TheBackShed.com

Forum Index : Microcontroller and PC projects : CMM2: V5.05.06RC19: Scotty more power

Page 4 of 4

Author

Message

LeoNicolas

Guru

Joined: 07/10/2020
Location: Canada
Posts: 444

Posted: 06:01am 17 Nov 2020

Copy link to clipboard

Print this post

I realized that my Octahedron code runs 20 ms slower on the RC20 compared to the RC19.

Is this difference expected?
Edited 2020-11-17 16:02 by LeoNicolas

matherp
Guru

Joined: 11/12/2012
Location: United Kingdom
Posts: 8578

Posted: 08:17am 17 Nov 2020

Copy link to clipboard

Print this post

V5.05.06RC21 posted:

http://geoffg.net/Downloads/Maximite/CMM2_Beta.zip

This includes a more aggressive crunch on load making the stored program even smaller and very slightly faster.
In addition mm.info(modified fname$) will return a valid datetime if fname$ is a directory (new) or a blank string if nothing is found (as in RC19 and before)

Quote I realized that my Octahedron code runs 20 ms slower on the RC20 compared to the RC19

I had to improve the handling of co-linear coordinates in the triangle draw routine as one combination could yield incorrect results. I've also tightened up the stack overflow handling.

NB: I'm about to redo a big part of the graphics engine. If it works the result will be V5.06.0 and V5.05.06 won't ever happen

JoOngle
Regular Member

Joined: 25/07/2020
Location: Sweden
Posts: 82

Posted: 08:24am 17 Nov 2020

Copy link to clipboard

Print this post

matherp said V5.05.06RC21 posted:

NB: I'm about to redo a big part of the graphics engine. If it works the result will be V5.06.0 and V5.05.06 won't ever happen

We're not worthy, we're not worthy (Wayne's world).
Thanks Peter, to follow your development is nothing short of mind-blowing.

thwill

Guru

Joined: 16/09/2019
Location: United Kingdom
Posts: 3839

Posted: 10:31am 17 Nov 2020

Copy link to clipboard

Print this post

Good morning,

Here is today's offering:

Colour Maximite 2
MMBasic Version 5.05.06RC21
Copyright 2011-2020 Geoff Graham
Copyright 2016-2020 Peter Mather

> list "bug.bas"
' Copyright (c) 2020 Thomas Hugo Williams

Option Explicit On
Option Default None

#Include "bug2.inc"

foo()
End

> list "bug2.inc"
Sub foo()
�Local j%

�j% = 0
�Print j%
�Inc j%
�Print j%
End Sub

> run "bug.bas"
0
6

The output should of course be:

0
1

Note that the #Include matters, if I flatten the example then it produces the correct output

EDIT: If you explicitly set the increment you don't see this issue:

Inc j%, 1

Best wishes,

Tom
Edited 2020-11-17 20:49 by thwill

Game*Mite, CMM2 Welcome Tape, Creaky old text adventures

matherp
Guru

Joined: 11/12/2012
Location: United Kingdom
Posts: 8578

Posted: 10:54am 17 Nov 2020

Copy link to clipboard

Print this post

Quote Here is today's offering:

Will fix
Edited 2020-11-17 20:58 by matherp

matherp
Guru

Joined: 11/12/2012
Location: United Kingdom
Posts: 8578

Posted: 11:03am 17 Nov 2020

Copy link to clipboard

Print this post

Fixed and re-uploaded - no version change

thwill

Guru

Joined: 16/09/2019
Location: United Kingdom
Posts: 3839

Posted: 11:03am 17 Nov 2020

Copy link to clipboard

Print this post

matherp said

Quote Here is today's offering:

Will fix

Thanks Peter.

I'm in the process of enhancing my unit-test framework so it is possible to run all the tests in a directory hierarchy with a single command. Once I've done that then perhaps you'd consider running that command as part of your release process since it seems that my code "reaches the parts that other beers cannot reach."

Best wishes,

Tom
Edited 2020-11-17 21:04 by thwill

Game*Mite, CMM2 Welcome Tape, Creaky old text adventures

LeoNicolas

Guru

Joined: 07/10/2020
Location: Canada
Posts: 444

Posted: 04:45am 18 Nov 2020

Copy link to clipboard

Print this post

Wow, the RC21 is running my octahedron 8 ms faster than the RC19 and 28 ms faster than the RC20.

Amazing job Peter.
Thank you

LeoNicolas

Guru

Joined: 07/10/2020
Location: Canada
Posts: 444

Posted: 05:09am 18 Nov 2020

Copy link to clipboard

Print this post

I found a bug with the INC command.
If I put more than one INC command in the same line with the increment equals to zero, all INC commands except the last one in the row will increment the variables by one.

dim i,x,y,z

for i=1 to 10
�inc x,0:inc y,0:inc z,0
�print str$(x)+", "+str$(y)+", "+str$(z)
next

The result will be:

1, 1, 0
2, 2, 0
3, 3, 0
4, 4, 0
....

