openvrml-develop — OpenVRML developers mailing list

Re: [Openvrml-develop] qtlookat

[bentley <be...@fw...>]

I don't expect the qt plugin will work with 0.9.0 without some
major changes. Particularly, it depended on
VrmlScene::loadFromFunction(...) which went away with the new
parser.


On Thu, 7 Dec 2000, clayton cottingham wrote:

> hey could someone give me some help on how to install this
> 
> id really like to have it as a plugin of course!!
> 
> 
> i read through README.build_from_cvs and tried both methods and got no
> where

---
Ken Bentley  <be...@fw...>

[Openvrml-develop] prob comp'in gtklookat

[clayton c. <cl...@ma...>]

following the READ.build_from_cvs:

built fine but on running the app:

Loading required GL library /usr/X11R6/lib/libGL.so.1.2
readWRL openvrml-0.9.0/models/snoman.wrl

** ERROR **: Can't create GtkGLArea widget

aborting...
Aborted

[Openvrml-develop] qtlookat

[clayton c. <cl...@ma...>]

clayton cottingham wrote:
> 
> 0.9 working on mandrake 7.2!!
> 
> lovely lovely
> 
> thanks to all for such hard work
> 
> should i get started on an rpm ??
> _______________________________________________
> Openvrml-develop mailing list
> Ope...@li...
> http://lists.sourceforge.net/mailman/listinfo/openvrml-develop


hey could someone give me some help on how to install this

id really like to have it as a plugin of course!!


i read through README.build_from_cvs and tried both methods and got no
where

[Openvrml-develop] yay!!

[clayton c. <cl...@ma...>]

0.9 working on mandrake 7.2!!

lovely lovely

thanks to all for such hard work

should i get started on an rpm ??

[Openvrml-develop] ANN: OpenVRML 0.9.0

[Braden M. <br...@en...>]

At long last, OpenVRML (formerly LibVRML97) version 0.9.0 has been
released. The library remains in a transitional phase and there remain
some significant regressions that have yet to be addressed. However, many
significant changes have occured since the 0.8.x series:

 - Name change: the library is now called OpenVRML.

 - Library is now covered by the GNU Lesser General Public License
   (LGPL).

 - New parser, removing the dependency on SGI parser code.
   (Braden McDaniel <br...@en...>)

 - View frustum culling and navigation fixes.
   (Christopher K. St. John <cs...@ac...>,
   S. K. Bose <bo...@pa...>)

 - Can now compile under CygWin.
   (Jean-Daniel Fekete <Jea...@em...>)

 - SphereSensor support.
   (Matt Giger <mg...@lu...>;.)

 - MPEG reader for MovieTexture nodes.
   (Chris Morley <cm...@ve...>;)

 - Updated the JavaScript support to use Mozilla JS 1.5rc2.
   (Braden McDaniel <br...@en...>;)

 - Lots of clean ups and fixes.


The project home page is

  <http://openvrml.sourceforge.net>

The source distribution can be downloaded from

  <http://sourceforge.net/project/showfiles.php?group_id=7151>


-- 
Braden N. McDaniel                              e-mail: br...@en...
<http://endoframe.com>                          Jabber: br...@ja...

Re: [Openvrml-develop] Java SAI support patch uploaded

[Jeff D. <ig...@po...>]

On Tue, Dec 05, 2000 at 11:55:21PM -0500, Tom Flynn wrote:
> This is great!!!

Thanks...  Hopefully, once it's in the CVS repository, people can
clean it up a bit, and complete the bits that I didn't get around to
writing (I'm gonna have to write the Browser node for my project after
all, mind you...)

> Haven't had a chance to play with it yet but probably will this weekend.
> Just wondering how hard you think it would be to add support for JDK 1.3 as
> well?

Hmm...  I haven't looked at the JNI changes between 1.1 and 1.3...  I
imagine it ought to work almost out-of-the-box.

I just installed 1.3, so I'll try to make this work, too.  Perhaps tomorrow.

-jeff

Re: [Openvrml-develop] Java SAI support patch uploaded

[Tom F. <tf...@us...>]

This is great!!!

Haven't had a chance to play with it yet but probably will this weekend.
Just wondering how hard you think it would be to add support for JDK 1.3 as
well?

