| Newsgroups |
gmane.comp.emulators.hercules390.general |
| Message-ID |
<[email protected]> |
Hi Paul
this commit, together with 0ccc672 https://github.com/SDL-Hercules-390/hyperion/commit/0ccc6720280ac5a0ba1d378c0b2a7ea9fe1eed31, brings the TWX 33/35 and 2741 devices to a satisfactory behaviour with SDL Hercules Hyperion 4.2. However, both commits do not solve the root cause of the addressed issues, which most probably is the same for both. For that reason I don't want to actively distribute them yet to the other Hercules repositories (namely Spinhawk 3.x and mainline Hyperion).
At some future point in time I'll go after the root cause and fix it once and for all. This will then be the point in time to commit to mainline Hyperion and to ask Roger to take it over to Spinhawk.
Cheers
Jürgen
---In [email protected], <kerravon86@...> wrote :
Thanks for finding that Juergen.
I tested the unmodified Hercules 4.2 a few
times and it worked fine.
I found your commit:
https://github.com/SDL-Hercules-390/hyperion/commit/f2a8a9f210db1f1b129026e76521348600f70cc4 https://github.com/SDL-Hercules-390/hyperion/commit/f2a8a9f210db1f1b129026e76521348600f70cc4
and this is something that should go into
Hercules 3.14 too - a bug in commadpt.c.
In Hercules/380 I'm actually using the
Hercules 3.13 code for this, and I was
able to update it with the same change
you made, and so now that is working too.
BFN. Paul.
---In [email protected] mailto:[email protected], <winkelmann@...> wrote :
Hi Paul, All
The new SDL Hercules Hyperion 4.2 release, Fish just announced, contains a fix for the "quit" problem discussed in this thread. So, it's no longer necessary to use the workaround for TK4- environments I mentioned below, if one intends to use enhanced shutdown techniques (clay pigeon shooting) on the system. Nonetheless, I don't encourage using such techniques _at all_.
Cheers
Jürgen
---In [email protected] mailto:[email protected], <winkelmann@...> wrote :
Hi Paul
Well, the same of course ;-). However with differing statistics. With all three sets of binaries it may well be that a whole series of runs comes out fine, and that all fail. So, our amount of samples is apparently far below statistical significance. Too bad, but not worthwhile to be pursued further at this point in time. I think the three major clues we got
- it is an old problem
- it is not TCAM related
- it is reliably reproducible (most of the time)
suffice to go back into an orderly mode of analysis. I'll dive into this again now, hoping for more luck than during the last try in May. It may take quite a while until I come back with results.
Until then you might still consider disabling the critical devices as a workaround. Simply create a file named tk4-.parm in tk4-\local_conf and enter these two lines into this file:
@ECHO OFF
SET REP101A=specific
Your "quit" problem will definitely be gone then.
Cheers
Jürgen
---In [email protected] mailto:[email protected], <kerravon86@...> wrote :
---In [email protected] mailto:[email protected] mailto:[email protected] mailto:[email protected], <winkelmann@...> wrote :
> Now I'm of course _very_ interested to see how
> this compares to the original TK4- Hercules binaries...
I have posted those results now.
I'm curious as to what happens on YOUR
system when you type "quit" after the IPL
is completed.
BFN. Paul.