http://sourceforge.net/project/showfiles.php?group_id=154306

Changes:
  * Added support for GCC 4.1.x
  * Support for GCC 3.3.x is deprecated
  * Updated to DMD 1.004
  * Fixed Bugzilla 836, 837, 838, 839, 841, 843, 844, 889, 896
  * Fixed problems with missing debug information
  * Added Rick Mann's -frameowrk patch for gdmd

SkyOS support will return in the next release.

David Friedman Wrote:

> http://sourceforge.net/project/showfiles.php?group_id=154306
> 
> Changes:
>    * Added support for GCC 4.1.x
>    * Support for GCC 3.3.x is deprecated
>    * Updated to DMD 1.004
>    * Fixed Bugzilla 836, 837, 838, 839, 841, 843, 844, 889, 896
>    * Fixed problems with missing debug information
>    * Added Rick Mann's -frameowrk patch for gdmd
> 
> SkyOS support will return in the next release.

Nice

David Friedman wrote:
> http://sourceforge.net/project/showfiles.php?group_id=154306
> 
> Changes:
>   * Added support for GCC 4.1.x
>   * Support for GCC 3.3.x is deprecated
>   * Updated to DMD 1.004
>   * Fixed Bugzilla 836, 837, 838, 839, 841, 843, 844, 889, 896
>   * Fixed problems with missing debug information
>   * Added Rick Mann's -frameowrk patch for gdmd
> 
> SkyOS support will return in the next release.

4.1.x support!  Hoorah!

 - Gregor Richards

Gregor Richards wrote:

> David Friedman wrote:
>> http://sourceforge.net/project/showfiles.php?group_id=154306
>> 
>> Changes:
>>   * Added support for GCC 4.1.x
>>   * Support for GCC 3.3.x is deprecated
>>   * Updated to DMD 1.004
>>   * Fixed Bugzilla 836, 837, 838, 839, 841, 843, 844, 889, 896
>>   * Fixed problems with missing debug information
>>   * Added Rick Mann's -frameowrk patch for gdmd
>> 
>> SkyOS support will return in the next release.
> 
> 4.1.x support!  Hoorah!
> 

Yeah!

Now all I need is x86_64.

Gregor Richards Wrote:

> David Friedman wrote:
> > http://sourceforge.net/project/showfiles.php?group_id=154306
> > 
> > Changes:
> >   * Added support for GCC 4.1.x
> >   * Support for GCC 3.3.x is deprecated
> >   * Updated to DMD 1.004
> >   * Fixed Bugzilla 836, 837, 838, 839, 841, 843, 844, 889, 896
> >   * Fixed problems with missing debug information
> >   * Added Rick Mann's -frameowrk patch for gdmd
> > 
> > SkyOS support will return in the next release.
> 
> 4.1.x support!  Hoorah!
> 
>   - Gregor Richards

What benefits does 4.1.x provide vs 3.4.x?

David Friedman wrote:

> http://sourceforge.net/project/showfiles.php?group_id=154306
> 
> Changes:
>   * Added support for GCC 4.1.x
>   * Support for GCC 3.3.x is deprecated
>   * Updated to DMD 1.004
>   * Fixed Bugzilla 836, 837, 838, 839, 841, 843, 844, 889, 896
>   * Fixed problems with missing debug information
>   * Added Rick Mann's -frameowrk patch for gdmd

Regarding the gdcmac/gdcwin/gdcgnu packages, they are
now offering the *same* GDC binaries as David provides.

D 1.0 was the end of the line for the deprecated builds,
like GCC 3.4.2 for MinGW and GCC 3.3 for Mac OS X 10.3.