Thanks,
Tom

----- Original Message -----
From: "Jeff Dubrule" <ig...@po...>
To: "OpenVRML Hackers" <ope...@li...>
Sent: Monday, December 04, 2000 1:37 PM
Subject: [Openvrml-develop] Java SAI support patch uploaded


> This patch adds preliminary, yet functional support for Java Script
> nodes. Currently, it only plays nicely with JDK 1.1, although it
> shouldn't be too hard to make Kaffe work (I started out using Kaffe,
> but Sun's JDK turned out to be easier to debug).
>
> To use this, you *must* set
> LD_LIBRARY_PATH=/usr/lib/jdk1.1/libs/ia32/native_threads
>
> You must also set
> CLASSPATH=/the/libvrml97/source/libvrml97java/java-src
>
> This patch handles most, but not all, of the SFFoo (in Java) => SFFoo
> (in C++) conversions. It'll blow up in a fairly intelligent manner if
> gets confused, and finishing them off, while tedious, is not hard.
>
> The Browser node is totally unimplemented, as are a number of
> functions in the other mostly-implemented nodes.
>
> I seriously recommand putting this on a branch, not the main trunk
> (also, I wouldn't mind CVS write access so I can debug as I go)
>
> This is the product of 3 weeks of work, where my goal was to make
> GeoVRML (www.geovrml.org) fly.
>
> Thanks go out to SRI International, who paid me to write this, then
> let me give it back to you.
>
> Share and enjoy
> -jeff
> _______________________________________________
> Openvrml-develop mailing list
> Ope...@li...
> http://lists.sourceforge.net/mailman/listinfo/openvrml-develop
>

[Openvrml-develop] Java SAI support patch uploaded

[Jeff D. <ig...@po...>]

This patch adds preliminary, yet functional support for Java Script
nodes. Currently, it only plays nicely with JDK 1.1, although it
shouldn't be too hard to make Kaffe work (I started out using Kaffe,
but Sun's JDK turned out to be easier to debug).

To use this, you *must* set
LD_LIBRARY_PATH=/usr/lib/jdk1.1/libs/ia32/native_threads

You must also set
CLASSPATH=/the/libvrml97/source/libvrml97java/java-src

This patch handles most, but not all, of the SFFoo (in Java) => SFFoo
(in C++) conversions. It'll blow up in a fairly intelligent manner if
gets confused, and finishing them off, while tedious, is not hard.

The Browser node is totally unimplemented, as are a number of
functions in the other mostly-implemented nodes.

I seriously recommand putting this on a branch, not the main trunk
(also, I wouldn't mind CVS write access so I can debug as I go)

This is the product of 3 weeks of work, where my goal was to make
GeoVRML (www.geovrml.org) fly.

Thanks go out to SRI International, who paid me to write this, then
let me give it back to you.

Share and enjoy
-jeff

Re: [Openvrml-develop] Trouble building Win32Lookat -- please help!

[S.K.Bose <bo...@pa...>]

On Wed, 29 Nov 2000, Joseph Sullivan wrote:

> I don't recall the version number of Lookat, but I downloaded it from
> www.openvrml.org.  Is this the fixed one?
> 

No, this is not the fixed one. Fixed on is on
http://sourceforge.net/projects/openvrml. I have given the mirror copy of
current CVS (by deleting platform specific compilation/installation for
Mac and Unixen) to Braden in zip. In next release there will be a single
distribution for windows, Mac and Unixen.

Thanks,

Bose

Re: [Openvrml-develop] cannot compile due to antlr.g ??

[S.K.Bose <bo...@pa...>]

On Sun, 26 Nov 2000, Rodrigo Brito wrote:

> Hi... I've taken the newest version of libVRML97 from CVS, and i can't compile libvrml97core
> the message from MSVC 6 is:
> Performing Custom Build Step on .\antlr.g
>         1 ficheiro(s) copiado(s)
> Sem espaço de ambiente
> Comando ou nome de ficheiro nao válido
> Compiling...
> Vrml97Parser.cpp
> fatal error C1083: Cannot open source file: 'C:\temp\Nova pasta\libvrml97\libvrml97core\win32\Vrml97Parser.cpp': No such file or directory
> Error executing cl.exe.
> 
> (The Portuguese is:
> 
> Performing Custom Build Step on .\antlr.g        
>         1 file(s) copied
> Without environment space
> Command or filename not valid
> compiling...
> )
> 
> 
> It seems that i can't really find antlr.g in my computer, although vrml97parser exists in the same directory as all other libvrml97core source files, but somehow is not loaded.
> Even when i pick the file up and subsitute it for the unexistent file, it still gives me compiling errors...
> 
> 
> 

