Re: HongKong 1.60 released

[Brian Howlett]

On 25 Jul, Steve Drain wrote:

> HongKong is a version of the Mahjong patience game Shanghai.

> It was written as a demonstration of Basalt and its Toolbox keywords.

26 Jul 00:55:49 000 80000000: Error from (unknown): Internal error:
undefined instruction at &20D487E2

Iyonix, RO 5.18
--
Brian Howlett
----------------------------------------------------------------------
I put instant coffee in the microwave, and almost went back in time...

[Steve Drain]

Brian Howlett wrote:
> 26 Jul 00:55:49 000 80000000: Error from (unknown): Internal error:
> undefined instruction at &20D487E2
>
> Iyonix, RO 5.18

Thanks for the report.

I do not have an Iyonix to test on, but it has been tested with RPCEmu
5.17. There does not seem to be a 5.18 rom version of that, or am I
missing something?

That address is in the RMA, I think. If you have any way to identify
which module it belongs to on your machine, I would be most grateful.

There is an issue with 5.19 that I know of, but that is a development
version and there are still other issues with that, so I will not pursue
it ATM.

I had a real problem with 256 colour modes on both emulators, but that
turned out to be caused by Paint on my RPC 4.39 doing something odd to
the mask when including an anti-aliased character in a sprite. All other
colour depths were fine, so was it Paint, ColourTrans or the Toolbox;
who knows?

These issues with ROOL development software are discouraging, because so
awkward to trace. I would be delighted to hear of anyone who uses
HongKong without problems. ;-)

Steve

[Peter Young]

On 26 Jul 2012 Steve Drain <st...@kappa.me.uk> wrote:

> Brian Howlett wrote:
>> 26 Jul 00:55:49 000 80000000: Error from (unknown): Internal error:
>> undefined instruction at &20D487E2
>>
>> Iyonix, RO 5.18

> Thanks for the report.

> I do not have an Iyonix to test on, but it has been tested with RPCEmu
> 5.17. There does not seem to be a 5.18 rom version of that, or am I
> missing something?

> That address is in the RMA, I think. If you have any way to identify
> which module it belongs to on your machine, I would be most grateful.

[snip]

> These issues with ROOL development software are discouraging, because so
> awkward to trace. I would be delighted to hear of anyone who uses
> HongKong without problems. ;-)

Sorry to let you down on that one. I've just tried it on this ARMini,
with RISC OS 5.19 (23 April 2012), and get:

26 Jul 10:53:32 000 80000000: Error from (unknown): Internal error:
undefined instruction at &20C52876

and:

*where
Address 20C52876 is not in any module
*

I imagine this means more to you than it does to me!

With best wishes,

Peter.

--
Peter \ / zfc Ta \ Prestbury, Cheltenham, Glos. GL52
and \/ __ __ \ England.
family / / \ | | |\ | / _ \ http://pnyoung.orpheusweb.co.uk
/ \__/ \_/ | \| \__/ \______________ pny...@ormail.co.uk

[Steve Drain]

Peter Young wrote:
> 26 Jul 10:53:32 000 80000000: Error from (unknown): Internal error:
> undefined instruction at &20C52876
>
> and:
>
> *where
> Address 20C52876 is not in any module
> *

Thanks. But it might be claimed by a module as workspace.

> I imagine this means more to you than it does to me!

I am not pursuing any bugs in 5.19 ATM, there are so many builds and a
lot of experimental development, even in BASIC. I refer to the
disclaimer that accompanies ROOL's software. ;-)

One piece of information that might be handy is if you are running
another application that uses the Toolbox, although even that does not
rule out that as a problem.

Steve

[Peter Young]

On 26 Jul 2012 Steve Drain <st...@kappa.me.uk> wrote:

Yes, several Toolbox modules in use here, but I can't remember which
app or apps use(s) them; Transient for one, I think.

[Alan Calder]

In article <778Qr.248771$hg.2...@fx21.am4>,
Steve Drain <st...@kappa.me.uk> wrote:

[Snip]

> These issues with ROOL development software are discouraging, because so
> awkward to trace. I would be delighted to hear of anyone who uses
> HongKong without problems. ;-)

Sadly not here. I get undefined intruction at &000086EC

RiscPC SA RO 4.39

Alan

--
Alan Calder, Milton Keynes, UK.

[Steve Drain]

Alan Calder wrote:
> Sadly not here. I get undefined intruction at &000086EC
>
> RiscPC SA RO 4.39

That is serious, because it the system I develop on, and that is an
address in the BASIC arguments, which Basalt has no reason to access. ;-(

Now, I wonder how the machines can differ to produce such an error.

The version 1.60 of HongKong was released using a !RunImage as a Basalt
file, which could possibly have repercussions beyond what I expect. I
have put a version with a BASIC !RunImage at:

http://kappa.me.uk/Examples/swHongKong161a.zip

but I am not hopeful.

