Re: HEADSUP: GCC 3.2.1 update is coming

[Nate Lawson]

On Thu, 10 Oct 2002, Kris Kennaway wrote:
> On Thu, Oct 10, 2002 at 10:32:15PM -0700, David O'Brien wrote:
> > On Thu, Oct 10, 2002 at 08:47:15PM -0700, Kris Kennaway wrote:
> > > Are these suitable for importing as a regression suite?
> > 
> > We could just import the GCC testsuite... it would add 25 MB (checked
> > out, more w/in /home/ncvs/src :-( )
> 
> Hmm, perhaps this should go in a new cvsup collection or somewhere
> else optional-but-accessible.
> 
> Kris

Errm, ports(7) umm portmgr.  Ok, nothing to add here.

-Nate


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message

Re: HEADSUP: GCC 3.2.1 update is coming

[Mike Barcroft]

David O'Brien <[EMAIL PROTECTED]> writes:
> On Thu, Oct 10, 2002 at 08:38:58PM -0400, Alexander Kabaev wrote:
> > Could you please try to compile libmsun with with GCC 3.3 snapshot David
> > just committed and see if that changes anything?
> 
> It doesn't compile on -current.  Mike and the standards guys are suspose
> to undo the breakage that crept into out headers.  For the daring, try to
> build the port and then change the two "u_long" that is complained about
> to "unsigned long".

I've committed this for now.  The `#ifndef _POSIX_SOURCE' conditional
needs to check for other standards, but I'll leave this until I have a
chance to go through the entire header to make it conformant.

This is a new problem because of rev 1.74 of  where a
conditional was updated to check for other standards, but had some
dependencies that I wasn't aware of and world didn't find them for me.
If there are any others I should find them as I go through each
standard header.

Best regards,
Mike Barcroft

To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message

Re: HEADSUP: GCC 3.2.1 update is coming

[Kris Kennaway]

On Thu, Oct 10, 2002 at 10:32:15PM -0700, David O'Brien wrote:
> On Thu, Oct 10, 2002 at 08:47:15PM -0700, Kris Kennaway wrote:
> > Are these suitable for importing as a regression suite?
> 
> We could just import the GCC testsuite... it would add 25 MB (checked
> out, more w/in /home/ncvs/src :-( )

Hmm, perhaps this should go in a new cvsup collection or somewhere
else optional-but-accessible.

Kris






msg44530/pgp0.pgp
Description: PGP signature

Re: HEADSUP: GCC 3.2.1 update is coming

[David O'Brien]

On Thu, Oct 10, 2002 at 08:47:15PM -0700, Kris Kennaway wrote:
> Are these suitable for importing as a regression suite?

We could just import the GCC testsuite... it would add 25 MB (checked
out, more w/in /home/ncvs/src :-( )

To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message

Re: HEADSUP: GCC 3.2.1 update is coming

[Steve Kargl]

On Thu, Oct 10, 2002 at 08:36:19PM -0700, Don Lewis wrote:
> On 10 Oct, Steve Kargl wrote:
> > On Thu, Oct 10, 2002 at 07:10:44PM -0400, Andrew Gallatin wrote:
> >> 
> >> This is a terrible example in that its impossible to tell what
> >> happened, but it sure shows that something is wrong.
> >> 
> >> Is there a floating point regression suite that you can point me at?
> >> That would be more useful for debugging.  
> >> 
> > 
> > ftp://ftp.netlib.org/fp/ucbtest.tgz
> > 
> > This is Kahan and his students floating point testsuite.
> > It normally check for conformance to IEEE-754.
> 
> For something a bit older, see .
> 
> 

Paranoia only checks things like rounding and epsilon.
ucbtest tests corner cases for the standard intrinsics.
ucbtest should be a much better test of libm than
paranoia.

-- 
Steve

To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message

Re: HEADSUP: GCC 3.2.1 update is coming

[Kris Kennaway]

On Thu, Oct 10, 2002 at 04:29:33PM -0700, Steve Kargl wrote:
> On Fri, Oct 11, 2002 at 01:26:47AM +0200, Martin Blapp wrote:
> > 
> > > Is there a floating point regression suite that you can point me at?
> > > That would be more useful for debugging.
> > 
> 
> I forgot another test suite.  See
> http://www.cs.berkeley.edu/~jhauser/arithmetic/TestFloat.html

Are these suitable for importing as a regression suite?

Kris



msg44527/pgp0.pgp
Description: PGP signature

Re: HEADSUP: GCC 3.2.1 update is coming

[Don Lewis]

On 10 Oct, Steve Kargl wrote:
> On Thu, Oct 10, 2002 at 07:10:44PM -0400, Andrew Gallatin wrote:
>> 
>> This is a terrible example in that its impossible to tell what
>> happened, but it sure shows that something is wrong.
>> 
>> Is there a floating point regression suite that you can point me at?
>> That would be more useful for debugging.  
>> 
> 
> ftp://ftp.netlib.org/fp/ucbtest.tgz
> 
> This is Kahan and his students floating point testsuite.
> It normally check for conformance to IEEE-754.

For something a bit older, see .


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message

Re: HEADSUP: GCC 3.2.1 update is coming

[David O'Brien]

On Thu, Oct 10, 2002 at 08:38:58PM -0400, Alexander Kabaev wrote:
> Could you please try to compile libmsun with with GCC 3.3 snapshot David
> just committed and see if that changes anything?

It doesn't compile on -current.  Mike and the standards guys are suspose
to undo the breakage that crept into out headers.  For the daring, try to
build the port and then change the two "u_long" that is complained about
to "unsigned long".