Did you give proper path for antlr-2.7.0 and java ? You have to set
it by means of setting antlr.g by performing Custom Build. By default 
"set classpath = c:\user\antlr-2.7.0" and java is
"c:\user\jdk1.2.2\bin\java".
Use "makefile" for avoiding all such problems. Please go through "Install"
file for details.

Bose

[Openvrml-develop] cannot compile due to antlr.g ??

[Rodrigo B. <rod...@cl...>]

Hi... I've taken the newest version of libVRML97 from CVS, and i can't =
compile libvrml97core
the message from MSVC 6 is:
Performing Custom Build Step on .\antlr.g
        1 ficheiro(s) copiado(s)
Sem espa=E7o de ambiente
Comando ou nome de ficheiro nao v=E1lido
Compiling...
Vrml97Parser.cpp
fatal error C1083: Cannot open source file: 'C:\temp\Nova =
pasta\libvrml97\libvrml97core\win32\Vrml97Parser.cpp': No such file or =
directory
Error executing cl.exe.

(The Portuguese is:

Performing Custom Build Step on .\antlr.g       =20
        1 file(s) copied
Without environment space
Command or filename not valid
compiling...
)


It seems that i can't really find antlr.g in my computer, although =
vrml97parser exists in the same directory as all other libvrml97core =
source files, but somehow is not loaded.
Even when i pick the file up and subsitute it for the unexistent file, =
it still gives me compiling errors...

Re: [Openvrml-develop] Never mind...

[Braden M. <br...@en...>]

On Tue, 21 Nov 2000, Braden McDaniel wrote:

> On Tue, 21 Nov 2000, Jeff Dubrule wrote:
>
> > This patch'll fix the list, and let fieldTypeName work for all types.
>
> Cool. This looks like something I must have botched a while back. Thanks
> for finding it before the 0.9.0 release.
>
> In the future, though, please use the patch manager:
>
>   <https://sourceforge.net/patch/?group_id=7151>
>
> We have the facility, and it's convenient to keep a history of
> submissions.

I've fixed the problem. I ended up making more substantial changes to the
code:

<http://cvs.sourceforge.net/cgi-bin/cvsweb.cgi/libvrml97/libvrml97core/src/vrml97/VrmlField.cpp.diff?cvsroot=openvrml&r1=text&tr1=1.8&r2=text&tr2=1.9&f=h>

-- 
Braden N. McDaniel                              e-mail: br...@en...
<http://endoframe.com>                          Jabber: br...@ja...

Re: [Openvrml-develop] More on compiling libjs with gcc 2.96

[Braden M. <br...@en...>]

On Wed, 15 Nov 2000, Braden McDaniel wrote:

> It looks like a bug in the SpiderMonkey code was triggering the compiler
> failure after all:
>
>   <http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=19211>
>   <http://bugzilla.mozilla.org/show_bug.cgi?id=59963>
>
> The good news is that the bug has been fixed. The bad news is that the fix
> lives only in CVS.
>
> I'll be updating us from RC1 to RC2 of JS 1.5 before the next release. In
> the absense of another JS release, I'll try to backport the patch to
> Mozilla bug 24892 as well.

Just to note, this issue has been resolved:

<http://sourceforge.net/bugs/?func=detailbug&bug_id=122654&group_id=7151>

-- 
Braden N. McDaniel                              e-mail: br...@en...
<http://endoframe.com>                          Jabber: br...@ja...

Re: [Openvrml-develop] Never mind...

[Braden M. <br...@en...>]

On Tue, 21 Nov 2000, Jeff Dubrule wrote:

> This patch'll fix the list, and let fieldTypeName work for all types.

Cool. This looks like something I must have botched a while back. Thanks
for finding it before the 0.9.0 release.

