www.digitalmars.com         C & C++   DMDScript  

digitalmars.D.bugs - [Issue 8667] New: selective import breaks normal overload resolution

reply d-bugmail puremagic.com writes:
http://d.puremagic.com/issues/show_bug.cgi?id=8667

           Summary: selective import breaks normal overload resolution
           Product: D
           Version: D2
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: critical
          Priority: P2
         Component: DMD
        AssignedTo: nobody puremagic.com
        ReportedBy: dmitry.olsh gmail.com



06:03:26 PDT ---
The consequences of this bug are commonly observed by unwary as spurious
template instantation fails around std.string split and (previously) replace if
std.regex is imported.

This happens because std.string publicly and selectively imports a bunch of
functions from std.array and that brings about a pack of bad side effects.

(marked as critical as it cripples Phobos usage in a very unfriendly way)

The bug simplified:

//2 modules with unambigiuos template function
module m2;

void split(T)(T k)
    if(is(T : int)){}

//second one
module m;

void split(T)(T k)
    if(is(T : string))
{
}

//another one to call them
import m;
import m2: split; //removing : split makes it work

void main(){
    split("abc");
    split(123);
}

Output:
tryit.d(5): Error: template m2.split does not match any function template
declar
ation
tryit.d(5): Error: template m2.split(T) if (is(T : int)) cannot deduce template
function from argument types !()(string)

So, apparently, selectively imported symbol hides all others. 
Tested on DMD v2.060, was there since at least 2.056.

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
Sep 16 2012
next sibling parent d-bugmail puremagic.com writes:
http://d.puremagic.com/issues/show_bug.cgi?id=8667




With current dmd implementation, this is an expected behavior.
(But, I'm not sure whether is an expected language design.)

A selective import adds an alias declaration in importing module. So:

  import m;
  import m2: split; //removing : split makes it work

is same as:

  import m;
  import m2;
  alias split = m2.split;
  // in here, the name `split` has just one overload m2.split
  // (does not contain m1.split)

Therefore, in main, split("abc") does not match any function template
declaration.

===

In addition, renamed import works as same way. With current implementation,

  import m : x = split;

behaves same as:

  import m;
  alias x = m.split;

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
Jan 09 2013
prev sibling next sibling parent d-bugmail puremagic.com writes:
http://d.puremagic.com/issues/show_bug.cgi?id=8667




1. If the current behavior is correct,

     import mod : name;

   is just equal to

     import mod : name = name;

2. If this bug report is correct (== current behavior is incorrect),

     import mod : name;

   just makes `name` visible in symbol look up. And,

     import mod : name = name;

   merges overload set in the importing module.
   Then the two are different.



-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
Jan 09 2013
prev sibling parent d-bugmail puremagic.com writes:
http://d.puremagic.com/issues/show_bug.cgi?id=8667


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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |pull



https://github.com/D-Programming-Language/dmd/pull/2256

I finally concluded that the current selective import behavior is not good.
Then I fixed this issue in the pull request, but it's a breaking change.
Dmitry, could you please comment your opinion in github discussion?

-- 
Configure issuemail: http://d.puremagic.com/issues/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
Jul 09 2013