Skip to content

v2.1.0

Latest

Choose a tag to compare

@jiggzson jiggzson released this 20 Nov 19:53
v 2.1.0

Release Nerdamer 2.1.0: new functions, solver improvements, and 1.x compatibility

Functions
- Add vars for listing expression variables, binom for exact binomial
  coefficients, nextprime, powmod, taylor, rank, and trace.
- Expose these functions through Nerdamer Notation and the appropriate
  JavaScript APIs.

Solving
- Improve exact exponential solving. For example, solve(2^x=8,x)
  now returns {3} instead of a logarithmic expression.
- Return complete periodic solution families for equations such as
  e^x=-1 instead of only the principal solution.
- Improve symbolic cubic solving, including cubics with three real
  roots that previously returned no solutions.
- Add Jenkins-Traub as an alternative polynomial root-finding method.
  Aberth remains the default.
- Allow an optional numerical search interval as the third argument
  to solve(), with a default of [-10,10]. The interval limits the
  numerical search, not exact symbolic solutions.
- Reject inequalities passed to solve(), solveeqs(), or roots()
  instead of incorrectly treating them as equations.
- Infer the variable for supported univariate solving operations.

Calculus
- Allow variable inference for differentiation and integration when
  the expression contains exactly one variable.
- Expand symbolic integration coverage, including additional products
  involving exponentials and trigonometric functions.
- Evaluate numerical definite integrals at the configured Decimal
  precision by default, with USE_NATIVE_DEFINT available for faster
  native-number evaluation.
- Detect certain divergent definite integrals, preserve the
  unevaluated expression, and report defint.singularity.
- Fix higher-order differentiation of standalone exponentials.
  For example, diff(e^(3*x),x,2) now gives 9*e^(3*x)
  instead of stopping after the first derivative.

Algebra and numerical correctness
- Improve logarithm-ratio simplification, including cancellation
  within products, powers, and sums.
- Preserve sign and branch-sensitive expressions when simplifying
  noninteger powers and products.
- Correct evaluation of sqrt(-x), avoiding an incorrectly repeated
  square-root operation.
- Improve comparisons involving infinity and max/min expressions
  containing infinite arguments.
- Evaluate large integer powers exactly, subject to the configured
  MAX_NUMBER_DIGITS limit.
- Correct handling of undefined operations, including negative
  powers of zero, log(0), gamma at nonpositive integers, and
  negative double factorials.
- Add configurable limits for factorials, double factorials,
  Fibonacci numbers, and radical factorization.

Sets and collections
- Restore Nerdamer 1.x functionality that was missing from the
  2.x rewrite: append, union, intersection, difference, intersects,
  and issubset.
- intersects(A,B) returns 1 when two sets share a value.
  issubset(A,B) returns 1 when every value of A belongs to B.
  Both return 0 otherwise.
- Make the restored operations available through Notation, the
  nerdamer JavaScript API, and nerdamer/structures, using isSubset
  as the TypeScript name for issubset.
- Preserve unresolved symbolic set calls and reject arguments
  that are not sets, consistent with existing set operations.
- Extend element-wise functions to finite sets, including abs
  and factor, while preserving set uniqueness.
- Fix parsing of function names immediately before closing brackets
  or commas.

Aliases and 1.x compatibility
- Add registry-based alias dispatch through registerAlias(name,
  target, { deprecated }). Aliases resolve to their canonical
  functions and preserve argument handling and errors.
- Permanent aliases inherit the target's numerical evaluator.
  Deprecated aliases use parser dispatch to preserve warnings
  during numerical evaluation. Connecting a numerical function
  through an alias connects its target.
- Support permanent aliases including ceiling, magnitude,
  factorial, dfactorial, phase, and ilt. The latter is now
  available in Notation as well as on nerdamer.
- Support all inverse trigonometric and hyperbolic arc names,
  including arcsin through arccot and arcsinh through arccoth,
  in Notation and on nerdamer.
- Restore deprecated Nerdamer 1.x names IF, is_in, is_subset,
  and solveEquations, mapping them to if, contains, issubset,
  and solveeqs respectively.
- Retain the corresponding JavaScript compatibility aliases,
  including nerdamer.is_subset and nerdamer.solveEquations.
- Record api.deprecatedFunction warnings identifying the
  replacement. Translate these messages into all seven
  supported languages.

TeX and warnings
- Improve TeX formatting of periodic solution parameters.
- Add TEX_PERIODIC_VAR for selecting the displayed parameter
  and tex.periodicVariable warnings when a conflict occurs.
- Document the new defint.singularity, tex.periodicVariable,
  and api.deprecatedFunction warning codes.

API and settings
- Rename Complex.add() and Complex.sub() to plus() and minus()
  in nerdamer/advanced for consistency with other arithmetic
  methods.
- Remove the obsolete MAX_FRAC_INT setting.
- Preserve the supported TypeScript package entry points and
  update the public type documentation.

Documentation
- Update release notes and the 1.x migration guide.
- Document restored set and collection functions, compatibility
  aliases, new mathematical functions, and solver changes.
- Update the Sets and Warnings guides, including deprecated
  function names and warning behavior.
- Regenerate parser function metadata and TypeDoc data,
  including deprecation information.

Issues resolved by this release:
Closes #696

Previously resolved in Nerdamer 2.x:
Closes #612
Closes #302
Closes #613
Closes #615
Closes #616
Closes #623
Closes #653
Closes #658
Closes #667

Note: #653 is resolved through the supported nerdamer.solve() API.
The former chainable solveFor() interface is not restored.