1. poke4 won't accept my 32bit values
- Posted by Chris Bensler <bensler at mailops.com>
Mar 18, 2001
-
Last edited Mar 19, 2001
Hi all,
OK,let's see if i can try to explain this..
As you might know, I'm working on the wrapper lib for exotica. I've
encountered a major hurdle though.
With exotica, when one writes code for it, it is wrapped in a function
and
passed to WinMain.. with my lib, I've managed to inline the code,
removing
the need for WinMain. This works fine and dandy for most things...
Except when I try the 3d demos.. they require poking 32bit color values
directly to memory.. when the code does not pass through WinMain, it
doesn't allow me to poke4 the values, resulting in incorrect colors..
I probably didn't get my point across, but PLEASE, if someone is willing
to
help me, or knows something about the workings of exotica, or even my
exoticaX lib, PLEASE reply..
I can't post my code here as it is much too involved. The problem itself
is
relatively straight forward though..
PLEASE, PLEASE,PLEASE HELP!!
I've been stuck on this for ages now..
I have made a workaround, but I don't like it..
TIA
Chris
2. Re: poke4 won't accept my 32bit values
- Posted by Robert Craig <rds at RapidEuphoria.com>
Mar 18, 2001
-
Last edited Mar 19, 2001
Chris Bensler writes:
> ...when the code does not pass through WinMain,
> it doesn't allow me to poke4 the values, resulting
> in incorrect colors..
I'm not sure what you are doing, but
keep in mind that Euphoria routines, such as poke4(),
deal with data in Euphoria's internal format.
Euphoria integers and C integers are the same,
as long as you are in the range of (roughly) -1 billion
to +1 billion. Outside of this range a Euphoria library
routine will interpret the number to be a kind of pointer to
a double or a sequence. To pass a 32-bit integer,
outside of the range, you would have to create a pointer
to a double, like what NewDouble() returns, and pass that.
This is very tricky, not for the faint of heart.
Also, WinMain() causes a lot of necessary stuff to
be initialized. If you skip it, some things won't work right.
Regards,
Rob Craig
Rapid Deployment Software
http://www.RapidEuphoria.com
3. Re: poke4 won't accept my 32bit values
>
> C_INT, what should it be?
>
> But even if this is the problem.. it is exclusive to my code.. No
> problems in the original Exotica demos..
>
I'm not sure, maybe Robert can chip in here, but C_UINT is a 32-bit unsigned
integer and C_INT is a 32-bit signed integer. Maybe when Euphoria invokes
the c_func() routine it converts the supplied parameters based on these
signatures. If it uses C_INT, it might be chopping off the high bits.
This is just a guess.
I'm also guessing that its Euphoria interpreter you are having trouble with.
Is it? Or is this the translated C code?
------
Derek Parnell
Melbourne, Australia
"To finish a job quickly, go slower."
4. Re: poke4 won't accept my 32bit values
Hi,
> This works fine as long as I don't need to poke the color values
> directly to memory..
> But when I do try to poke the values, it says the value is not 32bit.
>
> I hope this helps you to understand my dilemna..
Not a bit, Chris. Clear as mud. All you gave us was some code that works.
Why d'ya bother? How about showing us the code that actually fails!
Show the line(s) of code that you are using to POKE color values with.
Exactly what is telling you the "value is not 32bit"?
Is there an ex.err file you can send us?
Is it the poke4() routine? I've seen a similar message with that.
How about printing/displaying the value using sprintf("%g", colorvalue)?
Just to verify that what you think is being poked is in fact a 32-bit
integer.
What is the exact message?
Can you duplicate the error using a tiny bit of code?
Are there any colour values that work, or does every colour value fail it.
Why are you poking values anyhow?
------
Derek Parnell
Melbourne, Australia
"To finish a job quickly, go slower."
5. Re: poke4 won't accept my 32bit values
--383970696.985010192502.JavaMail.root@web577-mc
I was trying to explain how I've modified the WinMain Procedure..
but here is the exact error..