To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message

Re: HEADSUP: GCC 3.2.1 update is coming

[Alexander Kabaev]

On Fri, 11 Oct 2002 01:26:47 +0200 (CEST)
Martin Blapp <[EMAIL PROTECTED]> wrote:

> 
> Hi,
> 
> > Looks like we win some and we loose some.  I just noticed something
> > weird on my P4 desktop using:
> 
> I reported it too soon. I forgot that on remote Desktops the problem
> doesn't happen
> :P

Could you please try to compile libmsun with with GCC 3.3 snapshot David
just committed and see if that changes anything?

-- 
Alexander Kabaev

To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message

Re: HEADSUP: GCC 3.2.1 update is coming

[Steve Kargl]

On Fri, Oct 11, 2002 at 01:26:47AM +0200, Martin Blapp wrote:
> 
> > Is there a floating point regression suite that you can point me at?
> > That would be more useful for debugging.
> 

I forgot another test suite.  See
http://www.cs.berkeley.edu/~jhauser/arithmetic/TestFloat.html

-- 
Steve

To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message

Re: HEADSUP: GCC 3.2.1 update is coming

[Martin Blapp]

Hi,

> Looks like we win some and we loose some.  I just noticed something
> weird on my P4 desktop using:

I reported it too soon. I forgot that on remote Desktops the problem doesn't
happen
:P

So OpenOffice is still broken.

Yes, now mpg123 plays sound, but the output is horribly noisy. I'm at home
now and can hear the sound. From work I just saw that it played something,
since the output was not anymore zero.

> Is there a floating point regression suite that you can point me at?
> That would be more useful for debugging.

Yes please !

Martin


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message

Re: HEADSUP: GCC 3.2.1 update is coming

[Steve Kargl]

On Thu, Oct 10, 2002 at 07:10:44PM -0400, Andrew Gallatin wrote:
> 
> This is a terrible example in that its impossible to tell what
> happened, but it sure shows that something is wrong.
> 
> Is there a floating point regression suite that you can point me at?
> That would be more useful for debugging.  
> 

ftp://ftp.netlib.org/fp/ucbtest.tgz

This is Kahan and his students floating point testsuite.
It normally check for conformance to IEEE-754.

-- 
Steve

To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message

Re: HEADSUP: GCC 3.2.1 update is coming

[Andrew Gallatin]

Martin Blapp writes:
 > 
 > > There might still be a problem with libm and -march=pentium4. I do not
 > > have P4 myself and I cannot reproduce the problem locally.
 > 
 > This problem is also solved. libm compiles again fine with -march=pentium4.
 > OpenOffice and xmms, mpg123 work as they did before.
 > 

Looks like we win some and we loose some.  I just noticed something
weird on my P4 desktop using:

gcc version 3.2.1 [FreeBSD] 20021009 (prerelease)


If I build libm with -march=pentium4, then the mozilla theme "Oribit"
looks corrupt.  If I use -march=pentium3, the problem goes away.  I've
left xwindow dumps on freefall in ~gallatin/mozilla.bad.xwd
~gallatin/mozilla.good.xwd

This is a terrible example in that its impossible to tell what
happened, but it sure shows that something is wrong.

Is there a floating point regression suite that you can point me at?
That would be more useful for debugging.  

Drew

To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message

Re: HEADSUP: GCC 3.2.1 update is coming

[Martin Blapp]

Hi Alexander,

> I am going to import a new GCC 3.2.1 snapshot in about 10 minutes.
> This update fixes the problem with cpp0 dumping core while building a
> kernels many people reported on the list.

Yes, this is fixed. Just tested it with my config :P

> There might still be a problem with libm and -march=pentium4. I do not
> have P4 myself and I cannot reproduce the problem locally.

This problem is also solved. libm compiles again fine with -march=pentium4.
OpenOffice and xmms, mpg123 work as they did before.

Thank you very very much for your work !

Martin

Martin Blapp, <[EMAIL PROTECTED]> <[EMAIL PROTECTED]>
--
ImproWare AG, UNIXSP & ISP, Zurlindenstrasse 29, 4133 Pratteln, CH
Phone: +41 061 826 93 00: +41 61 826 93 01
PGP: 
PGP Fingerprint: B434 53FC C87C FE7B 0A18 B84C 8686 EF22 D300 551E
--


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message

Re: HEADSUP: GCC 3.2.1 update is coming

[Alexander Kabaev]

On Thu, 10 Oct 2002 00:13:04 -0400
Alexander Kabaev <[EMAIL PROTECTED]> wrote:

> I am going to import a new GCC 3.2.1 snapshot in about 10 minutes.
> This update fixes the problem with cpp0 dumping core while building a
> kernels many people reported on the list. 
> 
> Due to popular demand, it was decided to do another import today.
> 
> -- 
> Alexander Kabaev

The import is complete. People who had experienced compiler crashed due
to -march optimizations are encouraged to turn -march on again. Bug
reports are appreciated.

There might still be a problem with libm and -march=pentium4. I do not
have P4 myself and I cannot reproduce the problem locally.
-- 
Alexander Kabaev


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message

Re: HEADSUP: GCC 3.2.1 update is coming

[Robert Watson]

Wonderful.  Your work on the toolchain is much appreciated!

Robert N M Watson FreeBSD Core Team, TrustedBSD Projects
[EMAIL PROTECTED]  Network Associates Laboratories

On Thu, 10 Oct 2002, Alexander Kabaev wrote:

> I am going to import a new GCC 3.2.1 snapshot in about 10 minutes. This
> update fixes the problem with cpp0 dumping core while building a kernels
> many people reported on the list. 
> 
> Due to popular demand, it was decided to do another import today.
> 
> -- 
> Alexander Kabaev
> 
> 
> To Unsubscribe: send mail to [EMAIL PROTECTED]
> with "unsubscribe freebsd-current" in the body of the message
> 


To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message