In the future, though, please use the patch manager:

  <https://sourceforge.net/patch/?group_id=7151>

We have the facility, and it's convenient to keep a history of
submissions.

-- 
Braden N. McDaniel                              e-mail: br...@en...
<http://endoframe.com>                          Jabber: br...@ja...

Re: [Openvrml-develop] Re: trouble building Win32lookat.exe -- please help!

[S.K.Bose <bo...@pa...>]

On Tue, 21 Nov 2000, Joseph Sullivan wrote:

> S.K:
> 
>   Thank you for your reply!
> 
> I tried building in release mode.  In release, the file builds with the following warnings:
> ...locally defined symbol "__vsprintf" imported
> ...locally defined symbol "__setmbcp" imported
> 
>    But when I try to execute win32lookat.exe, it doesn't work.  The app cranks up OK, but when I try to open a .wrl file (say, xyz.wrl) I get the following error dialog boxes:
> 
> * Error near line 3: parse error 
> * Error: couldn't read c:\My Documents\xyz.wrl
> 
> I get the same error dialogs on all .wrl files I try to read.
> 
> 
> Any thoughts?  
> 

Which version are you using ? I hope you are using the old one at
<http://www.openvrml.org>. You download the recent one from
http://sourceforge.net/projects/openvrml. This problem was there earlier.
We are fixing all problems at sourceforge. I am going to put new zip
version this weekend for easy download. Otherwise you can download through
wincvs (windows version).

Thanx.

S. K. Bose

[Openvrml-develop] Re: trouble building Win32lookat.exe -- please help!

[Joseph S. <jo...@wi...>]

S.K:

  Thank you for your reply!

I tried building in release mode.  In release, the file builds with the =
following warnings:
...locally defined symbol "__vsprintf" imported
...locally defined symbol "__setmbcp" imported

   But when I try to execute win32lookat.exe, it doesn't work.  The app =
cranks up OK, but when I try to open a .wrl file (say, xyz.wrl) I get =
the following error dialog boxes:

* Error near line 3: parse error=20
* Error: couldn't read c:\My Documents\xyz.wrl

I get the same error dialogs on all .wrl files I try to read.


Any thoughts? =20

Joseph S.

[Openvrml-develop] Never mind...

[Jeff D. <ig...@po...>]

I figured out why it didn't work properly for field types beyond MFNode:

In the ftn array, which contains the field type names, it goes:
static char *ftn[] = {
  "SFBool",
  "SFColor",
  "SFFloat",
  "SFImage",
  "SFInt32",
  "SFNode",
  "SFRotation",
  "SFString",
  "SFTime",
  "SFVec2f",
  "SFVec3f",
  "MFColor",
  "MFFloat",
  "MFInt32",
  "MFNode"  <--- Note, no comma
  "MFRotation",
  "MFString",
  "MFTime",
  "MFVec2f",
  "MFVec3f",
};

Therefore fieldTypeName(new VrmlMFNode) => "MFNodeMVRotation"

This patch'll fix the list, and let fieldTypeName work for all types.
-jeff

Index: VrmlField.cpp
===================================================================
RCS file: /cvsroot/openvrml/libvrml97/libvrml97core/src/vrml97/VrmlField.cpp,v
retrieving revision 1.7
diff -u -r1.7 VrmlField.cpp
--- VrmlField.cpp       2000/09/10 00:45:48     1.7
+++ VrmlField.cpp       2000/11/21 18:17:34
@@ -86,7 +86,7 @@
   "MFColor",
   "MFFloat",
   "MFInt32",
-  "MFNode"
+  "MFNode",
   "MFRotation",
   "MFString",
   "MFTime",
@@ -99,7 +99,7 @@
 char const * VrmlField::fieldTypeName() const
 {
   int ft = (int) this->fieldType();
-  if (ft > 0 && ft <= (int) VrmlField::MFNODE)
+  if (ft > 0 && ft <= (int) VrmlField::MFVEC3F)
     return ftn[ft-1];
   return "<invalid field type>";
 }

[Openvrml-develop] VrmlField::fieldTypeName

[Jeff D. <ig...@po...>]

This function currently looks like this:

char const * VrmlField::fieldTypeName() const
{
  int ft = (int) this->fieldType();
  if (ft > 0 && ft <= (int) VrmlField::MFNODE)
    return ftn[ft-1];
  return "<invalid field type>";
}

Is there any reason why the cut off is 'MFNODE'?  This means that
MFRotation, MFString, MFTime, MFVec2f, and MFVec3f are considered to
be invalid types...

