I've gotten Phobos1 to at least pass the semantic phase of compiling with -m64. It would be nice if one of you would look at the checkin changes, and fold them correspondingly into Phobos2/Druntime. http://www.dsource.org/projects/phobos/changeset/1907

You'll need the latest dmd2 executable to build them, though don't expect them to get farther than the semantic pass.

Also, the latest zlib apparently supports 64 bit filesystems. The existing one does not. It would also be nice if someone would update Phobos' version. http://www.zlib.net/

Thanks!

Taking base64.d

Andrei

On 08/21/2010 03:18 PM, Walter Bright wrote:
> I've gotten Phobos1 to at least pass the semantic phase of compiling with -m64. It would be nice if one of you would look at the checkin changes, and fold them correspondingly into Phobos2/Druntime. http://www.dsource.org/projects/phobos/changeset/1907
>
> You'll need the latest dmd2 executable to build them, though don't expect them to get farther than the semantic pass.
>
> Also, the latest zlib apparently supports 64 bit filesystems. The existing one does not. It would also be nice if someone would update Phobos' version. http://www.zlib.net/
>
> Thanks!
> _______________________________________________
> phobos mailing list
> phobos at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/phobos

In fact I need to take a lot more than one module as compiling one entails compiling all of its dependents.

That's how e.g. I got to std.math, which has errors pointing to druntime...

Andrei

On 08/21/2010 03:18 PM, Walter Bright wrote:
> I've gotten Phobos1 to at least pass the semantic phase of compiling with -m64. It would be nice if one of you would look at the checkin changes, and fold them correspondingly into Phobos2/Druntime. http://www.dsource.org/projects/phobos/changeset/1907
>
> You'll need the latest dmd2 executable to build them, though don't expect them to get farther than the semantic pass.
>
> Also, the latest zlib apparently supports 64 bit filesystems. The existing one does not. It would also be nice if someone would update Phobos' version. http://www.zlib.net/
>
> Thanks!
> _______________________________________________
> phobos mailing list
> phobos at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/phobos

I'm out of town on family business, but I'll try to take care of some druntime changes.

Sent from my iPhone

On Aug 21, 2010, at 4:18 PM, Walter Bright <walter at digitalmars.com> wrote:

> I've gotten Phobos1 to at least pass the semantic phase of compiling with -m64. It would be nice if one of you would look at the checkin changes, and fold them correspondingly into Phobos2/Druntime. http://www.dsource.org/projects/phobos/changeset/1907
> 
> You'll need the latest dmd2 executable to build them, though don't expect them to get farther than the semantic pass.
> 
> Also, the latest zlib apparently supports 64 bit filesystems. The existing one does not. It would also be nice if someone would update Phobos' version. http://www.zlib.net/
> 
> Thanks!
> _______________________________________________
> phobos mailing list
> phobos at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/phobos

Yes, but the back end isn't started until all the modules on the command line pass semantic processing.

Andrei Alexandrescu wrote:
> In fact I need to take a lot more than one module as compiling one entails compiling all of its dependents.
>
> That's how e.g. I got to std.math, which has errors pointing to druntime...
>
> Andrei
>
> On 08/21/2010 03:18 PM, Walter Bright wrote:
>> I've gotten Phobos1 to at least pass the semantic phase of compiling with -m64. It would be nice if one of you would look at the checkin changes, and fold them correspondingly into Phobos2/Druntime. http://www.dsource.org/projects/phobos/changeset/1907
>>
>> You'll need the latest dmd2 executable to build them, though don't expect them to get farther than the semantic pass.
>>
>> Also, the latest zlib apparently supports 64 bit filesystems. The existing one does not. It would also be nice if someone would update Phobos' version. http://www.zlib.net/
>>
>> Thanks!
>> _______________________________________________
>> phobos mailing list
>> phobos at puremagic.com
>> http://lists.puremagic.com/mailman/listinfo/phobos
> _______________________________________________
> phobos mailing list
> phobos at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/phobos
>
>

Whoa, I changed like a million places only to compile base64, until I hit a brickwall:

mismatch between dmd and object.d or object.di found. Check installation and import paths with -v compiler switch.

I then changed the model back to 32 bits and made sure the unittests work. See the recent commit.

Andrei


On 08/21/2010 03:18 PM, Walter Bright wrote:
> I've gotten Phobos1 to at least pass the semantic phase of compiling with -m64. It would be nice if one of you would look at the checkin changes, and fold them correspondingly into Phobos2/Druntime. http://www.dsource.org/projects/phobos/changeset/1907
>
> You'll need the latest dmd2 executable to build them, though don't expect them to get farther than the semantic pass.
>
> Also, the latest zlib apparently supports 64 bit filesystems. The existing one does not. It would also be nice if someone would update Phobos' version. http://www.zlib.net/
>
> Thanks!
> _______________________________________________
> phobos mailing list
> phobos at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/phobos

Whoa, I changed like a million places only to compile base64, until I hit a brickwall:

mismatch between dmd and object.d or object.di found. Check installation and import paths with -v compiler switch.

I then changed the model back to 32 bits and made sure the unittests work. See the recent commit.

Andrei


On 08/21/2010 03:18 PM, Walter Bright wrote:
> I've gotten Phobos1 to at least pass the semantic phase of compiling with -m64. It would be nice if one of you would look at the checkin changes, and fold them correspondingly into Phobos2/Druntime. http://www.dsource.org/projects/phobos/changeset/1907
>
> You'll need the latest dmd2 executable to build them, though don't expect them to get farther than the semantic pass.
>
> Also, the latest zlib apparently supports 64 bit filesystems. The existing one does not. It would also be nice if someone would update Phobos' version. http://www.zlib.net/
>
> Thanks!
> _______________________________________________
> phobos mailing list
> phobos at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/phobos

Yeah, that's because the CLASSINFO_SIZE changes for 64 bit, and I haven't accounted for that in D2 yet. I'll see about a fix.

Andrei Alexandrescu wrote:
> Whoa, I changed like a million places only to compile base64, until I hit a brickwall:
>
> mismatch between dmd and object.d or object.di found. Check installation and import paths with -v compiler switch.
>
> I then changed the model back to 32 bits and made sure the unittests work. See the recent commit.
>
> Andrei
>

  What's the status of this?  Is the thing that's preventing progress
still the compiler, or should Phobos people start hacking this again?

On 8/23/2010 5:08 PM, Walter Bright wrote:
> Yeah, that's because the CLASSINFO_SIZE changes for 64 bit, and I haven't accounted for that in D2 yet. I'll see about a fix.
>
> Andrei Alexandrescu wrote:
>> Whoa, I changed like a million places only to compile base64, until I hit a brickwall:
>>
>> mismatch between dmd and object.d or object.di found. Check installation and import paths with -v compiler switch.
>>
>> I then changed the model back to 32 bits and made sure the unittests work. See the recent commit.
>>
>> Andrei
>>
> _______________________________________________
> phobos mailing list
> phobos at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/phobos
>

David Simcha wrote:
>  What's the status of this?  Is the thing that's preventing progress
> still the compiler, or should Phobos people start hacking this again?
>

Things got shelved while I worked on the linker problem. But that's solved now, so probably back to the 64 bit work.