Going forward, these are the supported targets for each:

gdcmac: Mac OS X 10.4 Universal, Apple GCC 4.0 (Xcode 2.4)
        from gdc-0.22-macosx-universal-gcc-5363.tar.bz2

gdcwin: MinGW32 i386, MinGW GCC 3.4.5 ("Candidate" 20060117-1)
        from gdc-0.22-mingw32-gcc-3.4.5.tar.bz2

gdcgnu: GNU/Linux i686/powerpc, FSF GCC 4.1.1 (20060524) *
        from gdc-0.22-i686-linux-gcc-4.1.1.tar.bz2 /
             gdc-0.22-powerpc-linux-gcc-4.1.1.tar.bz2

That is, whatever targets that the "dgcc" project offers...

--anders


Note: Fedora calls its GCC targets "i386-redhat-linux"
and "ppc-redhat-linux", while Ubuntu instead calls them
"i486-linux-gnu" and "powerpc-linux-gnu". So installing
the GDC files shouldn't replace any system GCC binaries.
(GDC uses "i686-pc-linux-gnu/"powerpc-unknown-linux-gnu")
The duplicated files are: libgcc.a, crt*.o and collect2

John wrote:

> Gregor Richards Wrote:
> 
>> David Friedman wrote:
>> > http://sourceforge.net/project/showfiles.php?group_id=154306
>> > 
>> > Changes:
>> >   * Added support for GCC 4.1.x
>> >   * Support for GCC 3.3.x is deprecated
>> >   * Updated to DMD 1.004
>> >   * Fixed Bugzilla 836, 837, 838, 839, 841, 843, 844, 889, 896
>> >   * Fixed problems with missing debug information
>> >   * Added Rick Mann's -frameowrk patch for gdmd
>> > 
>> > SkyOS support will return in the next release.
>> 
>> 4.1.x support!  Hoorah!
>> 
>>   - Gregor Richards
> 
> What benefits does 4.1.x provide vs 3.4.x?

I may be misunderstanding something, but my idea was that I'm using Fedora FC6, and everything on this system is built with gcc-4.1.1 (with additional patches).  I thought that if gdc is based on gcc-4.1.1 maybe that would simplify coexistence.  I was really hoping I could take the gdc patches and apply them to the redhat (fedora) gcc version.  It remains to be seen if this is true.  I haven't tried yet, because I'm using x86-64.

"Anders F Björklund" <afb@algonet.se> wrote in message news:eqaqsm$fat$1@digitaldaemon.com...
> D 1.0 was the end of the line for the deprecated builds, like GCC 3.4.2 for MinGW and GCC 3.3 for Mac OS X 10.3.
>
> Going forward, these are the supported targets for each:
>
...
> gdcwin: MinGW32 i386, MinGW GCC 3.4.5 ("Candidate" 20060117-1)
>         from gdc-0.22-mingw32-gcc-3.4.5.tar.bz2
>

Pardon if this is a stupid question, but I'm unclear on one thing regarding trying to compile GDC from sources: Can GDC be built on MSYS/MinGW using GCC 3.4.2, or will it fail unless I upgrade GCC to 3.4.5?

The reason I ask is that I'm currently getting build errors, and this snippit looks like it *might* be saying I need to go up to 3.4.5+, but I'm not sure.

(And yes, I see this GDC 0.22 is outdated since 0.23 just came out: I'll grab that in a bit)

Nick Sabalausky wrote:

>>gdcwin: MinGW32 i386, MinGW GCC 3.4.5 ("Candidate" 20060117-1)
>>        from gdc-0.22-mingw32-gcc-3.4.5.tar.bz2
> 
> Pardon if this is a stupid question, but I'm unclear on one thing regarding trying to compile GDC from sources: Can GDC be built on MSYS/MinGW using GCC 3.4.2, or will it fail unless I upgrade GCC to 3.4.5?

For my part I just got bored with building two sets of my packages when
there are now official tarballs from the actual GDC development project.

It used to build OK with MinGW GCC 3.4.2 when I did the GDC 0.21 builds,
but I haven't tried since. The post D 1.0 packages now all use David's.

