Please provide the defines, ready to be used:

// Define NO_EXTRUDRBOARD if you don't have an EXTRUDRBOARD and wish to
// reassign different functions to EXP1.
//#define NO_EXTRUDRBOARD
 
// Define NO_EXTRUDRBOARD_OUTPUT_SWAP if you have a REV F5 or lower and
// want to use EXTRUDRBOARD A for E1 and EXTRUDRBOARD B for E2.
//#define NO_EXTRUDRBOARD_OUTPUT_SWAP

Wouldn't those #defines belong in your Configuration.h? I tried to avoid any need for an individual builder to edit pins_PRINTERBOARD_REVF.h.

I suggest keeping them in the pins file, since they pertain to no other board.

For compatibility with Makefile and other build methods, these should be:

#if DISABLED(NO_EXTRUDRBOARD)
#if DISABLED(NO_EXTRUDRBOARD_OUTPUT_SWAP)

Will do.

You may want to double check these pin assignments.
Has anyone tested these pin assignments on a Printrboard Rev F with an Extrudrboard connected to Exp1? I have this configuration set up and the my pin assignments do not match the above configuration.

I am currently running this on my REV F5 with a stock extrudrboard and I've
checked this against the schematic for the REV F4 and REV F6 (available at
https://github.com/Printrbot/printrboard ). I'm guessing you're patching Printrbot's fork, which still uses the old fastio pin numbering and that's why these don't match your configuration. This is correct arduino numbering.

I just checked and the pins I was using were fastio.

Marlin no longer uses the so-called "Marlin mapping" or "fastio mapping" which numbers A0-A7 as 0-7, B0-B7 as 8-15, etc. We have killed that mapping. Thus the numbers marked in green on the linked page do not apply.

Marlin has switched to a unified Arduino-style pin mapping under fastio.h, including for all AT90USB processors. All Marlin code (with few exceptions) uses FastIO macros —such as READ, WRITE, etc.— allowing us to follow this mapping consistently.

Nor do we use the mapping provided by the Teensyduino headers. It's Arduino-style mapping all the way, in which the pin's "DIO number" is used as the pin number (except, 46 is pin E2 and 47 is pin E3.)

With that in mind, please make any necessary changes.

Yes, the linked page contains some additional pin numbers. But it also contains the arduino pin numbers, and that's what this comment explicitly references. I don't know why you are continuing to object to "fastio" when that doesn't appear anywhere in the patch.

The problem is there is some misinformation at that link, which is outdated and has shown to cause mounds of issues in the past. If history hadn't proven this over and over again (that link you found just one example of dozens online, and in this tracker), it wouldn't be such a stickling point.

If I have some time I may update the ASCII version of a similar chart that PrintrbotCo has at the top of their fork's pins.h file and incorporate it (or you can here !), but for now it is surely best to omit links that cite the old pins numbering in any form.

Correct. My comment is mainly due to the link being outdated and containing information that no longer pertains to Marlin. We have a website now, so if someone just throws together an update in Markdown format (e.g., a comment on Github will do), we may want to document this at marlinfw.org.

I asked about this previously only because of issues around this historically with various instabilities surrounding USBCON (see #6550, #6737, #4171, #2433, etc)

So it sounds like it compiled correctly without this because the Teensyduino extension must define USBCON when the correct board is selected in the AIDE ? If so, defining again (in case user is not using Teensyduino) certainly sounds correct. If not, it begs the question how this is currently working without it.

Thanks,

-=dave

Concerns over the existence of USBCON were resolved with #6737 (comment) and #6889.

I just tested compilation of the other boards (with SD and ABL enabled but no LCD): TEENSYLU, TEENSY2, SAV_MKI, PRINTRBOARD, BRAINWAVE_PRO, 5DPRINT, (and of course PRINTRBOARD_REVF) with USBCON define removed from respective pins_[board].h and all compiled w/o issue.

It seems removal of USBCON define from all pins_[board].h would be appropriate, allowing serial.h to pull it in as appropriate from Arduino.h ?

And just tested BRAINWAVE (AT90USB647) with @Bob-the-Kuhn 's Marlin extension and it too compiled as expected w/o USBCON defined in the board pins file (of course the resulting sketch wouldn't fit, but i'm sure it could be made to).

Submitted #7347 for this to save you a few seconds in case this way you want to go Scott. If this isn't correct, just reject.

Did you find that the compilation size was identical with the USBCON defines removed?

Didn't check before, but just tested PRINTRBOARD with SD, Viki2, 3-PT (essentially stock config) and it came out identical.

...and just checked BRAINWAVE (only 647 board and uses Bob's custom extension) and also identical 👍

So the consensus is to remove the #ifndef check and #error here then?

Yeah. USBCON defines for all other AT90USB boards have already been removed.

I get the general reasoning for doing it with intermediate named pins, such as for easier re-assignment, but really this is not needed. The pins file, once set, doesn't change, and if users need to edit the pins file for reassignment I'm sure they can figure out which of these blocks applies.