Many thanks

Steve

[Alan Calder]

In article <dxbQr.257990$L51.1...@fx33.am4>,

Thanks for the effort. The message I get now is undefined instruction at
&000086E8

Don't be downhearted. When I've got a moment later I'll try the program on
this machine on as clean a startup as I can manage just in case there is
something clashing.

Alan
[Snip]

[Steve Fryatt]

On 26 Jul, Steve Drain wrote in message
<c%8Qr.275869$N4.4...@fx07.am4>:

> I am not pursuing any bugs in 5.19 ATM, there are so many builds and a lot
> of experimental development, even in BASIC. I refer to the disclaimer that
> accompanies ROOL's software. ;-)

Does that extend to ARMini users who receive specific "semi-stable" (my
term) builds via R-Comp?

--
Steve Fryatt - Leeds, England

http://www.stevefryatt.org.uk/

[Steve Drain]

On 26/07/2012 23:15, Steve Fryatt wrote:
> On 26 Jul, Steve Drain wrote in message
> <c%8Qr.275869$N4.4...@fx07.am4>:
>
>> I am not pursuing any bugs in 5.19 ATM, there are so many builds and a lot
>> of experimental development, even in BASIC. I refer to the disclaimer that
>> accompanies ROOL's software. ;-)
>
> Does that extend to ARMini users who receive specific "semi-stable" (my
> term) builds via R-Comp?

ATM because I have no means of testing on 5.19, and seems to be
singularly troublesome. See my query in the ROOL forum.

Steve

[David Mc Dowell]

[Posted and mailed]

In message <ap.84093b52b5.a...@argonet.co.uk>

Steve Drain <st...@kappa.me.uk> wrote:

> HongKong is a version of the Mahjong patience game Shanghai.
>
> It was written as a demonstration of Basalt and its Toolbox keywords.
>

> It uses a drag-and-drop interface and differs in some other ways from
> alternative versions, so it might be worth a look if you are a fan.
>
> This version adds a choice of tile sets.
>
> HongKong can be downloaded as:
>
> http://kappa.me.uk/Examples/swHongKong160.zip
>
>

!HongKong Game Errors here on Iyonix RO5.18 with:-

"Abort on instruction fetch &1FE87650"

A *Where gives:-

*where &1FE87650
Address 1FE87650 is not in any module
Internal error: undefined instruction at &EC04B678

In turn another *Where

*where &EC04B678
Address EC04B678 is not in any module
Internal error: undefined instruction at &EC04B678

and so on....

Hope this is helpful,

Regards, David.



--
David Mc Dowell:
Acorn User since discovering the 'Atom' in 1980
St.Ives, Cambridgeshire, United Kingdom.

[Steve Drain]

On 30/07/2012 15:59, David Mc Dowell wrote:
> !HongKong Game Errors here on Iyonix RO5.18 with:-
>
> "Abort on instruction fetch &1FE87650"

Apart from any other problems, there does seem to be a particular
incompatibility with 5.18 and 5.19. Unfortunately, I cannot test with
either of these here; I am limited to RPCEmu 5.17, which is ok on at
least least a couple of systems, VARPC-SA 4.02, likewise, and an RPC
4.39, which may have problems elsewhere.

One test I could do easily is on the latest BASIC module, if anyone
could extract it from the rom and send me a soft copy. I do not think I
can download anything but the source from ROOL.

There is one immediate immediate issue with HongKong that might be worth
checking: the latest version of Basalt is not backward compatible with
the ancient one, 0.98, so if you have run the earlier HongKong and then
run the new one there will be unpredictable problems.

I have also recognised that making Basalt replaceable with running
programs is not going to work with the POLL keyword. Earlier programs
decoded events with the familiar CASE structure and did work, but the
POLL statement is a machine-code loop delivering events. The code runs
in the module and I have not found a reasonable way to change to the new
code loop when the module is replaced. I know this eventually causes
data aborts.

This means that I ought to change the way Basalt operates to the normal
"install module in System" method, which I hoped to avoid, but is easier
to implement. How valuable is it to have the module replaceable while in
use?

Steve

Follow-up set to programmer.

[Robert Heaton]

You can download a 5.19 ROM image for use with RPCEmu from here -
https://www.riscosopen.org/zipfiles/misc/IOMDHALDev.5.19.zip

I'll also extract the BASIC module and send that over to you.

Cheers,
Rob.

[Steve Drain]

Robert Heaton wrote:
> You can download a 5.19 ROM image for use with RPCEmu from here -
> https://www.riscosopen.org/zipfiles/misc/IOMDHALDev.5.19.zip

Thanks for that confirmation.

I identified that rom as a possibility and asked a question about its
compatibility in the ROOL form, but got no answer.

Steve

[Robert Heaton]

Sorry about that, I follow the ROOL forums, I must have missed your post.

Rob