborateAssign!T)

The bummer is that tuple's opAssign looks like:

        void opAssign(R)(auto ref R rhs)
        if (areCompatibleTuples!(typeof(this), R, "="))
        {
            import std.algorithm : swap;
            swap!(Tuple!Types)(this, rhs);

I'm not quite sure *how* the compiler resolves that:
1. To instantiate opAssign, we need swap...
2. For swap we need to know if assignable...
3. To know if assignable, we need to instantiate opAssign...

In particular, it gives "funny" scenarios such as:

//----
alias T = Tuple!int;

void main()
{
    T t;
    static assert(isAssignable!T); //FAILS
}
//----
alias T = Tuple!int;

void main()
{
    T t;
    t = T.init;
    static assert(isAssignable!T); //PASSES
}
//----

We can workaround the problem by checking "hasElaborateAssign" first, as that
check does not actually look into the implementation. Still, there is some bad
smell here.

-- 
Configure issuemail: https://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
Dec 31 2013
prev sibling next sibling parent d-bugmail puremagic.com writes:
https://d.puremagic.com/issues/show_bug.cgi?id=11853




Commits pushed to master at https://github.com/D-Programming-Language/phobos

https://github.com/D-Programming-Language/phobos/commit/5d583483e3a997562d036dac872ae8548e73483f
fix Issue 11853 - Tuples fail "isAssignable"

https://github.com/D-Programming-Language/phobos/commit/327085b8fed061853b0a9c8bdc42df9e495d0315


fix Issue 11853 - Tuples fail "isAssignable"

-- 
Configure issuemail: https://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
Dec 31 2013
prev sibling parent d-bugmail puremagic.com writes:
https://d.puremagic.com/issues/show_bug.cgi?id=11853


Walter Bright <bugzilla digitalmars.com> changed:

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


-- 
Configure issuemail: https://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
Dec 31 2013