[openMSX-devel] [ openmsx-Feature Requests-3218382 ] Run init scripts under the hardwareconfig.xml

[SourceForge.net <no...@so...>]

Feature Requests item #3218382, was opened at 2011-03-17 00:29
Message generated for change (Comment added) made by sd-snatcher
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=421864&aid=3218382&group_id=38274

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Open
Priority: 5
Private: No
Submitted By: FRS (sd-snatcher)
Assigned to: Nobody/Anonymous (nobody)
Summary: Run init scripts under the hardwareconfig.xml

Initial Comment:
This feature request is a subset of the RFE #3209252: For Acid2Test and Acid4Test

Please allow init script to be put inside the hardwareconfig.xml file. It would be used to issue openMSX shell commands on the machine (or extension) initialization.

<init>
         toggle_scc_viewer
         toggle_vu_meters
         toggle_vdp_access_test
</init>


The use case for this would have special machines configured for programming. They would quickly trigger the desired scripts used for the debugging.


----------------------------------------------------------------------

>Comment By: FRS (sd-snatcher)
Date: 2011-03-20 16:22

Message:
Patriek,

That's a sad and improper accusation. I would like to know what are those
"bombastic and false claims" you're talking about.

Why are you making personal attacks? Its it because you don't agree with
the viewpoints? What are you trying to accomplish with this?

Could you please explain your point of view using well based arguments and
referencing the standard where appropriate? This is the only rational way
to get into a common agreement.


----------------------------------------------------------------------

Comment By: Patriek Lesparre (guyver800)
Date: 2011-03-20 13:25

Message:
I've been following this discussion for a while, and I suggest FRS reads
the MSX standard properly and stop making bombastic and false claims (in
general).
(this comment rewritten 4 times to be less offensive)

----------------------------------------------------------------------

Comment By: FRS (sd-snatcher)
Date: 2011-03-20 12:48

Message:
I think you're right for the machines based on scripts, like the Acid4 and
Acid6 draft proposals.

Please don't get me wrong, but I think that some of the tests, like Acid1
and Acid2 wouldn't fit well as scripts. In fact, they're almost complete
as-they-are now, lacking only some minor features, in order of importance:

1) Acid2Test:
  a) Reconfigurable MemoryMapper I/O ports, like all other devices on
openMSX.
  b) Reconfigurable expanded-slot-register address

2) Acid1Test:
  a) Possibility to place each frame of the same memory-mapper under a
different slot.

The openMSX team don't need to include the machines by default on the
emulator if they don't want to. I can made them available on my site
coupled with install instructions. I'm just asking for the features needed
for the machines to get complete. Mainly the 1a and 1b of the above list,
if you ask me.    :)



----------------------------------------------------------------------

Comment By: Manuel Bilderbeek (manuelbi)
Date: 2011-03-19 06:57

Message:
I think that starting all tests as a script would be a solution. This makes
it easier to setup the right conditions for the test and doesn't need the
change you originally suggested. That change is not something that fits in
the design at all. Machine configurations are only to define hardware, they
have nothing to do with scripts. (Think about what happens if a script
starts a machine with a script... things you really don't want :)

So, making things uniform could be done by starting any test with a
script. And with that, I think all your plans would still be implementable.

----------------------------------------------------------------------

Comment By: FRS (sd-snatcher)
Date: 2011-03-17 18:09

Message:

Humm.  IMHO, it would be a bit counter intuitive and make some tests
different to run from others.

The idea is that each test is a machine that quickly assures the
programmer if his program is doing things right. He would use the AcidxTest
machine as any other openMSX machine, so he don't have to invest time on
learning how to use them (which would be against the quick and assertive
principle of an AcidTest).

Making some test as machines and other as scripts would break this
uniformity and simplicity?

Acid1Test would be just a machine
Acid2Test would be a machine, but require some scripts to be started
manually to find the unused ports and attach watchpoint for each unused
port
Acid3Test would be just a machine again. But maybe some scripts to set
breakpoints for the original BIOS positions would be used. This breakpoints
would issue warnings.
Acid4Test would require the user to boot any machine, then start a script
on the console that starts all the timing testing.
Acid5Test would be just a machine
Acid6Test would be any machine, but the user would run some scripts that
would setup traps for common mistakes


It's a bit awkward scenario, if you ask me. Nothing serious, but awkward
anyway.    :)

But I trust the openMSX team on the decision of the best way to implement
concept proposed by the tests. Please think of the tests more like
use-cases of aspects to be tested/debugged than a canned solution as it may
be sounding.


----------------------------------------------------------------------

Comment By: Manuel Bilderbeek (manuelbi)
Date: 2011-03-17 14:49

Message:
Instead of including script inside machine configs, I'd suggest to do it
the other way around: start a script which simply creates a machine with
the proper config and do the other stuff in the rest of the script.

So, I propose to reject this request. Do you agree?

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=421864&aid=3218382&group_id=38274