www.digitalmars.com         C & C++   DMDScript  

digitalmars.D.bugs - [Issue 1629] New: Link error: Previous Definition Different: blablah__initZ

reply d-bugmail puremagic.com writes:
http://d.puremagic.com/issues/show_bug.cgi?id=1629

           Summary: Link error: Previous Definition Different:
                    blablah__initZ
           Product: D
           Version: 1.022
          Platform: PC
        OS/Version: Windows
            Status: NEW
          Severity: normal
          Priority: P2
         Component: DMD
        AssignedTo: bugzilla digitalmars.com
        ReportedBy: wbaxter gmail.com


The error seems to require a sort of a diamond pattern of imports.  So it takes
several files to get it to happen.  Here's the minimal I've been able to get so
far:
---link_main.d---
module link_main;
import link_a;
import link_b;
link_a.Thing thinga;
link_b.Thing thingb;
void main()
{
}
---link_a.d---
module link_a;
import link_common;
alias link_common.CommonThing!() Thing;
---link_b.d---
module link_b;
import link_common;
alias link_common.CommonThing!() Thing;
---link_common.d---
module link_common;
import link_default;

struct CommonThing(T = DefaultT)
{
    struct InnerT {
        alias DefaultT.Type Type;
    }
    const ConstValue = DefaultT.ConstValue;
}
---link_default.d--
module link_default;
struct DefaultT
{
    alias float Type ;
    const uint ConstValue = 0;
}
---

Here's the exact compilation error:

rebuild -oqobjs link_main
f:\usr\pkg\d\dmd\bin\..\..\dm\bin\link.exe
objs\_link_main+objs\_link_a+objs\_link_common+objs\_link_default+objs\_link_b,link_main.exe,,user32+kernel32/noi;

OPTLINK (R) for Win32  Release 7.50B1
Copyright (C) Digital Mars 1989 - 2001  All Rights Reserved

objs\_link_b.obj(_link_b)  Offset 001ADH Record Type 0091 
 Error 1: Previous Definition Different :
_D11link_common17__T11CommonThingZ11CommonThing6InnerT6__initZ
--- errorlevel 1



At least in my actual case where this came up I found I could make the problem
go away if I created yet another module that link_a and link_b import that
looks like:

    public import link_common;
    alias link_common.CommonThing!() Thing;

This seems to clue the linker in that the two aliases are really instantiations
of the same template.


--
Oct 31 2007
next sibling parent d-bugmail puremagic.com writes:
http://d.puremagic.com/issues/show_bug.cgi?id=1629





------- Comment #1 from wbaxter gmail.com  2007-10-31 03:05 -------
Here's an NG post where I originally asked about this problem.
http://www.digitalmars.com/d/archives/digitalmars/D/learn/about_dsss_and_blablablah_initZ_previous_definition_different_9247.html

At the time I couldn't narrow it down to a small repro.


--
Oct 31 2007
prev sibling next sibling parent d-bugmail puremagic.com writes:
http://d.puremagic.com/issues/show_bug.cgi?id=1629





------- Comment #2 from wbaxter gmail.com  2007-11-03 23:47 -------
Note that the repro case uses 'rebuild' for compiling. It happens to use
version 0.73, which has the compile-one-file-at-a-time behavior that Gregor put
in to fix another linking error.

The compilation works fine with dmd passing all files at once, the compiling
separately one-at-a-time is what causes the failure.


--
Nov 03 2007
prev sibling next sibling parent d-bugmail puremagic.com writes:
http://d.puremagic.com/issues/show_bug.cgi?id=1629


matti.niemenmaa+dbugzilla iki.fi changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |matti.niemenmaa+dbugzilla ik
                   |                            |i.fi




------- Comment #3 from matti.niemenmaa+dbugzilla iki.fi  2008-07-19 08:35
-------
Just ran into this with Tango's XML package (tango.text.xml.DocPrinter to be
exact).This one's a bit different though, as one needs -inline and -release
thrown to trigger it.

Minimal testcase follows. Three files, no diamond pattern of imports, just a
straight line:

---------------------------
a.d:
----
module a;
import b;
void main() {}

b.d
---
module b;
import c;

// works - even fixes the error from below!
// C!(int) x;

// doesn't work
void foo() { C!(int) x; }

c.d
---
module c;
class C(T) { struct X {} }
---------------------------

Compile separately and link:

dmd -c a.d -inline -release
dmd -c b.d -inline -release
dmd -c c.d -inline -release
dmd a.obj b.obj c.obj

And get:

OPTLINK (R) for Win32  Release 8.00.1
Copyright (C) Digital Mars 1989-2004  All rights reserved.
b.obj(b)  Offset 0014EH Record Type 0091
 Error 1: Previous Definition Different : _D1c8__T1CTiZ1C1X6__initZ
--- errorlevel 1

Note the comment in b.d: adding an instantiation (or alias, thanks to the
original reporter for that hint) at global scope fixes the problem.


--
Jul 19 2008
prev sibling next sibling parent d-bugmail puremagic.com writes:
http://d.puremagic.com/issues/show_bug.cgi?id=1629


bugzilla digitalmars.com changed:

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




------- Comment #4 from bugzilla digitalmars.com  2009-03-11 14:51 -------
Fixed dmd 1.041 and 2.026


--
Mar 11 2009
prev sibling parent d-bugmail puremagic.com writes:
http://d.puremagic.com/issues/show_bug.cgi?id=1629


Don <clugdbug yahoo.com.au> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |2korden gmail.com




--- Comment #5 from Don <clugdbug yahoo.com.au>  2009-09-04 07:03:36 PDT ---
*** Issue 2530 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: -------
Sep 04 2009