This is Main module code, not wrapped in anything..
atom Color_Value Color_Value= allocate(4)
poke4(Color_Value,#FFFFFFFF)
this is the error message in EX.ERR...
poke4 is limited to 32 bit numbers
The code works fine in the exotica demo, and it is unchanged in my demo.
Shall I send you my code Derek?
EX.ERR attached
Chris
----- Original Message -----
From: Derek Parnell <ddparnell at bigpond.com>
To: EUforum <EUforum at topica.com>
Sent: Monday, March 19, 2001 8:30 PM
Subject: Re: poke4 won't accept my 32bit values
> Not a bit, Chris. Clear as mud. All you gave us was some code that works.
> Why d'ya bother? How about showing us the code that actually fails!
>
> Show the line(s) of code that you are using to POKE color values with.
>
> Exactly what is telling you the "value is not 32bit"?
>
> Is there an ex.err file you can send us?
>
> Is it the poke4() routine? I've seen a similar message with that.
>
> How about printing/displaying the value using sprintf("%g", colorvalue)?
> Just to verify that what you think is being poked is in fact a 32-bit
> integer.
>
> What is the exact message?
>
> Can you duplicate the error using a tiny bit of code?
>
> Are there any colour values that work, or does every colour value fail it.
>
> Why are you poking values anyhow?
>
> ------
> Derek Parnell
> Melbourne, Australia
> "To finish a job quickly, go slower."
6. Re: poke4 won't accept my 32bit values
Chris Bensler writes:
> test = 4294967296 <<<<<<<<< this is the value being
> poked... (#FFFFFFFF)
Not quite....
#FFFFFFFF = 4294967295
Your "test" variable is 1 higher than
the maximum 32-bit (unsigned) integer,
so poke4() gives you an error.
To get the low-order 32-bits you could do:
test = remainder(test, 4294967296)
which in this case is 0. In general you'll get
the low-order 32-bits, which is what a C program would
give you, since C doesn't worry about integer overflow.
Regards,
Rob Craig
Rapid Deployment Software
http://www.RapidEuphoria.com
7. Re: poke4 won't accept my 32bit values
Hi Chris,
I suspect that you have a bug in your code. Somehow the valriable called
'test' is being set to #FFFFFFFF and later getting 1 added to it before you
do the poke.
----- Original Message -----
From: "Chris Bensler" <godsgift at saintly.com>
To: "EUforum" <EUforum at topica.com>
Sent: Tuesday, March 20, 2001 1:35 AM
Subject: Re: poke4 won't accept my 32bit values
> atom Color_Value Color_Value= allocate(4)
> poke4(Color_Value,#FFFFFFFF)
This example you gave works fine. Are you saying you have the line ...
> poke4(Color_Value,#FFFFFFFF)
in two programs. It works in one and fails in the other?
According to the ex.err file, there is a variable called 'test' involved.
How about you scan for each use of 'test' and double(triple!) check that it
really does have the correct value in it JUST before getting poked. The
value in the ex.err file is #100000000 and NOT #FFFFFFFF
You can send me the source code if you wish me to debug it.
------
Derek Parnell
Melbourne, Australia
"To finish a job quickly, go slower."
8. Re: poke4 won't accept my 32bit values
Chris Bensler writes:
> Rob, I think maybe you should take a look at this one..
> I can send you the code, it's pretty straight forward..
I don't know what you are trying to do,
but I think you mentioned something about
skipping WinMain(), and I warned that you can't
necessarily do that, since WinMain() calls eu_startup()
and other routines which initialize a bunch of
important data structures. Data structures that you
might not need for integer (31-bit) calculations,
but you will definitely need for larger numbers.
If that's not what you are doing,
then please send me the code.
Thanks,
Rob Craig
Rapid Deployment Software
http://www.RapidEuphoria.com
9. Re: poke4 won't accept my 32bit values
How about chnaging the code like this ...
