Skip to content

Latest commit

 

History

History
122 lines (115 loc) · 5.29 KB

ChangeLog-5.44.md

File metadata and controls

122 lines (115 loc) · 5.29 KB
Raw

Maxima 5.44 change log

New items in core:

  • Additional float constants: LEAST-NEGATIVE-NORMALIZED-FLONUM, largest_negative_float, least_negative_float, least_positive_normalized_float, and least_negative_normalized_float.
  • Support for pre/post sub/superscripts (get_index_properties, declare_index_properties and remove_index_properties). (2) Make declare_index_properties
  • Mark non-trivial results from gcfactor with a GCFACTORED flag

New items in share:

  • pytranslate, which converts maxima to python
  • rtest_inteqn
  • descriptive: find_runs and find_runs_inverse

Changes in core:

  • Code clean ups and removal of functions that only had a historical meaning
  • In TEST-BATCH, ensure that list of unexpectedly passed tests is nonempty before reporting it.
  • :lisp-quiet can now be told not to use subsequent :lisp commands
  • defmfun now adds an ARG-LIST property that allows frontends to validate arguments
  • An improved version of float_approx_equal
  • Better handling of non-numerical input to find_root
  • Better handling of defint errors triggered by laplace()
  • Avoid variable capture in the WITH-NEW-CONTEXT macro
  • Evaluate the body of a WITH-NEW-CONTEXT form as an implicit PROGN
  • Numbers are no loger compared using EQ instead of EQL
  • Characters are no loger compared using EQ instead of EQL
  • Fixed bogus translations of apply2 and applyb2
  • Fixed bogus translations of atan and atan2 with float arguments
  • Fixed the bogus translation of atan2 with one float and one rational argument
  • Fixed bogus and inconsistent translations of max and min when the arguments were a mix of float and other modes
  • Fixed bogus translations of = and # with mixed float and numerical arguments
  • Fixed an internal error in the translator when attempting to convert a variable of mode rational to a float
  • Fixed the inconsistent translation of signum with a float argument
  • Fixed the bogus translation of random with a float argument
  • Fixed the translations of log and sqrt so they honor the flag tr_float_can_branch_complex
  • Updated gentran to v.5
  • Improved the pattern matcher for "+" and "*" expressions, see maxima-discuss, circa 2019-12-27: "Function to recognise series"
  • allow zn_log to work on subgroups
  • gamma_incomplete_lower is now returned in the noun form in many places
  • operatorp(expr,op) was inconsistent with operatorp(expr,[op])
  • sqrtdenest is now built-in, not a share function.
  • load(sqdnst) now is a no-op.

Changes in share:

  • Interpol used funmake instead of apply
  • An updated version of COMA
  • A testbench for draw()
  • plot2d with a single function now uses xlabel
  • Many minor corrections
  • In package amatrix, protect against empty matrices when converting amatrix to matrix and vice versa
  • Fixed contragrad in ctensor, added tests for cograd/contragrad

Bug fixes:

  • 3412: Corrected the translation of conditionals
  • 3587, +3489: Better handling of limits of logarithms
  • 3521: In simplify_sum, avoid double-counting -min_ni term in to_hypergeometric
  • 3265: gcfactor(0) -> division by zero
  • 2839: gcfactor(9) => 9
  • 3287: Cross product of scalar zero versus vector yields zero result
  • 3583: Stack overflow for equality testing with assumptions
  • 3608: logic_simplify handles inequalities incorrectly
  • 3607: printprops displays matchdeclare properties incorrectly when value is a lambda
  • 2174: Bogus translation of declare
  • 3513: Emacs: wxdraw3d not working
  • 3048: notequal is not translated properly

Unnumbered bugs:

  • plot now works correctly with function names containing _, ^, {, } and @
  • Maxima now no longer becoms unusable with errset=t
  • FREE-LISP-VARS property for LET was broken
  • translator: improved error message when a throw is not in a catch
  • PATCH-UP-MEVAL-IN-FSET had a bogus BARFO call
  • standard-input was broking during gentranin runs
  • Problem loading abs_integrate in imaxima mode (emacs)
  • compgrind:true caused compfile to not write to the output file
  • compgrind:true didn't ensure pretty-printing of compfile output
  • Taylor expansion for gamma_incomplete_lower(1/2,z) now works
  • compgrind:true didn't ensure pretty-printing of compfile output
  • gamma_incomplete_lower now respects gamma_expand

Additional improvements:

  • The documentation now is build use a stock makeinfo
  • Many other improvements to the way documentation is generated
  • The RPM files now end up in the build directory
  • A way to run many draw() and plot() examples for manual tests
  • More out-of-source-build improvements
  • "make dvi" was subject to bit-rot and should now work again
  • Resolved a few cases in which common variable names in arguments caused unexpected behavior
  • Better ABCL integration
  • Added a help message (Option: /?) to the windows installer
  • Fix errors when installing on MacOS via Homebrew
  • run_testsuite(share_tests=only) now doesn't fail due to unicode-sniffer.lisp not having been loaded
  • Draw now translates the coordinates of labels, too
  • Use DLMF 8.8.9 to expand gamma_incomplete(a+n,z)
  • Expand gamma_incomplete for rational order
  • Many code cleanups
  • Many improvements to the documentation
  • Windows Installer: Updated wxMaxima to the new release.