K-3D

m=E5n 2003-06-23 klockan 10.41 skrev Richard K=F6rber:
> Sorry, but Google couldn't help and you don't have an archive where I=20
> could have looked up...

There is a mailing list archive on the SourceForge project page: Go to
http://k3d.sourceforge.net and click on 'Mailing List' in the one of the
boxes to the left, then click on 'k3d-deveopment'. To search the archive
use the regular SF.net search form in the left corner and choose Mailing
List Archive from the drop down box.

Just noted that the list "is not yet archive" ... that happen from time
to time when SourceForge do maintance and upgrade the system.

> I have compiled and installed the latest k3d sources using gcc3.1. When=20
> I start it, it will show the splash screen, but then it terminates with=20
> this error message:

Exactly what do you meen with "latest sources", latest cvs or latest
stable release? Please provide us with a version number.

:-)

--=20
Anders Dahnielson
<anders@...>

GnuPG 1024D/D424211B 8FDC 38E7 D8E9 67AE 85F4  F2A1 D71D 7FCE D424 211B

"I think the biggest threat to privacy is Moore's law. The=20
human population does not double every eighteen month but=20
the possibility to keep track is."
-- Phil Zimmermann

Richard K=F6rber writes:
 > Hi!
 >=20
 > Sorry, but Google couldn't help and you don't have an archive where =
I=20
 > could have looked up...
 >=20
 > I have compiled and installed the latest k3d sources using gcc3.1. W=
hen=20
 > I start it, it will show the splash screen, but then it terminates w=
ith=20
 > this error message:
 >=20
 > k3dScriptMenuItem: line 1 column 0: Premature end-of-document at dep=
th 0
 >=20
 > SDPGTK-CRITICAL **: file sdpgtkobjectcontainer.cpp: line 415 (bool=20=

 > sdpGtkObjectContainer::Load (sdpxml::Document &)): assertion `XML.Na=
me()=20
 > =3D=3D "gtkml"' failed.
 >=20
 > K3DSDK-CRITICAL **: file k3dscriptmenuitem.cpp: line 124 (bool=20
 > k3dScriptMenuItem::Create (sdpGtkIObjectContainer *, sdpxml::Documen=
t &,=20
 > sdpxml::Element &)): assertion `Load(uitemplate, "k3dScriptMenuItem"=
)'=20
 > failed.
 >=20
 > SDPGTK-ERROR **: file sdpgtkobjectcontainer.cpp: line 177 (void=20
 > sdpGtkObjectContainer::NotifyObject (GtkObject *)): assertion failed=
:=20
 > (Object)
 > aborting...
 >=20
 > What to do? :-)
 >=20

If I remember well, I had the same with a very recent CVS check out (41=
?)
and the splash screen gtml files where missing from the installed dirs.=


Fred

P.S.  Sorry for forgetting to report that before ;-)

Hi!

> Just noted that the list "is not yet archive" ... that happen from time
> to time when SourceForge do maintance and upgrade the system.

So I suppose this was when I was entering the page. :-)

> Exactly what do you meen with "latest sources", latest cvs or latest
> stable release? Please provide us with a version number.

Sorry, my mistake. I meant the source of the latest stable release=20
(0.2.5.4).

My configure line was

   ./configure --with-gcc-2-96 --without-plib

--with-gcc-2-96 was required because I have both gcc 2.96 and 3.1=20
installed on my system. I set CC=3Dgcc3, and it was also properly accepte=
d=20
by configure, but later it complained just because of the existence of=20
gcc 2.96.

--without-plib was required because I don't have plib (well, I even=20
don't know what it is) and configure suggested to use this option.

Regards
--=20
Richard K=F6rber

Richard Körber wrote:

> I have compiled and installed the latest k3d sources using gcc3.1. When I start it, it will show the splash screen, but then it terminates with this error message:
> 
> k3dScriptMenuItem: line 1 column 0: Premature end-of-document at depth 0
> 
> SDPGTK-CRITICAL **: file sdpgtkobjectcontainer.cpp: line 415 (bool sdpGtkObjectContainer::Load (sdpxml::Document &)): assertion `XML.Name() == "gtkml"' failed.
> 
> K3DSDK-CRITICAL **: file k3dscriptmenuitem.cpp: line 124 (bool k3dScriptMenuItem::Create (sdpGtkIObjectContainer *, sdpxml::Document &, sdpxml::Element &)): assertion `Load(uitemplate, "k3dScriptMenuItem")' failed.
> 
> SDPGTK-ERROR **: file sdpgtkobjectcontainer.cpp: line 177 (void sdpGtkObjectContainer::NotifyObject (GtkObject *)): assertion failed: (Object)
> aborting... 

