Eldar Insadutdinov Wrote:

> Another release of QtD is out. This time, it's Linux-only because OPTLINK refuses to link the project with debug info on Windows (see bug http://d.puremagic.com/issues/show_bug.cgi?id=2436). If anybody wants QtD on Windows, please keep pushing Walter until he does something with OPTLINK (bury it). Voting for the bug may help, too.
> I would repeat once again the request, that was raised here numerous of times, Walter, please change object file format, for me the ones used by MSVC or MinGW would be just fine.
> 
> In this release:
> 
> * Lots of bugfixes
> * Switched to Qt 4.5
> * ldc supported (for both X86 and X86-64)
> * All imports inside Qt are now public to avoid import hell
> * Implemented API dealing with containers
> * All classes from Gui, OpenGL, Xml, Svg, Network and Webkit packages
> are wrapped
> * Build system is now based on CMake to be crossplatform and more
> flexible
> * New signals and slots implementation (pretty limited but we are
> working on a better one, which will support queued connections,
> connections by name at runtime, etc).
> * Ported duic, the tool for generating code out of xml representation
> * Ported drcc, the resources compiler
> 
> Thanks to everyone involved in the project.

So apparently I made it working on Windows as well. The thing that solved the problem was to feed source files to the compiler in the different order. Is it a correct behaviour?

On Fri, May 29, 2009 at 12:44 AM, Eldar Insafutdinov <e.insafutdinov@gmail.com> wrote:
> Eldar Insadutdinov Wrote:
>
>> Another release of QtD is out. This time, it's Linux-only because OPTLINK refuses to link the project with debug info on Windows (see bug http://d.puremagic.com/issues/show_bug.cgi?id=2436). If anybody wants QtD on Windows, please keep pushing Walter until he does something with OPTLINK (bury it). Voting for the bug may help, too.
>> I would repeat once again the request, that was raised here numerous of times, Walter, please change object file format, for me the ones used by MSVC or MinGW would be just fine.
>>
>> In this release:
>>
>> * Lots of bugfixes
>> * Switched to Qt 4.5
>> * ldc supported (for both X86 and X86-64)
>> * All imports inside Qt are now public to avoid import hell
>> * Implemented API dealing with containers
>> * All classes from Gui, OpenGL, Xml, Svg, Network and Webkit packages
>> are wrapped
>> * Build system is now based on CMake to be crossplatform and more
>> flexible
>> * New signals and slots implementation (pretty limited but we are
>> working on a better one, which will support queued connections,
>> connections by name at runtime, etc).
>> * Ported duic, the tool for generating code out of xml representation
>> * Ported drcc, the resources compiler
>>
>> Thanks to everyone involved in the project.
>
> So apparently I made it working on Windows as well. The thing that solved the problem was to feed source files to the compiler in the different order. Is it a correct behaviour?
>

The DMD frontend has quite a few "order dependencies". It's not correct behaviour, but a result of how it does semantics.

On Fri, 29 May 2009 13:54:58 +0400, Tomas Lindquist Olsen <tomas.l.olsen@gmail.com> wrote:

> On Fri, May 29, 2009 at 12:44 AM, Eldar Insafutdinov <e.insafutdinov@gmail.com> wrote:
>> Eldar Insadutdinov Wrote:
>>
>>> Another release of QtD is out. This time, it's Linux-only because OPTLINK refuses to link the project with debug info on Windows (see bug http://d.puremagic.com/issues/show_bug.cgi?id=2436). If anybody wants QtD on Windows, please keep pushing Walter until he does something with OPTLINK (bury it). Voting for the bug may help, too. I would repeat once again the request, that was raised here numerous of times, Walter, please change object file format, for me the ones used by MSVC or MinGW would be just fine.
>>>
>>> In this release:
>>>
>>> * Lots of bugfixes
>>> * Switched to Qt 4.5
>>> * ldc supported (for both X86 and X86-64)
>>> * All imports inside Qt are now public to avoid import hell
>>> * Implemented API dealing with containers
>>> * All classes from Gui, OpenGL, Xml, Svg, Network and Webkit packages
>>> are wrapped
>>> * Build system is now based on CMake to be crossplatform and more
>>> flexible
>>> * New signals and slots implementation (pretty limited but we are
>>> working on a better one, which will support queued connections,
>>> connections by name at runtime, etc).
>>> * Ported duic, the tool for generating code out of xml representation
>>> * Ported drcc, the resources compiler
>>>
>>> Thanks to everyone involved in the project.
>>
>> So apparently I made it working on Windows as well. The thing that solved the problem was to feed source files to the compiler in the different order. Is it a correct behaviour?
>>
>
> The DMD frontend has quite a few "order dependencies". It's not correct behaviour, but a result of how it does semantics.

Is it fixable, or a new frontend is required?

