I2c master slave example

[suculent]

Fresh merge against current master. May fail with type changes to i2c core.

---

This change is 

[devyte]

Fixes #5288

[d-a-v]

Would you like help to remove bearssl + lwip2 changes from this PR ?

[suculent]

No, you’re kind, thanks, I’ve created it from scratch and it’s been already merged. This was supposed to be closed already. Odesláno z iPhonu 22. 12. 2018 v 9:35, david gauchard <[email protected]>:

…

Would you like help to remove bearssl + lwip2 changes from this PR ? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or mute the thread.

[devyte]

@suculent your PR has not been merged yet. The lwip changes are still included, and are now conflicting. What needs to happen is one the following:

• you need to revert the lwip stuff in your branch and push the update
• you need to redo the PR and make sure only the examples are inccluded, and not lwip changes
• @d-a-v or I can pull this PR, revert the lwip stuff locally, and make a new PR

@d-a-v was referring to the third point 😃

[suculent]

I do understand it’s urgent. Odesláno z iPhonu 22. 12. 2018 v 16:09, Develo <[email protected]>:

…

@suculent your PR has not been merged yet. The lwip changes are still included, and are now conflicting. What needs to happen is one the following: you need to revert the lwip stuff in your branch and push the update you need to redo the PR and make sure only the examples are inccluded, and not lwip changes @d-a-v or I can pull this PR, revert the lwip stuff locally, and make a new PR @d-a-v was referring to the third point 😃 — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

[d-a-v]

I do understand it’s urgent.

Not that urgent

I’ve created it from scratch and it’s been already merged.

This one is not merged yet. You have to be careful of what is committed with git.
If you use command line, do not

git commit -a

prefer

git add -p
git commit

If you use a graphical interface, then next time maybe unselect changes that are not yours (everything you did not modify) before committing.
Either way, I think it is easier to make a new pull-request.

edit
The command line way would be

cd root-of-the-git-repository
git checkout i2c_master_slave_x
git diff master > patch
git checkout master
git pull
git checkout -b i2c_master_slave_y
patch -p1 < patch
git add -p
git commit
git push suculent/i2c_master_slave_y

(suculent/i2c_master_slave_y or whatever/i2c_master_slave_y, your local name of your personal distant fork)

[devyte]

I tried the two examples here, and couldn't get them to work even after playing with the code for a while.
Notes:

• start sequence should be first get the slave up, then the master, otherwise the sequence handling errors out (and doesn't recover)
• comment about the pin numbers seem to have the numbering reversed (wemos d1 mini r2)
• the msg struct has size 32 bytes, but only 30 bytes of members, which means there are two bytes of padding. This is obvious from lookong at the struct ember ordering.
• The xmit includes those padding bytes, which are garbage, because the structs are never zeroed out.
• When receiving the msg, bytes are copied into a char array of size 32, and then a null term is assigned at the end of the array. I think the struct is then copied over, and if the last padding byte is garbage, the null term could get lost. Printing then happens with garbage at the end.
• Retransmit request doesn't seem to work.

[devyte]

Pushing back, as this needs more work.

[d-a-v]

@suculent Do you plan to fix your PR ?

I also wonder if void (*TwoWire::user_onReceive)(int => size_t); is a breaking change.

[devyte]

Pushing back again.

[suculent]

Sorry, I cannot find time for this. I'm not pursuing software I2C on ESP8266 anymore as it seems to be a chimera. My time is tight so I've switched to SPI-driven CAN-bus to chain many devices (4 for starters). int => size_t was required to pass the 2.5.0 automated review and it makes sense to me – it was suggested by you in a code-review @devyte (cheers anyway!) The example was working for me. If anyone has suggestions to better solutions, please file a PR on your own to keep it fluent.

…

On 29 Oct 2019, at 23:35, Develo ***@***.***> wrote: Pushing back again. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#5464?email_source=notifications&email_token=AABWFR5KNS6SLICRDJXAX2LQRC3D3A5CNFSM4GJJDTO2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOECSJ3YQ#issuecomment-547659234>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AABWFR4DKAWQXN2BGCVSQQTQRC3D3ANCNFSM4GJJDTOQ>.

[devyte]

CC @Tech-TX

[devyte]

Reducing 2.7.0 scope. This is to be handled as part of the ongoing Wire rework.