[Issue 11648] rangeerror when adding element to associative array and value is implicitely converted by alias this

[d-bugmail at puremagic.com]

https://d.puremagic.com/issues/show_bug.cgi?id=11648


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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID


--- Comment #1 from Kenji Hara <k.hara.pg at gmail.com> 2013-11-30 09:29:22 PST ---
(In reply to comment #0)
> dmd 2.063 compiles and runs this code correctly:
> 
> //////////////////
> import std.stdio;
> 
> struct S
> {
>     int _payload;
>     alias _payload this;
> }
> 
> void main()
> {
>     S[string] symbols;
>     symbols["a"] = 1; // works if using S(1)
>     writeln(symbols); // ["a":1]
> }
> ////////////////
> 
> but dmd 2.064 causes a range error when executing:
> 
> core.exception.RangeError at test(12): Range violation

This is intended behavior introduce by fixing bug 6178.

Test case in compiler test suite.
https://github.com/D-Programming-Language/dmd/pull/2539/files#diff-a137107b45f82e7d3b516e2aff79f1d5R940

First, 'alias this' does not work for object construction.

struct S { int val; alias val this; }
S s = 1;   // compile error

And in your case, AA does not have a storage corresponding to the key "a" yet.
So, `symbols["a"] = 1;` should invoke _construction of S from integer 1_. But
as I shown, S cannot construct from int. So, compiler always treat the line as
_an assignment through alias this_. And of course, if the specified key does
not exist in AA, the indexing should throw RangeError.

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