http://d.puremagic.com/issues/show_bug.cgi?id=4841

           Summary: An array()/map inlining problem
           Product: D
           Version: D2
          Platform: x86
        OS/Version: Windows
            Status: NEW
          Severity: normal
          Priority: P2
         Component: DMD
        AssignedTo: nobody@puremagic.com
        ReportedBy: bearophile_hugs@eml.cc


--- Comment #0 from bearophile_hugs@eml.cc 2010-09-08 04:08:00 PDT ---
This is a D2 program:


import std.algorithm: map;
import std.array: array;
void main() {
    int c;
    array(map!((x){return c;})([1]));
}


It works if you compile it with dmd 2.048 with:
dmd test.d

But if use inlining:
dmd -inline test.d
It doesn't compile and dmd returns:
test.d(5): Error: function D main is a nested function and cannot be accessed
from array

I think this is a compiler bug because inlining should not affect the compilability of a program.

-- 
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=4841


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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |clugdbug@yahoo.com.au
            Summary|An array()/map inlining     |-inline wrecks delegate
                   |problem                     |literals as alias
                   |                            |parameters (An array()/map
                   |                            |inlining problem)


--- Comment #1 from Don <clugdbug@yahoo.com.au> 2010-09-10 00:02:24 PDT ---
Original title: "An array()/map inlining problem"
Here's a simpler case that gives the same error message, only with -inline:

---
struct Struct4841(alias pred)
{
    this(int k) { }
}

void bug4841a() {
      Struct4841!(   (t) { any_old_garbage;} )(  1 );
}

void bug4841b() {
   bug4841a();
}
---
This is a problem with delegates and alias templates. It's not a regression. There also seems to be an accepts-invalid bug in this. You can put any old garbage inside the alias, which seems odd. And you cannot declare a variable of type Struct4841!(   (t) { any_old_garbage;} ) --- it's rejected at the parsing stage.

-- 
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=4841


Trass3r <mrmocool@gmx.de> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |mrmocool@gmx.de


--- Comment #2 from Trass3r <mrmocool@gmx.de> 2012-01-16 12:41:31 PST ---
I think the following test case also belongs to this issue:

R1 find(alias pred = "a == b", R1, R2)(R1 haystack, R2 needle)
{
        return simpleMindedFind!pred(haystack, needle);
}
R1 simpleMindedFind(alias pred, R1, R2)(R1 haystack, R2 needle)
{
    bool haystackTooShort()
    {
            return true;
    }

    return haystack;
}

sizediff_t indexOf(Char1, Char2)(const(Char1)[] s, const(Char2)sub)
{
    const(Char1)[] balance = find!({})(s, sub);
    return -1;
}

string extStr;
void main()
{
    extStr.indexOf("bla");
}

-- 
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=4841


David Simcha <dsimcha@yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |dsimcha@yahoo.com


--- Comment #3 from David Simcha <dsimcha@yahoo.com> 2012-01-24 13:41:43 PST ---
*** Issue 4724 has been marked as a duplicate of this issue. ***

-- 
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=4841


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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |2.059
            Summary|-inline wrecks delegate     |-inline wrecks nested
                   |literals as alias           |struct with alias template
                   |parameters (An array()/map  |parameter (An array()/map
                   |inlining problem)           |inlining problem)


--- Comment #4 from Don <clugdbug@yahoo.com.au> 2012-01-27 00:35:53 PST ---
Further reduced shows neither constructor nor delegate is required.

struct Struct4841(alias pred)
{
    void unused_func();
}

void bug4841a() {
   int w;
   Struct4841!(  w ) m;
}

void bug4841b() {
   bug4841a();
}

----------------
The unused function is required only because it makes Struct4841 into a nested
struct.
I think it is because moving a templated nested struct changes its type - the
enclosing function is part of the mangled name, or something like that. It
can't trivially be moved around.

-- 
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=4841



--- Comment #5 from github-bugzilla@puremagic.com 2012-02-05 12:34:31 PST ---
Commit pushed to master at https://github.com/D-Programming-Language/dmd

https://github.com/D-Programming-Language/dmd/commit/7766bbd9857ab5ad5b1d5ff0444cb7088823a793
fix Issue 4841 - -inline wrecks nested struct with alias template parameter (An
array()/map inlining problem)

-- 
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=4841


Walter Bright <bugzilla@digitalmars.com> changed:

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


-- 
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=4841


David Simcha <dsimcha@yahoo.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |REOPENED
         Resolution|FIXED                       |


--- Comment #6 from David Simcha <dsimcha@yahoo.com> 2012-02-05 21:07:57 PST ---
I just tested the fix that was merged recently.  It only fixes some cases such as Don's and Tass3r's.  Bearophile's case is still broken, as is the case I reported in Bug 4724.

-- 
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=4841



--- Comment #7 from Walter Bright <bugzilla@digitalmars.com> 2012-02-05 22:08:20 PST ---
I believe this is the same as bug 5939.

-- 
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=4841



--- Comment #8 from Don <clugdbug@yahoo.com.au> 2012-02-06 10:00:46 PST ---
(In reply to comment #7)
> I believe this is the same as bug 5939.

I don't think bug 5939 involves -inline, but this one definitely does. Unless there are plans to make compiler changes are to fix bug 5939, the inliner shouldn't be making this transformation.

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