Final report Thursday 08-Nov-01: Question about CGM support in pl otutils-2.4.1

[David Humphrey]

I have finally worked around this problem.  First, I needed to open the
output file in binary mode.  Then, I was having trouble with passing an I/O
handle from a main program into a Windows DLL (libplot linked as a .dll).
This problem is still open, but it's a Windows linking and runtime library
problem, nothing wrong with plotutils.  For the moment, I'm opening the
output file in apinewc.c, a workaround, for sure.  So, I consider this case
closed as far as plotutils goes.

Regards,
David L. Humphrey
Director, Technical Services
Bell Geospace, Inc

>  -----Original Message-----
> From:         David Humphrey  
> Sent: Monday, November 05, 2001 10:27 AM
> To:   'address@hidden'
> Subject:      Status report Monday 05-Nov-01: Question about 
> CGM support in plotutils-2.4.1
> 
> I've spent some time walking through the code in the debugger 
> and now believe that the extra byte '0d' is introduced into 
> the output stream when libplot writes the output cgm file to 
> disk.  Therefore, it must be the way I'm opening the output file.
> 
> It now dawns on me that I left out some important 
> information.  I've taken plotutils-2.4.1 and compiled libplot 
> on Windows NT using Microsoft Visual Studio 6.0.  So, this 
> bug is probably a case of misuse on my part since Windows and 
> UNIX treat files a bit differently.  I'll resolve this 
> problem and submit the fix for your information.
> 
> Regards,
> David L. Humphrey
> Director, Technical Services
> Bell Geospace, Inc
> 
> >  -----Original Message-----
> > From:       David L. Humphrey [mailto:address@hidden 
> > Sent:       Saturday, November 03, 2001 3:02 PM
> > To: address@hidden
> > Subject:    Additional Information: Question about CGM 
> > support in plotutils-2.4.1
> > 
> > I have some additional information on this problem.  I sat 
> > down and decoded the binary CGM file by hand in a effort to 
> > discover the problem.  It seems that there is an error in the 
> > encoding of the first Restricted Text command starting at 
> > byte 0x198.  The 0x0d byte at 0x19a does not belong there.  I 
> > removed it and the Microsoft products read the file properly.
> > 
> > It's odd that the first Restricted Text command is erroneous 
> > while the following two are correct.  I'll peruse the code 
> > Monday in an effort to spot the bug.  If I can find it, I'll 
> > submit a correction.  However, I strongly request that you 
> > look for this one, too.
> > 
> > Regards,
> > David L. Humphrey
> > Director, Technical Services
> > Bell Geospace, Inc
> > 
        > >  -----Original Message-----
        > > From:       David Humphrey  
        > > Sent:       Friday, November 02, 2001 12:10 PM
        > > To: 'address@hidden'
        > > Subject:    Question about CGM support in plotutils-2.4.1
        > > 
        > > Attached are two test files I produced from plotutils-2.4.1.  
        > > The only difference is that I changed the encoding from 
        > > "binary" to "cleartext".  Microsoft Word & Powerpoint will 
        > > display the cleartext version, but report an error for the 
        > > binary one.  At first, I thought the Microsoft products would 
        > > not display binary CGMs at all, but this is not the case as I 
        > > have displayed several intricate ones without problems.
        > > 
        > > Would you please advise me on where the problem resides?  Am 
        > > I misusing plotutils in some way?
        > > 
        > > Regards,
        > > David L. Humphrey
        > > Director, Technical Services
        > > Bell Geospace, Inc
        > > 
        > >  << File: binV1.cgm >>  << File: ctV1.cgm >>