Ok, I figured out what went wrong with the last beta, the new posix makefiles put the libraries in a different directory. This is corrected now.


http://ftp.digitalmars.com/dmd1beta.zip http://ftp.digitalmars.com/dmd2beta.zip

The biguint internal functions are in the wrong place. They should be
in std/internal/math/, instead they're just in std/.
This is probably just a problem in the zip maker.
Applies to biguintcore, biguintnoasm, biguintX86.

Bigint appears in the docs, but it's not in the changelog. There's another fixed bug, as well.

    $(WHATSNEW
	$(LI std.bigint has been completely replaced with a faster implementation.
	   Multiplication is now 5 times faster, division is
	   300 times faster, and squaring is 10 times faster. For large numbers
	   (~5000 words), the speedup is 5 times larger than this.)
	 )

	$(LI $(BUGZILLA 4023): std.math.hypot() returns infinity when either
argument is zero)

---


On 7 April 2010 08:22, Walter Bright <walter at digitalmars.com> wrote:
> Ok, I figured out what went wrong with the last beta, the new posix makefiles put the libraries in a different directory. This is corrected now.
>
>
> http://ftp.digitalmars.com/dmd1beta.zip
> http://ftp.digitalmars.com/dmd2beta.zip
> _______________________________________________
> dmd-beta mailing list
> dmd-beta at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/dmd-beta
>

That reminds me, I didn't check, but wasn't there some work done to AAs?  Is that in there?

BTW, the OS I was using was Linux when trying to build/test phobos.

-Steve




----- Original Message ----
> From: Don Clugston <dclugston at googlemail.com>
> To: Discuss the dmd beta releases for D <dmd-beta at puremagic.com>
> Sent: Wed, April 7, 2010 3:27:26 AM
> Subject: Re: [dmd-beta] dmd 1.058 and 2.043 beta
> 
> The biguint internal functions are in the wrong place. They should be
in
> std/internal/math/, instead they're just in std/.
This is probably just a
> problem in the zip maker.
Applies to biguintcore, biguintnoasm,
> biguintX86.

Bigint appears in the docs, but it's not in the changelog.
> There's
another fixed bug, as well.


> $(WHATSNEW
    $(LI std.bigint has been completely replaced
> with a faster implementation.
       Multiplication is
> now 5 times faster, division is
       300 times faster,
> and squaring is 10 times faster. For large numbers

> (~5000 words), the speedup is 5 times larger than this.)

> )

    $(LI $(BUGZILLA 4023): std.math.hypot() returns
> infinity when either
argument is zero)

---


On 7 April 2010
> 08:22, Walter Bright <
> href="mailto:walter at digitalmars.com">walter at digitalmars.com>
> wrote:
> Ok, I figured out what went wrong with the last beta, the new
> posix
> makefiles put the libraries in a different directory. This is
> corrected now.
>
>
> 
> http://ftp.digitalmars.com/dmd1beta.zip
> 
> http://ftp.digitalmars.com/dmd2beta.zip
> 
> _______________________________________________
> dmd-beta mailing
> list
> 
> href="mailto:dmd-beta at puremagic.com">dmd-beta at puremagic.com
> 
> http://lists.puremagic.com/mailman/listinfo/dmd-beta
>
_______________________________________________
dmd-beta
> mailing list

> href="mailto:dmd-beta at puremagic.com">dmd-beta at puremagic.com

> href="http://lists.puremagic.com/mailman/listinfo/dmd-beta" target=_blank
> >http://lists.puremagic.com/mailman/listinfo/dmd-beta

On 07/04/10 07:22, Walter Bright wrote:
> Ok, I figured out what went wrong with the last beta, the new posix makefiles put the libraries in a different directory. This is corrected now.
>
>
> http://ftp.digitalmars.com/dmd1beta.zip http://ftp.digitalmars.com/dmd2beta.zip

Before the final release of this is there any chance you could include some of the patches in bugzilla, or comment on why they haven't been included? There's 68 issues that aren't enhancements in bugzilla that have patches, I know you've commented on a few of them, it would be nice to have some feedback on them though. I wouldn't expect you to include all of them for obvious reasons, it'd be nice if you could gradually filter through them though.

http://goo.gl/ViGD (Sorry for shortening it, it spanned 5 lines otherwise though)

I know this release is primarily to fix the memory corruption issues, I'd still like to see some of these included though (such as the debug info patches I wrote, otherwise I'll have to keep using my patched version of dmd :s)

> http://goo.gl/ViGD (Sorry for shortening it, it spanned 5 lines otherwise though)