FROM:
poke4(pvCube+16, {#FFFFFFFF,0})
TO:
atom spareatom
. . .
spareatom = #FFFFFFFF
poke4(pvCube+16, {spareatopm, 0})
just as an experiment.
------
Derek Parnell
Melbourne, Australia
"To finish a job quickly, go slower."
10. Re: poke4 won't accept my 32bit values
Hi Al,
it's still early yet, but just before you posted this I'd tracked down the
location of the problem. Superficially it appears to be a bug in the direct
draw library (create3d_device) that is corrupting a part of the interpreter.
I found this out but placing the offending line of code in various parts of
the program until I isolated the line of code that was causing the effect.
I appears that once the call to create3d_device has been executed, poke4()
would no longer work for any values between #FFFFFF01 and #FFFFFFFF. Instead
it would also convert these to #100000000, even if the values were in
variables or expressions (eg. #FFFFFF00 + #FF).
The workaround was to place all the code containing the poke4() statements
physically before the create3d_device function was executed.
Of course, none of this is conclusive yet. I've given this info to Robert as
well and maybe he can confirm it or find a deeper reason.
----- Original Message -----
From: "Al Getz" <Xaxo at aol.com>
To: "EUforum" <EUforum at topica.com>
Sent: Wednesday, March 21, 2001 4:39 PM
Subject: RE: poke4 won't accept my 32bit values
[big snip]
>
> Good luck with it.
> --Al
>
------
Derek Parnell
Melbourne, Australia
"To finish a job quickly, go slower."
11. Re: poke4 won't accept my 32bit values
>
> Thanks Derek, but it still doesn't explain why it would work for the
> original Exotica Demos, and not for ExoticaX.. Tells me it has something
> to do with my code, but I can't find it..
>
> I even used the original Exotica demo, and changed only the necessities
> to ExoticaX.. that would be: exotica_init(), and exotica_error().
> That's why I looked there..
>
> create3D_device doesn't cause the program to fail, only when I change
> the lines.. c_proc(EXOTICA_INIT,{NULL}) to exotica_init(), and
> c_func(ON_ERROR,{NULL}) to exotica_error().. the respective ExoticaX
> counterparts.. does the program give me that error..
> Only when I split the WinMain function like I have in ExoticaX do I get
> that error..
>
Well, except that the problem seems to start happening if create3d_device()
is EXECUTED before Euphoria intepreter has PROCESSED (but not executed) the
FrameMove() procedure definition. And this is the case in the original demo
but in your modified demo.
In the original demo, the Init_Loop function, which contains the
create3d_device() call is processed by the interpreter, then the FrameMove
procedure is processed etc... and the create3d_device() function only gets
executed after Euphoria finally gets to the end of the file at the
exotica_execute() function.
But, in your demo, the Init_Loop function has been converted to inline code
and is executed prior to Euphoria interpreter getting to the FrameMove()
procedure definition. If one moves the FrameMove procedure definition nearer
to the front of the file and before the create3d_device() invocation, your
demo works just fine. Move it to just AFTER the create3d_device() call and
your demo fails.
The key to it seems to be that if the create3d_device is EXECUTED prior to
Euphoria processing the definition of FrameMove() then the problem begins to
happen. If create3d_device() is EXECUTED after Euphoria processes the
procedure definition then the problem doesn't happen.
Try it yourself. In the original (working) demo, move the first few lines up
to, and including, the create3d_device() call, out of the Init_Loop()
function to somewhere before the FrameMove() definition. Make sure these
moved lines are executed before the FrameMove() definition is seen by the
interpreter. The problem is now in the original demo too.
------
Derek Parnell
Melbourne, Australia
"To finish a job quickly, go slower."
