Fred's firmware development diary

[Fred]

The serial code works and works well as far as I can tell. Something has introduced a change to the way the 0xfffe vector is treated. I'm unsure as to why that is and do not currently understand it, but have reimplemented soft reset instead for now. I'll work on it and the docs tomorrow with a view to a release tomorrow night. I have a bit of cleaning up to do, but it's pretty good now.

Fred.

[Fred]

http://i260.photobucket.com/albums/ii15 ... -comms.png

Thought I'd better post it here too :-)

Two way comms with a dedicated app on a pc :-)

No release from me tonight. It's still not ready and I need some sleep. Hopefully there is internet where I'm going tomorrow, if so I'll keep you informed with what is happening.

Fred.

[Fred]

This morning I added most of the code to return the contents of any useful flash block via serial. RAM and modify in ram are next.

Good night.

Fred.

[Fred]

Worked on this more today and tonight. It's quite close. I think the work left to integrate in ram and burning etc is not huge so I may end up with that in as well.

so :

• return a block of flash by ID
• return a block of ram by ID
• write a block of ram by ID
• write a block of ram to a block of flash by ID
• soft reset as well as hard reset (because the latter isn't working for some reason)
• Update the tables cell at a time by ID and coordinates.
• Update the tables axis values by ID and index
• Modify the tables mechanism to keep the max indices outside the table

Additionally, the case statements are making the comms file somewhat unwieldy so I may try to split them out at some stage.

At the least I want the flash return working
At a stretch I want all of it working
Realistically I'll get some subset of it done.

None of it is a hold up for the release really. I do need to bring the docs up to date though and that could take some time.

It's going to be a good release! :-)

Fred.

[Fred]

My boss used to say "get the low hanging fruit first" so I now have a soft reset function. I guess this means we'll need an orange/amber button to go with the red one LOL.

I need to map out the memory locations for the block send receive burn stuff and see if my cunning code size reducing plan is possible :-)

Fred.

[Fred]

Memory is mapped and cunning plan implemented. A few finishing touches are required and some debugging of an issue where it stops receiving and some testing of the block return function and we are good to go :-)

I'll be working on this for 4 hours in the car tomorrow and then in the evening for a release unless things go horribly wrong.

Sean!! :-) Flash please, any update is good, even if its just a partial fix ;-)

Fred.

[Fred]

The code compiles again!

Which means there are bugs to be found.

I'm going to be smart tonight and do the dishes first. That way when I get tired I can just go to bed...

Hopefully I can debug the small issue with the code and update the docs before bed. I REALLY want to release it tonight, but may not have it ready. I won't put out a flaky release if I can help it. There is always tomorrow night.

Good news is that this code is ready to send back blocks of flash and ram and close to being ready to receive blocks of the same and do useful things with them.

I'm going to try to make it happen, but if friday is anything to go by i may not have it in me. Just about collapsed on friday from sheer tiredness.

Will post again tonight.

Fred.

[Fred]

Problem solved. The issue was a single missing line. It is now semi ready to go. Needs a tidy up and docs. It would be nice if I could add some further features too. I'll see how I feel about that in a while after I've done those chores...

Fred.

[Fred]

It can now dump tables from ram and flash in some form :-)

Fred.

[Fred]

Fixed a minor bug and tested different packets. It sends out 2k lookup tables too now.

I need to fine tune and layout the packets more nicely, but it's all pretty sweet!

No release tonight, off to bed.

Should be good for one tomorrow though as it's all very very close, just tidy up and document really. Perfect work for the train.

Fred.