> Sorry, my mistake. I meant the source of the latest stable release 
> (0.2.5.4).

Go to the top-level k3d source directory, and do

	make test

does that work?

Tim

Hi!

> Go to the top-level k3d source directory, and do
>     make test
> does that work?

No... It's the same. The splash screen appears, disappears, and then I 
get an error on the console:

----------------------- 8< ------------------------
PATH=`pwd`/renderjob:`pwd`/renderframe:`pwd`/shaderpreprocessor:$PATH 
application/k3d --basepath `pwd` --libpath `pwd`/modules
k3dScriptMenuItem: line 1 column 0: Premature end-of-document at depth 0

SDPGTK-CRITICAL **: file sdpgtkobjectcontainer.cpp: line 415 (bool 
sdpGtkObjectContainer::Load (sdpxml::Document &)): assertion `XML.Name() 
== "gtkml"' failed.

K3DSDK-CRITICAL **: file k3dscriptmenuitem.cpp: line 124 (bool 
k3dScriptMenuItem::Create (sdpGtkIObjectContainer *, sdpxml::Document &, 
sdpxml::Element &)): assertion `Load(uitemplate, "k3dScriptMenuItem")' 
failed.

SDPGTK-ERROR **: file sdpgtkobjectcontainer.cpp: line 177 (void 
sdpGtkObjectContainer::NotifyObject (GtkObject *)): assertion failed: 
(Object)
aborting...
make: *** [test] Abgebrochen
----------------------- 8< ------------------------

(Abgebrochen = aborted)

Regards

Richard Körber wrote:

> My configure line was
> 
>   ./configure --with-gcc-2-96 --without-plib
> 
> --with-gcc-2-96 was required because I have both gcc 2.96 and 3.1 
> installed on my system. I set CC=gcc3, and it was also properly accepted 
> by configure, but later it complained just because of the existence of 
> gcc 2.96.

Richard:

This didn't catch my eye at first, but is a problem: "CC" is the C 
compiler; overriding it has no effect when compiling a C++ application 
like K-3D.  You have to override CXX instead.

Having said that, you may be in for disappointment - the problem is that 
  binaries (applications and libraries) compiled with gcc 2.96 and 3.1 
are not mutually compatible.  Presumably your system is going to have 
some of each, and the problem is making sure that K-3D and every library 
it loads have all been compiled with gcc 3.1.  This is a hugely 
difficult task unless you know exactly what you're doing.  I assume 
since you have gcc 2.96 that you're working with a RedHat or derivative 
distro, which means that every binary package you've installed has been 
built against gcc 2.96.  My suggestion is that you completely remove gcc 
2.96 and 3.1, and install gcc 3.0.2.  The advantage to this is that 
binaries built with 3.0.2 are compatible with binaries built with 2.96, 
but don't suffer from any of the known issues we've identified in 2.96. 
  This is the configuration I have on the machine where I do most of my 
K-3D development, BTW.

A second alternative would be to get the absolute latest version of gcc 
2.96 and remove all traces of 3.1.  Then, you *might* get a good build 
of K-3D using the --with-gcc-2-96 option.

Cheers,
Tim

Hi!

> Having said that, you may be in for disappointment - the problem is tha=
t=20
>  binaries (applications and libraries) compiled with gcc 2.96 and 3.1=20
> are not mutually compatible. [...]

Thanks for the explanation! Anyhow I don't want to downgrade to gcc 3.0=20
because I need gcc 3.1 to compile mplayer.

I will challenge my luck and give K3d another try by overriding CXX as=20
well. If that fails too, well... Bad luck to me. Frankly, there are=20
moments when I think that Linux sucks a big deal, and I wish I had my=20
Amiga back. ;-)

Regards
--=20
Richard K=F6rber

Hi!

> I will challenge my luck and give K3d another try by overriding CXX as=20
> well. If that fails too, well... Bad luck to me.

