From: Han on
The extable library produces slightly different code depending on
whether it is compiled on the calculator or on a computer. The problem
lies in the fact that the table (hptab.hp) is saved as an HP48 file
(with the HPHP48-X header). When downloaded onto the HP50G, the header
remains there so that INCLOB treats the table as if it were a string
object, stripping only the string and length. The source accounts for
this with an offset of 16 (8 characters in the header, 2 nibbles
each).

When compiled with, say Debug4x, however, INCLOB strips the header, so
that the object is included (verbatim). Thus, no offset is required.

Perhaps table creation program that is included with extable could be
updated to create both an HP48 _and_ an HP49 file (with their
respective headers) so that the source files could be kept consistent
across compilers (be it on a PC or on the calculator). Is that library
package still being maintained at all?
From: John H Meyers on
On 7/2/2010 6:25 PM, Han wrote:

> the table (hptab.hp) is saved as an HP48 file
> (with the HPHP48-X header). When downloaded onto the HP50G,
> the header remains there

Change "48" to "49" (with a hex editor)
to make an HP49/50 series binary object header
(caution: do this only for compatible objects, such as strings and grobs)

A free hex editor for Windows:
http://www.softcircuits.com/cygnus/fe/

An alternative, to run on the calculator,
after transferring an object having wrong binary header
(operates on a string beginning with "HPHP4x-y" on the stack):

FIXOB 1.1 (by Werner Huysegoms)
http://www.hpcalc.org/details.php?id=4207

JYA's UserRPL suggestion (unsafe, no check for valid object):
http://www.hpcalc.org/details.php?id=3396
[do HOME 256 ATTACH before creating this program]
\<< MEM DROP \->A 26 + A\-> NEWOB \>>


[r->] [OFF]
 | 
Pages: 1
Prev: HP 49G \->TAG Command
Next: HP42s manual