In case you are using a ROM expansion, latest versions of Minerva have trouble 
with that.

Daniele
________________________________
From: Ql-Users <ql-users-boun...@lists.q-v-d.com> on behalf of Martyn Hill via 
Ql-Users <ql-users@lists.q-v-d.com>
Sent: Sunday, February 25, 2018 6:51 AM
To: ql-us...@q-v-d.com
Cc: Martyn Hill
Subject: [Ql-Users] Source code availability for Minerva v1.92 or v1.89?

Hi everyone!

I'm looking to get a hold of the *source *for either of *Minerva v1.92
*or *v1.89*. We have v1.98 source available, which I have poured over
for /many /an hour...

Is anyone able to contact the QView team to ask, or else has a copy to
hand of either version of source code and has permission to release to a
enthusiastic tester?

Whilst I can continue using Min v1.92 +TK2 with full & reliable
networking, my aim is to run the latest Minerva**available - possibly
after hacking v1.98 or TK2 to restore /full /QLNet
compatibility**between them**- see below.*
*

*Context:*
In my on-going exploration and testing of the network on the BBQL, I
discovered an apparent incompatibility between Minerva v1.97 and 1.98
whenever TK2 (any version) is loaded when it comes to using LBYTES
specifically. Other uses of the NET driver work as more-or-less as
expected (OPEN, LOAD, SBYTES, etc), albeit with some additional latency
sometimes observed.

I posted a thread on the forum a while back but, at that time, I had
reservations about the hardware state of my own 2x QL motherboards (an
Iss7 and an Iss5 - both 'hacked' quite a bit.)

Since then, I have tested with other QL motherboards (Iss6 and another
Iss5 - no hacks) and made adjustments to the TK2 code to validate and
now have a degree of confidence that there really is something odd with
Minerva + TK2 /after /Minerva v1.92.

Specifically, when receiving via LBYTES (and SBYTES at the remote end),
Minerva 1.97+ on its own receives perfectly, but once the NET driver is
effectively replaced with TK2 (upto v2.23), issuing LBYTES will
completely hang the /receiving /QL - immediately and regardless of
whether the sending end has even started.

It doesn't seem to make a jot of difference what is running at the
/sending/ end - nor which issue of motherboard are at either end...

ROM (Rx)        w/o TK2    with TK2
-----------------  ------------   -----------
JS and MG      OK              OK
Min v1.89       OK              OK
Min v1.92       OK              OK
Min v1.97       OK              Hangs on LBYTES
Min v1.98       OK              Hangs on LBYTES

(I don't know how well that table will render...)

I have already hacked the TK2 (v2.23) NET driver source to allow both
the Minerva NET and the TK2 NET driver (renamed to 'TNT') to co-exist,
and also to allow both Minerva's LBYTES and TK2's equivalent procedure
(again, renamed.)

When using the Minerva NET driver, but still with TK2 loaded, normal
service resumes. The moment LBYTES is issued (again on Min v1.97/1.98)
but against the renamed 'TNT' driver, the receiver hangs again. As one
might expect from the above, Minerva v1.92 and 1.89 execute LBYTES
perfectly well using /either /NET or the TNT variant.

I suspect some combination of register or stack usage/dependencies that
are no longer as TK2 expects /after /Min v1.92, but after literally
man-days of cross comparing the various source codes available, I can't
nail it down and its bugging me :-)

What I am missing is a decent disassembly of one of the 'working'
Minerva versions, or better still, the source. Hence my request.

If I can't lay my hands on the source, I'll resort to running DEA or
IDIS against the Minerva v1.98 ROM, but that's tedious as hell :-)

Any takers?

Regards,
*Martyn Hill.*

--
"There are 10 types of people in this world. Those who understand binary and 
those who don't."

_______________________________________________
QL-Users Mailing List
_______________________________________________
QL-Users Mailing List

Reply via email to