Considering the votes and creation time (resp. dmd version) might also be interesting.

http://tinyurl.com/yenoj24

Don Clugston wrote:
> The biguint internal functions are in the wrong place. They should be
> in std/internal/math/, instead they're just in std/.
> This is probably just a problem in the zip maker.
> Applies to biguintcore, biguintnoasm, biguintX86.
>
> Bigint appears in the docs, but it's not in the changelog. There's another fixed bug, as well.
>
>     $(WHATSNEW
> 	$(LI std.bigint has been completely replaced with a faster implementation.
> 	   Multiplication is now 5 times faster, division is
> 	   300 times faster, and squaring is 10 times faster. For large numbers
> 	   (~5000 words), the speedup is 5 times larger than this.)
> 	 )
>
> 	$(LI $(BUGZILLA 4023): std.math.hypot() returns infinity when either
> argument is zero)
>
> 

Got it fixed now.

Robert Clipsham wrote:
>
> Before the final release of this is there any chance you could include some of the patches in bugzilla, or comment on why they haven't been included? There's 68 issues that aren't enhancements in bugzilla that have patches, I know you've commented on a few of them, it would be nice to have some feedback on them though. I wouldn't expect you to include all of them for obvious reasons, it'd be nice if you could gradually filter through them though.
>
> http://goo.gl/ViGD (Sorry for shortening it, it spanned 5 lines otherwise though)
>
> I know this release is primarily to fix the memory corruption issues, I'd still like to see some of these included though (such as the debug info patches I wrote, otherwise I'll have to keep using my patched version of dmd :s)
>

The reason for this release being in a bit of a hurry is because:

1. Memory corruption problems are a nasty, ugly, soul-sucking problem for anyone using the compiler. They make the language look really bad. Rolling out a fix for them is and should be first priority.

2. I'm going to be gone to the ACCU next week, and so won't be working on the compiler.  I don't think this update can wait.

The dwarf patches you wrote are very important and I wanted to get them in, but there isn't time. The existing patches have already destabilized the build & test enough :-(

Also, I'm spending my time preparing for the ACCU (Andrei and I were invited to do a full day tutorial on D, yay!) which is important as we want to get D's best foot forward. It's a great opportunity for D.

On 07/04/10 19:23, Walter Bright wrote:
> The reason for this release being in a bit of a hurry is because:
>
> 1. Memory corruption problems are a nasty, ugly, soul-sucking problem for anyone using the compiler. They make the language look really bad. Rolling out a fix for them is and should be first priority.

I agree, I'm glad to see it's so important for you too :)

> 2. I'm going to be gone to the ACCU next week, and so won't be working on the compiler. I don't think this update can wait.

Again, I'd have to agree, other non-critical patches can wait, particularly if you have other commitments.

> The dwarf patches you wrote are very important and I wanted to get them in, but there isn't time. The existing patches have already destabilized the build & test enough :-(

This is unfortunate, I guess they can wait though. Could you maybe specify which patches are causing issues so people can look at them and help take some of your work load? There's a lot of people who are willing to help test/tweak patches out there, they're generally in the less vocal minority though.

> Also, I'm spending my time preparing for the ACCU (Andrei and I were invited to do a full day tutorial on D, yay!) which is important as we want to get D's best foot forward. It's a great opportunity for D.

Getting the word out about D is a good idea, particularly with D2 stabilizing and becoming feature frozen. I hope there's video's for some/all of this so those of us who cannot attend don't miss out, I've enjoyed watching previous video's from various D based presentations!

You've mentioned the time constraints are becoming more of an issue for you, which is holding back patches... Maybe it's time to take the next step in the open source development model and give others commit access to the dmd repository? You could of course review all commits, and set strict regulations on what gets in, eg bigger changes/patches must be run by you first, the test suite must pass etc, whatever you deem fit to make sure nothing too radical gets in. An ideal candidate for commit access would of course be Don who has been providing excellent patches for a long time now, and I'm sure he'd be happy to go along with whatever regulations you put in place... Just something for you to consider, particularly if you're going to have less time to work on the compiler here and there :)

David Simcha reports that this beta fixes his corruption issues.

-Steve



----- Original Message ----
> From: Walter Bright <walter at digitalmars.com>
> Ok, I figured out what went wrong with the last beta, the new posix makefiles
> put the libraries in a different directory. This is corrected
> now.

Steve Schveighoffer wrote:
> David Simcha reports that this beta fixes his corruption issues.
>
>
> 

Good. If nothing else crops up, I'll plan on releasing it tonight.