Well, as you have expected, ld fails eventually. I give up... ;( But I=20
just wonder why it is not possible to run K3d on a major distribution=20
like RedHat 7.3.

Regards
--=20
Richard K=F6rber

Richard K=F6rber wrote:

> I will challenge my luck and give K3d another try by overriding CXX as=20
> well. If that fails too, well... Bad luck to me. Frankly, there are=20
> moments when I think that Linux sucks a big deal, and I wish I had my=20
> Amiga back. ;-)

Don't we all...?

-- Brett
                                           http://www.chapelperilous.net/
------------------------------------------------------------------------
An exotic journey in downtown Newark is in your future.

Richard K=F6rber wrote:
> Hi!
>=20
>> I will challenge my luck and give K3d another try by overriding CXX as=
=20
>> well. If that fails too, well... Bad luck to me.
>=20
>=20
> Well, as you have expected, ld fails eventually. I give up... ;( But I=20
> just wonder why it is not possible to run K3d on a major distribution=20
> like RedHat 7.3.

Maybe you should switch to Debian... it compiles just fine on Debian Sid=20
using gcc 3.2...

-- Brett
                                           http://www.chapelperilous.net/
------------------------------------------------------------------------
An exotic journey in downtown Newark is in your future.

Hi!

> Maybe you should switch to Debian... it compiles just fine on Debian Si=
d=20
> using gcc 3.2...

Yeah, I will definitely use Debian the next time I set up my computer=20
from scratch.

Regards
--=20
Richard K=F6rber

Brett W. McCoy writes:
 > Richard K=F6rber wrote:
 > > Hi!
 > >=20
 > >> I will challenge my luck and give K3d another try by overriding C=
XX as=20
 > >> well. If that fails too, well... Bad luck to me.
 > >=20
 > >=20
 > > Well, as you have expected, ld fails eventually. I give up... ;( B=
ut I=20
 > > just wonder why it is not possible to run K3d on a major distribut=
ion=20
 > > like RedHat 7.3.
 >=20
 > Maybe you should switch to Debian... it compiles just fine on Debian=
 Sid=20
 > using gcc 3.2...
 >=20

Or upgrade RedHat.  It also compiles just fine under RedHat 8.0 (and
presumably 9.0).  The problem is that gcc as provided in RH7.x is not
terribly good (and apparently was never meant as a release version...).=


Fred

Fred Labrosse wrote:
> 
> Brett W. McCoy writes:
>  > Richard Körber wrote:
>  > > Hi!
>  > >
>  > >> I will challenge my luck and give K3d another try by overriding CXX as
>  > >> well. If that fails too, well... Bad luck to me.
>  > >
>  > >
>  > > Well, as you have expected, ld fails eventually. I give up... ;( But I
>  > > just wonder why it is not possible to run K3d on a major distribution
>  > > like RedHat 7.3.
>  >
>  > Maybe you should switch to Debian... it compiles just fine on Debian Sid
>  > using gcc 3.2...
>  >
> 
> Or upgrade RedHat.  It also compiles just fine under RedHat 8.0 (and
> presumably 9.0).  The problem is that gcc as provided in RH7.x is not
> terribly good (and apparently was never meant as a release version...).

[OT] off topic

The story is, we will still be under version 3.00 of gcc.
The same kind of story that is happening with Xfree86 since january.
We require people with some drive.

Release the thing, we will make it work.  Is that bold enough.

Thing that Red Hat is doing now, do now release unofficial release
before it has receive benediction of the pope.  But the version
number tell it all. ex.: 2.3.4.5.90, means you are very, very slow.
ex.: before gcc 3.3, we were 2.2.96

I find very funny people, complaining about gcc 2.96.
<bold> It was the first gcc doing sensible thing with template</bold>
That is compliance with ISO/IEC 14882, many years after the
official release of the standard.
Where were you, when the avalanche of patches to fix non
compliant application were flying around.

The thing(gcc 2.96) must work, I still have Netscape which require the 
libstdc++ compile with 2.96. and my system is rawhide
where the current version is gcc 3.3

=====

Another point, or another fight.
With a distribution release, never upgrade thing like gcc, binutils.
From an administrative point of view, that is a big no no.
Redhat have made that point very clear.
If new functionality is require, it must be backport.
And of discussion.  Exception for people with lot of money and time,
to prove that every thing work with this new set of tool.