12. Re: poke4 won't accept my 32bit values
--- Derek Parnell <ddparnell at bigpond.com> wrote:
> Well, except that the problem seems to start happening if create3d_device()
> is EXECUTED before Euphoria intepreter has PROCESSED (but not executed) the
> FrameMove() procedure definition. And this is the case in the original demo
> but in your modified demo.
This all reminds me of the problems I had with GetChildFromWindow. Probably no
one remembers, but this function used to fail if you had allocated memory (and
not released it) before calling GetChildFromWindow. Some strange things seem
to happen sometimes...
Matt Lewis
13. Re: poke4 won't accept my 32bit values
------=_NextPart_000_0074_01C0B24E.6CCAD800
charset="iso-8859-1"
Chris:
I finally got around to trying your program.
Since you aren't using Win32Lib,
I tried your program using safe.e in place
of the regular machine.e.
I did the following:
copy \euphoria\include\safe.e .
rename safe.e machine.e
Then I ran Derek's "working demo.exw".
After a few seconds safe.e reported a call
to poke() that is poking out of bounds, and it gave
me an ex.err dump (attached).
Here's what's happening.
In poke_words() in exotica_api.ew
you have a loop that is poking values.
On the final iteration of the loop you (or Todd?) are
stepping out of bounds. Although you are
storing 2-byte values, you are actually poking
4-bytes at a time (from int_to_bytes()), while
advancing the address only by 2.
This is strange, but it works, except
that on the final iteration you write 2 bytes
out of bounds. pwCubeIndices is allocated a block
that's 2-bytes too small. I gave it a couple of extra
bytes and the safe.e report went away. However,
I think I have an old version of directx, so my system
hangs beyond that point.
What does this have to do with poke4 and FFFFFFFF?
Well, the interpreter can't store FFFFFFFF as a
Euphoria 31-bit integer, so it allocates a place in the heap
memory for a double. It's just possible that you are
overwriting this double, thereby changing the value
of this literal "constant" in the program.
Any change to the program, such as shuffling statements
around, do things in a slightly different way, could possibly
change what is stored just after your block of memory,
and could make the program either work or not.
This naturally leads to all sorts of wild theories as
to what might be happening.
Anyway, fix this bug and carry on.
There may be other bugs. I can't tell.
Regards,
Rob Craig
Rapid Deployment Software
http://www.RapidEuphoria.com
------=_NextPart_000_0074_01C0B24E.6CCAD800
name="save.err"
14. Re: poke4 won't accept my 32bit values
------=_NextPart_000_01C0B300.97F101A0
Chris Bensler wrote:
> Thank you Robert,
> I fixed that bug..
> instead of poking all four bytes, I changed it to only poke the first
> two bytes from int_to_bytes.. This works with Safe.e
> But like you said, after this fix, the system hangs on Safe.e
> When I run the fixed demo with Machine.e, I get the same error..
> poke4 is limited to 32-bit numbers
> I think what you found was just ANOTHER bug, not the one I'm trying to
> solve..
On my favourite 386-25 8M RAM Win95 machine,
win32lib.ew v0.45r works wery well.
But v 0.50c++ can not open comctl32.dll,
if I run both standard machine.e or save.e.
In this case, error message appeas in the
Win's crash message box without ex.err file(!),
and, in this box, text is
"Win32lib Error
Common controls could not be initialized!"
I think for day and night - Why not ?
My mainframe-386?
But, if I use machine.e then v 0.55.1 crashes
whith the next message:
"Win32lib App Window - Fatal Error
Error code 498
Common controls could not be initialized!
Win32lib v 0.55.1"
And if I run v 0.55.1 with save.e then
save.e says me about incorrect *poke*
and gives ex.err before this message above
as well (See attachment).
Then I begin to understand that
8M RAM of my mainframe may be full
of starting win32lib.
And I think - what about virtual memory?
Note, I run pure win32lib.ew whithout
any app, without just one button.
And again I love very much the simplest
solution - 386+Eu_DOS32.
Can you say me what RAM I need to
have 500 buttons (at all) in 5
windows on Win32 ?
