TheBackShed.com

Forum Index : Microcontroller and PC projects : Micromite Firmware V5.03

Page 3 of 3

Author

Message

Phil23
Guru

Joined: 27/03/2016
Location: Australia
Posts: 1667

Posted: 12:21pm 01 May 2017

Copy link to clipboard

Print this post

Geoffg said

Phil23 said Just wondering if the SSD not connected issue can be addressed at some point in 5.3.

I cannot reproduce the "bug" that you refer to and there is nothing that I can see in the code that would make the console stop working when a display is removed.

Thanks Geoff,

I'm about to be back playing with my E100's soon.

Will try again to see if I can reproduce the issue again with 100% repeatability & minimum code.

If I can do that on both my E100's I'll post the procedure.

Phil.

(And yes, I'd then consider the issue low priority).

Geoffg

Guru

Joined: 06/06/2011
Location: Australia
Posts: 3292

Posted: 06:35pm 01 May 2017

Copy link to clipboard

Print this post

PicFan said Hello Geoff !
Please try this !

...

Thank you !
Wolfgang

I tried it and it worked perfectly!!!
What device were you running on?

Geoff

> print mm.device$, mm.ver
Micromite MkII 5.0302
> list
Option autorun on
Option explicit
Option default integer

Dim test1 = 3

On error skip
RTC gettime
If (MM.Errno <> 0) Then
Print "RTC NICHT VORHANDEN, interne Uhr wird verwendet !"
Print ""
EndIf
Print test1

End
> run
RTC NICHT VORHANDEN, interne Uhr wird verwendet !

3
>

Geoff Graham - http://geoffg.net

TassyJim

Guru

Joined: 07/08/2011
Location: Australia
Posts: 6283

Posted: 08:06pm 01 May 2017

Copy link to clipboard

Print this post

No problems on a MMX running Ver 5.0318
and I did remember to unplug the RTC first.

Jim

VK7JH
MMedit

PicFan
Senior Member

Joined: 18/03/2014
Location: Austria
Posts: 133

Posted: 10:00pm 01 May 2017

Copy link to clipboard

Print this post

Geoffg said

PicFan said Hello Geoff !
Please try this !

...

Thank you !
Wolfgang

I tried it and it worked perfectly!!!
What device were you running on?

Geoff

I have the board (Exp. 64 MM470) updated today with the last version of your homepage and it works. (Example with "ON ERROR SKIP"). Is it possible that the download with the MAC problems caused?

Thank you and best regards

Wolfgang

CaptainBoing

Guru

Joined: 07/09/2016
Location: United Kingdom
Posts: 2170

Posted: 10:52pm 01 May 2017

Copy link to clipboard

Print this post

PicFan said
Is it possible that the download with the MAC problems caused?

Hi Wolfgang.

It's unlikely the download platform was responsible (PC/MAC/*nix etc..) more likely it is on-chip corruption. I have not seen any problems with MZ series or indeed any of the "simpler" 12F, 16F, 18F stuff (limit of my usage), but odd things with a 28 pin MX170 and two 44pin MX170s worked fine after re-flashing the firmware.

It's not my first tactic, but I have got into the habit of re-flashing whenever something *really* odd is happening. Going back through this forum you will find recent examples of stuff being fixed with a re-flash or a change of chip and personally I have had three instances of weird stuff happening that was fixed by flashing the firmware.

<speculation> Maybe there is "something" with the MX <\speculation>

PicFan
Senior Member

Joined: 18/03/2014
Location: Austria
Posts: 133

Posted: 03:09am 02 May 2017

Copy link to clipboard

Print this post

Hello Captain!

Thanks for your quick reply. My problem is as follows:
I have a program which calculates only numbers and requires time and date. The program runs for about 10 days on the same MM-470 board with V5.0302 without error. On Saturday I added only the "ON ERROR SKIP" command (see examble) after the initialization of about 20 variables and "RUN". ERROR, none of the variables was more declared? I remove the command, the Programm runs, i added the command, program fail. ERROR "xyz" not declared.

