newRPL: [UPDATED April 27-2017] Firmware for testing available for download

[Claudio L.]

A new firmware image is ready for download.
Use the same link from the first post in this thread to get it.

For EXISTING newRPL users only:
----------------- Skip this if you don't have a previous alpha version installed --------------------------
This image had some major internal reorganization to accommodate units, so your existing objects might become invalid after an update. For this reason, a complete memory wipe is recommended after installing this ROM.
To do that:
Press simultaneously On-A-C
The option in the middle is "Clear Mem". However, to avoid accidentally nuking your memory, this requires all 3 shift keys to be pressed simultaneously, so to clear your memory, do:
L Shift + R Shift + Alpha + C (or D, same thing)
You should see a message that install will wipe your memory.
New users obviously need not do this, this is only for people upgrading from the previous alpha. I believe this will be the last major reorganization, as all main object types are now implemented.
----------------- Now here's the important part -----------------------------------------------

New in this version: UNITS!

All units in the AUR are supported. Commands UBASE, UVAL and CONVERT are already implemented.

Features of newRPL's unit system:
* Completely open system, you can define any units you like (but not yet, the command to do that is not included in this version).
* User defined units are system-wide, and don't clutter your directories, so a unit and a variable of the same name will never conflict.
* Syntax 1_[m] is now accepted, is optional on normal objects, but mandatory within symbolics. Previously, typing '1_m/X+2' would include the X as part of the unit, the new syntax allows to properly define what's unit and what's variable, and also resolves conflicts: '1_[m]/m+2' will use unit 'm' and variable 'm' as completely separate things.
* Units can be applied to other objects, like vectors, matrices, etc.
* Exponents in units are fractions, not real numbers, so 1_m^(1/3) will stay 1/3, not 0.3333333.
* Operator _ applies a unit after the fact, to an object in the stack (same as ->UNIT but cleaner). << 2 _m >> will apply the 'm' unit to the number 2 that's already in the stack.
*Temperatures work as described in a recent thread.
* All unit definitions are updated as per latest CODATA recommended values, so they may differ from the values given by the HP48.
* All units are as exact as possible for the selected system precision. For example, the conversion factor between Rankine and Kelvin is 5/9 at any precision, not just a fixed-precision constant ( like 0.55555556).

I invite all testers to check this new module and report any problems. Thank you!