Threaded Mode | Linear Mode

J-F Garnier · (This post was last modified: 02-05-2015 08:27 PM by J-F Garnier.)

(02-05-2015 07:13 PM)Christoph Giesselink Wrote: for getting the 12 byte media physical attributes from the drive you have send a DDT 7 to the disc drive before fetching the data with SDA. ... The last DDT setting was DDT 6 for getting the 2 byte last sector number.

Christoph mixed the DDT 6 and 7 numbers but the principle is there.

There is also a second bug in the INITIALIZE function of HP-IL module 1A: the directory size is incorrect. The detection method of Christoph covers both bugs, and fixes them of course.

Quote:Detecting this behavior of the 1A module was the work of Jean-Francios. My job was fixing a bug in ILPer 1.45 and earlier, so that ILPer behave in the same way like the real 9114 drive on the buggy initialization sequence.

This was really a team work: Sylvain reported the bug and provided HP-IL traces, I investivated the source cause, and Christoph devised the right fix (which was not obvious to cover all the use cases on the HP-41C, HP-71B, HP-75C/D and the Portable Plus!).

Actually the bugs are documented (and fixed) in HP-IL IDS source code of the module 1B, look at the cassette routine module (NZ&CAS) lines 367 and 552.

What was really surprising was to observe that the HP-71B with HP-IL 1A was writing incorrect disc parameters on disc sector 0 during INITIALIZE on the HP9114, then a read was showing that all information was correct!
The only possible explanation was that the HP9114 was correcting the disc parameters, so we just had to do the same in ILPer!

This is a very particular example of a major bug on one machine that has been fixed by a workaround on an other one. I don't know any other example in the field of HP calculators and computers.

This is also the reason why this bug was not known by the HP-71B user community (at least by me!), since it is completely hidden by the HP9114 firmware.

J-F