-jeff

Re: [Openvrml-develop] Trouble building Win32Lookat -- please help!

[S.K.Bose <bo...@pa...>]

On Fri, 17 Nov 2000, WilTech wrote:

> Hello folks,
>   I downloaded 0.8.1 Win32Lookat.  I compiled LibVRMLgl, core, and js, as well as libjpeg ,zlib, and png.  GL came with my copy of VC++ 6.0, but I added GLUT.
> 
> Now the problem:  when I try to compile the debug version of Lookat, I get LNK 2005 ("already defined") and LNK2001 ("unresolved external") errors!   Can anyone give me any assistance here?  The following is a copy of the messages:
> 
> Linking...
> libci.lib(ostream.obj) : error LNK2005: "public: class ostream & __thiscall ostream::operator<<(char const *)" (??6ostream@@QAEAAV0@PBD@Z) already defined in msvcirtd.lib(MSVCIRTD.dll)
> libci.lib(ostream.obj) : error LNK2005: "public: class ostream & __thiscall ostream::flush(void)" (?flush@ostream@@QAEAAV1@XZ) already defined in msvcirtd.lib(MSVCIRTD.dll)
> libci.lib(_ios.obj) : error LNK2005: "public: virtual __thiscall ios::~ios(void)" (??1ios@@UAE@XZ) already defined in msvcirtd.lib(MSVCIRTD.dll)

In debug mode such problem exist for making lookat.exe but it will make
all libraries (libvrml97core.lib and libvrml97gl.lib). No such problem
exist in release mode. For getting rid of this problem in debug mode, I
can suggest to put comment in all I/O statements in lookat.cpp

> libvrml97core.lib(VrmlScene.obj) : error LNK2001: unresolved external symbol _errno
> libvrml97core.lib(VrmlNodeInline.obj) : error LNK2001: unresolved external symbol _errno
> zlib.lib(gzio.obj) : error LNK2001: unresolved external symbol _errno
> libci.lib(filebuf.obj) : error LNK2001: unresolved external symbol ___pioinfo
> Debug/win32lookat.exe : fatal error LNK1120: 2 unresolved externals
> Error executing link.exe.

Please go through the portion "How to build third party libraries?" in
"Install", where it is discussed (specifically zlib.lib). 
> 
> win32lookat.exe - 8 error(s), 0 warning(s)

I hope if you make the zlib.lib again, then there will not be any problem
to make win32lookat.exe in both mode (release or debug)

Thanks,

S. K. Bose

[Openvrml-develop] Trouble building Win32Lookat -- please help!

[WilTech <in...@wi...>]

Hello folks,
  I downloaded 0.8.1 Win32Lookat.  I compiled LibVRMLgl, core, and js, =
as well as libjpeg ,zlib, and png.  GL came with my copy of VC++ 6.0, =
but I added GLUT.

Now the problem:  when I try to compile the debug version of Lookat, I =
get LNK 2005 ("already defined") and LNK2001 ("unresolved external") =
errors!   Can anyone give me any assistance here?  The following is a =
copy of the messages:

Linking...
libci.lib(ostream.obj) : error LNK2005: "public: class ostream & =
__thiscall ostream::operator<<(char const *)" =
(??6ostream@@QAEAAV0@PBD@Z) already defined in =
msvcirtd.lib(MSVCIRTD.dll)
libci.lib(ostream.obj) : error LNK2005: "public: class ostream & =
__thiscall ostream::flush(void)" (?flush@ostream@@QAEAAV1@XZ) already =
defined in msvcirtd.lib(MSVCIRTD.dll)
libci.lib(_ios.obj) : error LNK2005: "public: virtual __thiscall =
ios::~ios(void)" (??1ios@@UAE@XZ) already defined in =
msvcirtd.lib(MSVCIRTD.dll)
libvrml97core.lib(VrmlScene.obj) : error LNK2001: unresolved external =
symbol _errno
libvrml97core.lib(VrmlNodeInline.obj) : error LNK2001: unresolved =
external symbol _errno
zlib.lib(gzio.obj) : error LNK2001: unresolved external symbol _errno
libci.lib(filebuf.obj) : error LNK2001: unresolved external symbol =
___pioinfo
Debug/win32lookat.exe : fatal error LNK1120: 2 unresolved externals
Error executing link.exe.

win32lookat.exe - 8 error(s), 0 warning(s)

Thanks much for any help!

