Re: ERROR #22 texdef2d: ERR_BADINDEX

New Message Reply Date view Thread view Subject view Author view

Rob Jenkins (robj++at++barney.reading.sgi.com)
Thu, 30 Nov 1995 17:15:08 +0000

On Nov 30, 8:35am, Robin Rowe wrote:
> Subject: Re: ERROR #22 texdef2d: ERR_BADINDEX
> Rob,
>
> << There were a couple of known problems running on a non-Reality Engine
> platform using TX_FAST_DEFINE or TX_INTERNAL_FORMAT,TX_RGBA_4 but I
> thought they were fixed for Irix 5.3. What machine are you using...>>
>
> This problem manifests on an Indigo. On a Onyx RE I don't get the error
> message, but still have problems. All my compilers and libraries are
> current releases.
>
> <<...can you reproduce with a small piece of example code ? >>
>
> I wish. It's about 200 kloc of code, mixed C++, ANSI C, K&R C, and
> Fortran. The Fortran is -align16. I suspect that the executable is not
> being made properly (bad C prototype or alignment). Visually, I have four
> wide ragged horizontal bands of rogue haze. After making some (apparently)
> unrelated changes the haze went away, but some other changes and now it is
> back again.
>
Please run through gldebug < your app > and break on GL errors - you may get
some errors that don't normally show up, if you use the -C option you should be
able to make some C code with the relevant GL calls in, this may help make an
example if you know roughly where/when in your app the texdef2d happens.

> If I crawl around in the program using dbx I reach a call to gversion(), the
> code then magically jumps to the line of code three lines above, next it
> executes the gversion() line again, then hurls. No core file is
> generated. On an Onyx RE it doesn't crash. It bus errors just on the
> Indigo.
>
Are you compiling with -g ?

> I don't expect you to debug this for me, just wondered if the unusual
> error message at the console is a clue.
>
> One more question, what libraries am I supposed to have when I build
> using -align? Do I need to call handle_unaligned_traps()? If so, what is
> its include file?
>
from man f77:
     The following three options when used at compile time generate various
     degrees of misaligned data in common blocks, and the code to deal with
     the misalignment. You must include these options to f77 in the
     compilation of all modules that reference or define common blocks with
     misaligned data. Failure to do so could cause core dumps (if the trap
     handler is not used), or mismatched common blocks.

     To load the system libraries capable of handling misaligned data, use the
     -L/usr/lib/align switch at load time. The trap handler may be needed to
     handle misaligned data passed to system libraries not included in the
     /usr/lib/align directory (see fixade(3f) and unaligned(3x)).

so you do need to handle any align errors.
What happens if you don't use -align ? NOTE from man fixade ( check it out if
you haven't already ):

         NOTE: the use of this trap handler is intended for diagnostic
         purposes only. Program efficiency may be severely impacted by
         its use.

> Thanks.
>
> Robin
>-- End of excerpt from Robin Rowe

-- 
________________________________________________________________
Rob Jenkins, Software Support Group, Silicon Graphics UK Ltd.       
Forum 1, Station Road, Theale, Reading, UK, RG7 4SB. 
tel 01734 257736, fax 01734 257553, email robj++at++reading.sgi.com,

New Message Reply Date view Thread view Subject view Author view

This archive was generated by hypermail 2.0b2 on Mon Aug 10 1998 - 17:52:04 PDT

This message has been cleansed for anti-spam protection. Replace '++at++' in any mail addresses with the '@' symbol.