I have now again version V5.0302 (MAC download) programmed, it also works this version without error?

Many thanks to all and greetings from Austria!

WolfgangEdited by PicFan 2017-05-03

CaptainBoing

Guru

Joined: 07/09/2016
Location: United Kingdom
Posts: 2170

Posted: 03:53am 02 May 2017

Copy link to clipboard

Print this post

I have pasted your code into a MX170 beast and removed the RTC to trigger the ON ERROR SKIP

this is at version 5.3

here is the screen dump:

> rtc gettime
RTC not responding
> list
Option autorun on
Option explicit
Option default integer

Dim test1 = 3

On error skip
RTC gettime
If (MM.Errno <> 0) Then
Print "RTC NICHT VORHANDEN, interne Uhr wird verwendet !"
Print ""
EndIf
Print test1

End
> run
RTC NICHT VORHANDEN, interne Uhr wird verwendet !

3
>

it works no problem. I think you had a corrupt flash and the act of taking it to 5.2 cleaned it up. 5.3 just puts you back where you were.

Edited by CaptainBoing 2017-05-03

MicroBlocks

Guru

Joined: 12/05/2012
Location: Thailand
Posts: 2209

Posted: 05:21am 02 May 2017

Copy link to clipboard

Print this post

When you suspect a corrupt flash it might be best to read the chip first, then reprogram it and compare the two. MMBasic saves stuff in flash and it there is a small change it might be a bug. If you have the 'corrupted' flash then maybe something can be found.
I used the mx chips extensively (not MMBasic) and never had a corrupt flash before.

Microblocks. Build with logic.

CaptainBoing

Guru

Joined: 07/09/2016
Location: United Kingdom
Posts: 2170

Posted: 06:15am 02 May 2017

Copy link to clipboard

Print this post

yep - that is a good idea.

sort of a "crash dump". I haven't had a problem for a while but I am going to do this next time (if there is one)

Geoffg

Guru

Joined: 06/06/2011
Location: Australia
Posts: 3292

Posted: 11:28am 02 May 2017

Copy link to clipboard

Print this post

I find it hard to envisage a corruption that would just cause a "variable not found" error but otherwise leave the firmware running perfectly. It would require a few bits to be flipped in specific locations but everything else left untouched - the odds against that happening would be astronomical.

But, there is no denying that Wolfgang experienced something like that and I have heard of other cases of where people have corrected weird behavior by reflashing the chip. It might be that reflashing resets everything (program space, options, etc) and that is the fix. Who knows?

Geoff

Geoff Graham - http://geoffg.net

CaptainBoing

Guru

Joined: 07/09/2016
Location: United Kingdom
Posts: 2170

Posted: 09:41pm 02 May 2017

Copy link to clipboard

Print this post

something similar/related

I remember an email exchange where a FOR/NEXT loop wouldn't count backwards with a local integer variable - A re-flash fixed that.

MMBasic is a massive complex beast - there will always be obscure "surprises" in anything this complex. The space shuttle's code was the most tested/debugged code ever written and they were still finding odd things right up to the last.

The pro's far outweigh the odd wrinkle and I know from my job that often the time taken to hunt down a specific cause is not worth the time/spend - especially if you get it for nothing - BTW, hows the paypal donate button coming? (HINT HINT)Edited by CaptainBoing 2017-05-04

piclover
Senior Member

Joined: 14/06/2015
Location: France
Posts: 134

Posted: 02:01pm 03 May 2017

Copy link to clipboard

Print this post

Geoffg said I have finally got the time to start on the bug list and this is one of the first that I have found and fixed - the fix will be in the next version.

But I just wanted to say a big thanks for the succinct demo code, it was a huge help in nailing a subtle bug.

Most welcome. As a programmer myself, I always seek/ask for some reproduction procedure whenever people find bugs in my software.

Bug-hunting without a repro can become "painful"...

Page 3 of 3

Print this page

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