Just interesting to know
without programming :::---)))
Regards,
Igor Kachan
kinz at peterlink.ru
------=_NextPart_000_01C0B300.97F101A0
Content-Description: Ex-55-1.err (ERR )
15. Re: poke4 won't accept my 32bit values
-------Phoenix-Boundary-07081998-
Hola, Igor Kachan escribi=F3 el 22/03/2001 10:57:39 a.m.:
>On my favourite 386-25 8M RAM Win95 machine,
>win32lib.ew v0.45r works wery well.
>
>But v 0.50c++ can not open comctl32.dll,
>if I run both standard machine.e or save.e.
>In this case, error message appeas in the
>Win's crash message box without ex.err file(!),
>and, in this box, text is
>"Win32lib Error
>Common controls could not be initialized!"
>
Igor,
Theres an update for common controls that let you run the last win32lib
progs even with 8MB of ram.
The update is:
50comupd.exe
It is 499K, and I've used on a 486DX2 66MHZ and a Laptop AMD 486 in both
I've been able to run the last version of the IDE (XIDE)
Good luck,
Fabio
-------------------------------------
Fabio Ramirez R.
Administrador de Redes - CSI
-------------------------------------
-------Phoenix-Boundary-07081998---
16. Re: poke4 won't accept my 32bit values
Hi Fabio,
>Theres an update for common controls that
>let you run the last >win32lib
>progs even with 8MB of ram.
>The update is:
>50comupd.exe
>It is 499K, and I've used on a 486DX2 66MHZ
>and a Laptop AMD 486 in both
>I've been able to run the last version
>of the IDE (XIDE)
>Good luck,
>Fabio
>-------------------------------------
>Fabio Ramirez R.
>Administrador de Redes - CSI
>-------------------------------------
Thank you very much!
No problem!
Win32lib 0.55.1 works VERY WELL
on 386-25 8M RAM with updated comctl32.dll
Regards,
Igor Kachan
kinz at peterlink.ru
17. Re: poke4 won't accept my 32bit values
Dear Eu users:
I wrote:
> Hi Fabio,
>
> >Theres an update for common controls that
> >let you run the last >win32lib
> >progs even with 8MB of ram.
> >The update is:
>
> >50comupd.exe
>
> >It is 499K, and I've used on a 486DX2 66MHZ
> >and a Laptop AMD 486 in both
> >I've been able to run the last version
> >of the IDE (XIDE)
>
> Thank you very much!
> No problem!
>
> Win32lib 0.55.1 works VERY WELL
> on 386-25 8M RAM with updated comctl32.dll
And I can to continue.
Lib comctl32.dll v.5.80 for ie 5
is much *slower* than my old ones
for ie 3 and ie 4 that was on my
machine in the past.
This new dll looks like my progs developed
on 386 for 486 with delays turned on
on 386 ---- is this passage clear-cut?
So, I wanted just to have 3 different
comctl32.dll in my directory
D:\windows\system\
under the different names for
the different jobs, rename needed
one into comctl32.dll on *plain*
DOS and then run win command to
make the needed job.
(It seems comctl32.dll for ie 4
was the fastest one on my old
and kind mainframe-386).
But MS doesn't allow this tric
***(!!! See EULA !!!)***.
(Or I don't understand that refined English?)
And I have decided to delete new .dll
and just copy my Win95's old one from
CD cab onto my HD.
Do you notice what kind of problems
we all are solveing with Eu now ?
Old and white bugs in the old C's dlls ?
But new dll is just *slower* !
My 386 ?
There are no bad hardware in the World.
All hardware is good.
I can not to throw out my old
BK-0010-01 (like PDP-10),
Poisk-1.06 (like 8088),
286-20, 386SX-33, 386DX-25 - all work as new ones.
And again I love very much my simplest,
powerful and kind friends -- 386 & Eu.
Regards,
Igor Kachan
kinz at peterlink.ru