Tomas Lindquist Olsen Wrote:

> On Fri, May 29, 2009 at 12:44 AM, Eldar Insafutdinov <e.insafutdinov@gmail.com> wrote:
> > Eldar Insadutdinov Wrote:
> >
> >> Another release of QtD is out. This time, it's Linux-only because OPTLINK refuses to link the project with debug info on Windows (see bug http://d.puremagic.com/issues/show_bug.cgi?id=2436). If anybody wants QtD on Windows, please keep pushing Walter until he does something with OPTLINK (bury it). Voting for the bug may help, too.
> >> I would repeat once again the request, that was raised here numerous of times, Walter, please change object file format, for me the ones used by MSVC or MinGW would be just fine.
> >>
> >> In this release:
> >>
> >> * Lots of bugfixes
> >> * Switched to Qt 4.5
> >> * ldc supported (for both X86 and X86-64)
> >> * All imports inside Qt are now public to avoid import hell
> >> * Implemented API dealing with containers
> >> * All classes from Gui, OpenGL, Xml, Svg, Network and Webkit packages
> >> are wrapped
> >> * Build system is now based on CMake to be crossplatform and more
> >> flexible
> >> * New signals and slots implementation (pretty limited but we are
> >> working on a better one, which will support queued connections,
> >> connections by name at runtime, etc).
> >> * Ported duic, the tool for generating code out of xml representation
> >> * Ported drcc, the resources compiler
> >>
> >> Thanks to everyone involved in the project.
> >
> > So apparently I made it working on Windows as well. The thing that solved the problem was to feed source files to the compiler in the different order. Is it a correct behaviour?
> >
> 
> The DMD frontend has quite a few "order dependencies". It's not correct behaviour, but a result of how it does semantics.

Why is it working on linux then? With both ldc and dmd?

On Fri, May 29, 2009 at 11:57 AM, Denis Koroskin <2korden@gmail.com> wrote:
> On Fri, 29 May 2009 13:54:58 +0400, Tomas Lindquist Olsen <tomas.l.olsen@gmail.com> wrote:
>
>> On Fri, May 29, 2009 at 12:44 AM, Eldar Insafutdinov <e.insafutdinov@gmail.com> wrote:
>>> Eldar Insadutdinov Wrote:
>>>
>>>> Another release of QtD is out. This time, it's Linux-only because OPTLINK refuses to link the project with debug info on Windows (see bug http://d.puremagic.com/issues/show_bug.cgi?id=2436). If anybody wants QtD on Windows, please keep pushing Walter until he does something with OPTLINK (bury it). Voting for the bug may help, too. I would repeat once again the request, that was raised here numerous of times, Walter, please change object file format, for me the ones used by MSVC or MinGW would be just fine.
>>>>
>>>> In this release:
>>>>
>>>> * Lots of bugfixes
>>>> * Switched to Qt 4.5
>>>> * ldc supported (for both X86 and X86-64)
>>>> * All imports inside Qt are now public to avoid import hell
>>>> * Implemented API dealing with containers
>>>> * All classes from Gui, OpenGL, Xml, Svg, Network and Webkit packages
>>>> are wrapped
>>>> * Build system is now based on CMake to be crossplatform and more
>>>> flexible
>>>> * New signals and slots implementation (pretty limited but we are
>>>> working on a better one, which will support queued connections,
>>>> connections by name at runtime, etc).
>>>> * Ported duic, the tool for generating code out of xml representation
>>>> * Ported drcc, the resources compiler
>>>>
>>>> Thanks to everyone involved in the project.
>>>
>>> So apparently I made it working on Windows as well. The thing that solved the problem was to feed source files to the compiler in the different order. Is it a correct behaviour?
>>>
>>
>> The DMD frontend has quite a few "order dependencies". It's not correct behaviour, but a result of how it does semantics.
>
> Is it fixable, or a new frontend is required?
>

Of course it's fixable :) But how much work it is, that's the real question ...

On Fri, May 29, 2009 at 12:22 PM, Eldar Insafutdinov <e.insafutdinov@gmail.com> wrote:
> Tomas Lindquist Olsen Wrote:
>
>> On Fri, May 29, 2009 at 12:44 AM, Eldar Insafutdinov <e.insafutdinov@gmail.com> wrote:
>> > Eldar Insadutdinov Wrote:
>> >
>> >> Another release of QtD is out. This time, it's Linux-only because OPTLINK refuses to link the project with debug info on Windows (see bug http://d.puremagic.com/issues/show_bug.cgi?id=2436). If anybody wants QtD on Windows, please keep pushing Walter until he does something with OPTLINK (bury it). Voting for the bug may help, too.
>> >> I would repeat once again the request, that was raised here numerous of times, Walter, please change object file format, for me the ones used by MSVC or MinGW would be just fine.
>> >>
>> >> In this release:
>> >>
>> >> * Lots of bugfixes
>> >> * Switched to Qt 4.5
>> >> * ldc supported (for both X86 and X86-64)
>> >> * All imports inside Qt are now public to avoid import hell
>> >> * Implemented API dealing with containers
>> >> * All classes from Gui, OpenGL, Xml, Svg, Network and Webkit packages
>> >> are wrapped
>> >> * Build system is now based on CMake to be crossplatform and more
>> >> flexible
>> >> * New signals and slots implementation (pretty limited but we are
>> >> working on a better one, which will support queued connections,
>> >> connections by name at runtime, etc).
>> >> * Ported duic, the tool for generating code out of xml representation
>> >> * Ported drcc, the resources compiler
>> >>
>> >> Thanks to everyone involved in the project.
>> >
>> > So apparently I made it working on Windows as well. The thing that solved the problem was to feed source files to the compiler in the different order. Is it a correct behaviour?
>> >
>>
>> The DMD frontend has quite a few "order dependencies". It's not correct behaviour, but a result of how it does semantics.
>
> Why is it working on linux then? With both ldc and dmd?
>

I can't answer that, it's entirely possible it's a whole other issue.

However, it could be that a version block is pulling in some trouble code, that is otherwise skipped on other platforms than windows.

In the end you should really construct a test case to get it fixed.