Re: [Flashforth-devel] problems with recent source build (FlashForth 5 PIC18F46K22 13.06.2019)

[craig b. <dab...@ya...>]

 Thanks, Mike, but I'm guessing that's not my problem... Results aren't necessarily the same sending the same file to the same processor after sending the "empty" word. I emptied it last night before going home, started everything back up this morning, added extended version of immediate free at every marker and loaded the file up. It finished successfully.

Here's the starting point from "empty":

 flash hi dup u. here dup u. - u. cr ebff 2300 c8ff 
 ok<$,flash>
 eeprom hi dup u. here dup u. - u. cr efff ec0c 3f3 
 ok<$,eeprom>
 ram hi dup u. here dup u. - u. cr ff37 f1a1 d96 


Here's the ending stats:

 flash hi dup u. here dup u. - u. cr ebff 640e 87f1 
 ok<$,flash>
 eeprom hi dup u. here dup u. - u. cr efff ec2c 3d3 
 ok<$,eeprom>
 ram hi dup u. here dup u. - u. cr ff37 f22e d09 


I'm still under the half-way point on flash space according to this log.
Also, this same file loaded to a Pic18F45K22 build of the same source behaved just as badly.

If I were to empty the chip and try to load it again at this point, it's highly likely that things would blow up again. instead, I'm going to shut everything down, let it cool off, and try to load the interrupt system words later.

I've added 528 words to the user dictionary beyond the 294 of the system. As I remember, there was a way in the older forths to daisy-chain multiple used dictionaries. That probably isn't as useful in an empeded context, but is it worthwhile to do an "iflush" every once in a while? Will this force the flash dp to a 64-byte boundry or will it carry any already-defined bytes in the buffer over and add on from that point when more comes in?

I'll keep playing with it to see if I can spot more behavioral patterns. If it matters, I'm developing on an old Gateway P3 based laptop running Win-XP Home SP2 using TerraTerm v4.75...

Oh, one thing that struck me as odd with this board is when TeraTerm connects. I routinely plug its onboard FTDI into the computer, start TerraTerm (which is already set to its virtual com port), then apply power to the board. Normally I expect to see the FF5 self-announce at this time, but on this board I see nothing at all until I hit enter. If I hit enter before applying board power, there's a "?" vefore the OK prompt. Odd, and I don't do board design so I have no idea why it acts like that.

Many thanks for your time,
craig


     On Friday, August 30, 2019, 5:36:39 AM EDT, Mikael Nordman <mik...@fl...> wrote:  
 
 Hi Craig.

If your flash DP grows over $ec00 it will overwrite the system DPs in 
eeprom.

There is no check or warning about this limit.

You can put 'flash here cr u. cr' in your source files to see what 
happens.

BR Mike

On 2019-08-29 23:19, craig bair via Flashforth-devel wrote:
> Am I violating some limit I'm unaware of yet? (It IS a lot of words...)
> Is there an earlier known-stable version of the source that I should be 
> using?
> Could I be puting something into a colon-definition that could trash a
> system pointer?
> 
> Any thoughts and/or suggestions appreciated gratly.
> Thanks.
> craig bair
> 
> 
> _______________________________________________
> Flashforth-devel mailing list
> Fla...@li...
> https://lists.sourceforge.net/lists/listinfo/flashforth-devel

-- 
--
Mikael


_______________________________________________
Flashforth-devel mailing list
Fla...@li...
https://lists.sourceforge.net/lists/listinfo/flashforth-devel