> The reason I ask is that I'm currently getting build errors, and this snippit looks like it *might* be saying I need to go up to 3.4.5+, but I'm not sure.

As far as I know, all versions of GCC 3.4.x are supported by GDC.
Maybe your build errors are about something else ? (any details ?)

I used this: http://gdcwin.sourceforge.net/gcc-3.4.2-build.sh
after unpacking GCC and GDC and running the setup-gcc.sh script.

--anders

"Anders F Björklund" <afb@algonet.se> wrote in message news:eslrbm$15q8$1@digitalmars.com...
> Nick Sabalausky wrote:
>
>>>gdcwin: MinGW32 i386, MinGW GCC 3.4.5 ("Candidate" 20060117-1)
>>>        from gdc-0.22-mingw32-gcc-3.4.5.tar.bz2
>>
>> Pardon if this is a stupid question, but I'm unclear on one thing regarding trying to compile GDC from sources: Can GDC be built on MSYS/MinGW using GCC 3.4.2, or will it fail unless I upgrade GCC to 3.4.5?
>
> For my part I just got bored with building two sets of my packages when there are now official tarballs from the actual GDC development project.
>
> It used to build OK with MinGW GCC 3.4.2 when I did the GDC 0.21 builds, but I haven't tried since. The post D 1.0 packages now all use David's.
>
>> The reason I ask is that I'm currently getting build errors, and this snippit looks like it *might* be saying I need to go up to 3.4.5+, but I'm not sure.
>
> As far as I know, all versions of GCC 3.4.x are supported by GDC. Maybe your build errors are about something else ? (any details ?)
>
> I used this: http://gdcwin.sourceforge.net/gcc-3.4.2-build.sh after unpacking GCC and GDC and running the setup-gcc.sh script.
>
> --anders

I got errors at different places depending on whether or not I include c++.

With c++, it errored around libcpp "0 [main] make 2264 open_stackdumpfile: Dumping stack trace to make.exe.stackdump". I think I did something sightly different that time than in subsequent runs though. I may want to retry this.

If I omit c++, then it fails while compiling phobos (std/c/windows/com.d:240: error: '*LTHUNK0' aliased to undefined symbol '_D3std1c7windows3com9ComObject14QueryInterfaceMWPS3std1c7windows3com4GUIDPPvZi')

If I omit gdc, and build just c, or just c and c++, then it compiles fine.

That was all with a plain "make" with no paramaters and most, but not all, of the configure paramaters you used in that build script you just posted. I'm trying c/c++/d with all the configure and make paramaters you used right now. We'll see what happens.

A few things I should note:

1. I'm not sure which version of GCC you were patching GDC against, but I'm
building
 against version 4.1.2. To clarify all the versionings: I'm using MSYS/MinGW
GCC 3.5.2
 to build GDC 0.22 (old now, I know) against the GCC 4.1.2 sources.

This did originally create one problem: I would always get the following error no matter what languages (c++,d) were included or excluded:

Makefile:1280: *** target pattern contains no '%'. Stop.

After googling, I found that after running configure, I needed to go into the makefile and change the line:

LD = D:{a big giant mess here}ld.exe

To:

LD = ld

Which solved that problem. (I have no idea why it was happening, but it seems to be a common problem when trying to build GCC 4.x.x on MSYS/MinGW)


2. Not sure if it matters, but configure scripts tell me that someting about symbolic links isn't working. It reports "ln" works, but "ln -s" doesn't. Not sure if that matters. I have only a passing familiarity with symlinks


3. Also not sure if it matters, but at times I've noticed the compiler spitting out the following (I know warnings are normal while building gcc, but these are regarding commandline paramaters rather than the source being compiled):

cc1plus.exe: warning: command line option "-Wstrict-prototypes" is valid for
Ada/C/ObjC but not for C++
cc1plus.exe: warning: command line option "-Wmissing-prototypes" is valid
for Ada/C/ObjC but not for C++