Skip to content

Commit

Permalink
Merge pull request #14 from bxparks/develop
Browse files Browse the repository at this point in the history
merge 0.9.0 into master
  • Loading branch information
bxparks authored Jan 6, 2024
2 parents 4c9002f + 0310c8c commit 2172579
Show file tree
Hide file tree
Showing 250 changed files with 13,005 additions and 8,336 deletions.
76 changes: 75 additions & 1 deletion CHANGELOG.md
Original file line number Diff line number Diff line change
@@ -1,6 +1,80 @@
# Changelog

- Unreleased
- 0.9.0 (2024-01-06)
- **Breaking**: Change names and internal formats of various appVars
- `STK` list variable replaced with `RPN83STK`
- `REGS` list variable replaced with `RPN83REG`
- `RPN83SAV` remains unchanged
- the old `STK` and `REGS` variables can be manually removed
- see [TI-OS Interaction](docs/USER_GUIDE.md#ti-os-interaction) for more
details
- **Breaking**: Add `CPLX` menu at the previous location of `CONV`
- `CPLX` menu seemed most convenient on row 1 of the `ROOT` menu where
the old `CONV` was located
- `CONV` got pushed to row 2, where the `UNIT` menu was previously
located
- `UNIT` got pushed to row 3, into an empty slot
- the RPN83P home menu row now has exactly the same items as the `MATH`
menu on the TI-OS:
- RPN83P: `MATH`, `NUM`, `PROB`, `CPLX`
- TI-OS: `MATH`, `NUM`, `CPX`, `PROB`
- Support more than 14 digits during edit/input
- When more than 14 digits are entered, the left most digits scroll off
to the left, with the left most digit replaced with an ellipsis
character indicates existence of extra digits.
- Normal mode:
- accepts maximum of 20 digits, which supports entering all 14
digits encoded by the TI-OS floating point number format
- Complex mode:
- accepts maximum of 41 digits to allow 2 floating point numbers
- `BASE BIN` mode
- accepts up to 32 digits, to allow a 32-bit binary number when the
`WSIZ` is 32.
- see [Input Limits and Long
Numbers](docs/USER_GUIDE.md#input-limits-and-long-numbers) for more
details
- `PROB`
- Expand range of `COMB(n,r)` and `PERM(n,r)` arguments to `n,r<=65535`
from `n,r<=255`.
- Improve performance of `COMB(n,r)` when `r>(n-r)` by taking advantage
of the symmetry of `COMB(n,r)==COMB(n,n-r)`.
- Eliminate floating point round-off errors in computing `COMB(n,r)` by
incrementing the divisor from 1 to `r`, instead of decrementing it
from `r` to 1.
- Complex Numbers
- redesign RPN stack and storage registers to support both real and
complex numbers
- extend arithmetic, algebraic, transcendental handlers to support
complex numbers
- add explicit CPLX menu group with: REAL, IMAG, CONJ, CABS, CANG
- support RRES (real result) and CRES (complex result) menu settings
- support RECT (rectangular), PRAD (polar radian), PDEG (polar degree)
menu settings
- display complex numbers in RPN stack in rect and polar modes
- support complex numbers in SHOW in rect and polar modes
- support Linking/Unlinking a complex number and its 2 real components
using 2ND LINK (equivalent to COMPLEX button on the HP-42S)
- support entry of complex numbers on a single line in RECT, PRAD, and
PDEG modes using the `2ND i` and `2ND ANGLE` keys.
- see [Complex Numbers](docs/USER_GUIDE.md#complex-numbers) for more
details.
- Save and restore app MODE settings independently from the TI-OS settings
- decouple the TI-OS MODE settings from the RPN83P MODE settings
- the TI-OS MODE settings are saved upon app start, and restored upon
app exit
- the app MODE settings are saved and restore independently
- for example, it is now possible to set the TI-OS to FIX(2) and DEG,
while setting RPN83P to SCI(4) and RAD and the 2 settings are managed
separately, even though there is only a single set of global OS
settings
- **Bug Fix**: Render 3-digit EE exponents correctly in `SHOW` mode.
- 3-digit exponents can only be shown for complex numbers, so the bug
was latent until complex numbers were added
- **Bug Fix**: Fix overflow in rectangular to polar conversion `>POL`
- the built-in TI-OS `RToP()` function has a bug which throws an
exception when `x^2+y^2` becomes `>=1e100`, which can happen for `x`
or `y` as low as `7.07e49`.
- reimplement using a custom `rectToPolar()` without the scaling bug
- 0.8.0 (2023-12-03)
- **Breaking**: Flip the order of polar-rectangular conversion menu function
(`>POL` and `>REC`) so that they are consistent with the HP-42S. I don't
Expand Down
174 changes: 0 additions & 174 deletions DEVELOPER.md

This file was deleted.

Loading

0 comments on commit 2172579

Please sign in to comment.