  Joseph Sullivan

Re: [Openvrml-develop] Segfault on rotation_toy.wrl

[Braden M. <br...@en...>]

On Thu, 16 Nov 2000, Braden McDaniel wrote:

> We are currently segfaulting on rotation_toy.wrl. This is a model I tested
> with heavily when I was doing the initial integration work with the new
> parser, so I'm sure this has started sometime since then. Can anyone help
> me draw a firmer bead on it? Unless someone already has a good idea of
> what's going on here and wants to take it, I'll probably tackle this one
> over the weekend.

Nevermind. I already fixed it. :-)

-- 
Braden N. McDaniel                              e-mail: br...@en...
<http://endoframe.com>                          Jabber: br...@ja...

[Openvrml-develop] Segfault on rotation_toy.wrl

[Braden M. <br...@en...>]

We are currently segfaulting on rotation_toy.wrl. This is a model I tested
with heavily when I was doing the initial integration work with the new
parser, so I'm sure this has started sometime since then. Can anyone help
me draw a firmer bead on it? Unless someone already has a good idea of
what's going on here and wants to take it, I'll probably tackle this one
over the weekend.

-- 
Braden N. McDaniel                              e-mail: br...@en...
<http://endoframe.com>                          Jabber: br...@ja...

Re: [Openvrml-develop] Doing the Right Thing with libantlr

[Braden M. <br...@en...>]

On Sat, 4 Nov 2000, Braden McDaniel wrote:

> [Note: None of what I'm about to say applies to any changes before
> 0.9.0. But I wanted to go ahead and ask about this stuff while it's fresh
> on my mind.]
>
> Right now, we export a bunch of symbols and install a bunch of header
> files associated with libantlr. Since it's conceivable that The Real
> libantlr could coexist on a system with our library, I think this is
> Bad Form.
>
> I think the need to do all of this could be eliminated if there's
> agreement that we don't need to be making the parser public. If we don't
> need *those* symbols/headers public, I don't think we need any of the
> ANTLR stuff public either.

On second thought, going ahead with this for 0.9.0 seems like a good idea.
Unless there are objections, I'm going to at least tweak the build scripts
to stop installing those headers. I'll probably put off cleaning up any
symbols we might be inadvertently leaking publicly.

-- 
Braden N. McDaniel                              e-mail: br...@en...
<http://endoframe.com>                          Jabber: br...@ja...

[Openvrml-develop] More on compiling libjs with gcc 2.96

[Braden M. <br...@en...>]

It looks like a bug in the SpiderMonkey code was triggering the compiler
failure after all:

  <http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=19211>
  <http://bugzilla.mozilla.org/show_bug.cgi?id=59963>

The good news is that the bug has been fixed. The bad news is that the fix
lives only in CVS.

I'll be updating us from RC1 to RC2 of JS 1.5 before the next release. In
the absense of another JS release, I'll try to backport the patch to
Mozilla bug 24892 as well.

Now that Netscape 6 has been released and several major Linux
distributions include Mozilla, I'd like to try to stop incorporating our
own copy of libjs post 0.9.0.

-- 
Braden N. McDaniel                              e-mail: br...@en...
<http://www.endoframe.com>                      Jabber: br...@ja...

[Openvrml-develop] ANN: OpenVRML 0.9.0pre2

[Braden M. <br...@en...>]

OpenVRML 0.9.0pre2 has been released.

This release includes a number of miscellaneous bug fixes since
0.9.0pre1. Notably, the normal build process nolonger requires ANTLR.

-- 
Braden N. McDaniel                              e-mail: br...@en...
<http://www.endoframe.com>                      Jabber: br...@ja...