EDIT:
The first two INC commands are always incrementing by one, even if I change the increment value for a value greater than 1.
Edited 2020-11-18 15:13 by LeoNicolas

TassyJim

Guru

Joined: 07/08/2011
Location: Australia
Posts: 5901

Posted: 06:07am 18 Nov 2020

Copy link to clipboard

Print this post

The problem is bigger than it first looks.
Any multi-statement line with any INCrement reverts to INC of 1
The INCrement can be a literal or variable. Both will fail.

dim i,x,y,z, f
f = -1
for i=1 to 10
inc x,f : ' test
inc y,f
inc z,0
print str$(x)+", "+str$(y)+", "+str$(z)
next
1, -1, 0
2, -2, 0
3, -3, 0
4, -4, 0
5, -5, 0
6, -6, 0
7, -7, 0
8, -8, 0
9, -9, 0
10, -10, 0

Pushing the limits was always going to be a bumpy ride!

Jim
Edited 2020-11-18 16:08 by TassyJim

VK7JH
MMedit MMBasic Help

TassyJim

Guru

Joined: 07/08/2011
Location: Australia
Posts: 5901

Posted: 07:13am 19 Nov 2020

Copy link to clipboard

Print this post

A strange problem with READ

OPTION EXPLICIT
OPTION DEFAULT NONE
DIM INTEGER n, s
DIM p1$, p2$

READ p1$, P2$
PRINT P1$+">> "+P2$
READ p1$, P2$
PRINT P1$+">> "+P2$

END

DATA "I2C ","I2C parameter"
DATA "I2C CLOSE ","I2C CLOSE"

gives:

I2C 1 >> I2C 1 parameter
I2C 1 CLOSE >> I2C 1 CLOSE

It gets even crazier with I2C2 in the DATA lines.

With 650 data statements I estimate that 90% went through OK but that still leaves a lot of oddities.

Jim

VK7JH
MMedit MMBasic Help

matherp
Guru

Joined: 11/12/2012
Location: United Kingdom
Posts: 8578

Posted: 10:23am 19 Nov 2020

Copy link to clipboard

Print this post

Jim

I've noted and fixed your two bugs but the code is in bits at the moment as I'm re-writing a big part of the graphics engine so no updates for the moment - sorry

William Leue
Guru

Joined: 03/07/2020
Location: United States
Posts: 382

Posted: 03:49pm 19 Nov 2020

Copy link to clipboard

Print this post

Speaking of bugs, I had a weird thing happen to me yesterday.

While editing and testing a short MMBasic program, I decided to change a line at the top of the program that set the graphics mode. Initially it said

mode 1,8

But I didn't like the very limited gray-scale colors that are available in this mode, so I changed that single line to

mode 1,16

and then ran the program. To my surprise, the program, instead of running and showing some graphics, did nothing for several seconds and then terminated. But that is not the weird part.

Going back into edit, I discovered to my horror that several lines of the program source had been corrupted! And even worse, after I fixed them and re-ran the program, the same lines of the source were corrupted again.

Power-cycling the computer did not stop the problem from happening yet more times. If I changed the mode back to 1,8, everything ran correctly.

It sounds as if the firmware on 5.05.06RC15 has a bug that causes the editor to go wonky when the graphics mode is changed to 16-bit.

I realize this is going to be tough to debug since it is so complicated. Hopefully a more recent release candidate fixes the issue.

-Bill

jirsoft

Guru

Joined: 18/09/2020
Location: Czech Republic
Posts: 532

Posted: 04:07pm 19 Nov 2020

Copy link to clipboard

Print this post

Hi Bill,
I have similar problem, but didn't realised the connection to 16 bit mode. Since change back to RC10 (or so) and later to RC21, the problem was not there...

Jiri
Napoleon Commander and SimplEd for CMM2 (GitHub), CMM2.fun

William Leue
Guru

Joined: 03/07/2020
Location: United States
Posts: 382

Posted: 05:03pm 19 Nov 2020

Copy link to clipboard

Print this post

Thanks, Jiri! I will update to a new release candidate soon. In the mean time I am keeping away from 16-bit graphics.

-Bill

markboston36
Regular Member

Joined: 27/10/2020
Location: United States
Posts: 76

Posted: 02:04am 25 Nov 2020

Copy link to clipboard

Print this post

when will 5.05.06 be released? i see we already have 5.06 beta's

panky

Guru

Joined: 02/10/2012
Location: Australia
Posts: 1094

Posted: 02:21am 25 Nov 2020

Copy link to clipboard

Print this post

matherp said V5.05.06RC21 posted:

http://geoffg.net/Downloads/Maximite/CMM2_Beta.zip

This includes a more aggressive crunch on load making the stored program even smaller and very slightly faster.
In addition mm.info(modified fname$) will return a valid datetime if fname$ is a directory (new) or a blank string if nothing is found (as in RC19 and before)

Quote I realized that my Octahedron code runs 20 ms slower on the RC20 compared to the RC19

... almost all of the Maximites, the MicromMites, the MM Extremes, the ArmMites, the PicoMite and loving it!

Page 4 of 4

Print this page

To reply to this topic, you need to log in.