Releases: kalmarek/Arblib.jl
v1.2.1
Arblib v1.2.1
This is a patch release which fixes a few minor issues and improves the CI
- Fixes some issues with
show
forArb
andAcb
withcompact
printing - Add constructor for
Mag
,Arf
andAcb
fromComplex
and constructor ofArb
fromAcb
- Fix issues with
minimum/maximum
forArbRef
and mixtures betweenArb
and other types. - The CI now runs Aqua.jl and the Github actions are more up to date (thanks to @devmotion)
Merged pull requests:
- Fix admonitions (#184) (@devmotion)
- Add dependabot and update CI (#185) (@devmotion)
- Add Aqua tests and fix unbound type parameters (#186) (@devmotion)
- Fix
setprecision
errors on Julia nightly (#187) (@devmotion) - Bump codecov/codecov-action from 1 to 4 (#188) (@dependabot[bot])
- Bump actions/checkout from 2 to 4 (#189) (@dependabot[bot])
- Bump julia-actions/setup-julia from 1 to 2 (#190) (@dependabot[bot])
- Fix codecov integration (#191) (@devmotion)
- Fix ambiguities (#192) (@Joel-Dahne)
- Fix some edge cases in
string
andshow
forArb
andAcb
(#193) (@Joel-Dahne) - Release version 1.2.1 (#194) (@Joel-Dahne)
v1.2.0
Arblib v1.2.0
This is a minor release with a few updates:
- The
ArbRoundExact
rounding mode is removed since it never actually existed in Flint. #181 - Flint is upgraded to 3.1. #182
- The
show
andstring
methods have been updated forMag
,Arf
,Arb
andAcb
. Thestring
method now takes arguments that affect the formatting. The default formatting forshow
has also changed, in particularMag
is now printed as a floating point number and the output is shortened for compact printing #183
Merged pull requests:
- README: Update Zenodo link to generic version instead of 1.0.0 (#180) (@Joel-Dahne)
- Remove the non-existing rounding mode ArbRoundExact (#181) (@Joel-Dahne)
- Flint 3.1.0 (#182) (@Joel-Dahne)
- Improve
show
andstring
(#183) (@Joel-Dahne)
v1.1.0
Arblib v1.1.0
This is a minor release with a few new features:
- Add
typemin
andtypemax
forMag
,Arf
andArb
. - Add
frexp
andldexp
forArf
andArb
. - Add support for
T == Float64
inradius
. This doesn't allocate and can be useful for example when heuristically choosing algorithm based on the radius. SinceFloat64
has a bounded exponent this will overflow (rounding up) for extremely small or large radius. - Add support for
T == Arf
ingetball
. - Add
log2
andlog10
forArb
andAcb
. - Improve conversion to
Int
andBigInt
. The conversion toInt
would previously crash on overflow. - Add
load_string
, non-inplace version ofload_string!
.
It also improves the handling of rounding arguments when converting to other float types. In particular it handles rounding arguments for conversion to BigFloat
, Float64
, Float32
and Float16
. For Arb
the conversion is done for the midpoint and the radius does therefore not play a role in the rounding. In the future it is possible that rounding for Arb
will consider the full ball.
Merged pull requests:
- add badges: doi, ci, coverage (#176) (@kalmarek)
- Float interface improvements - and some other things (#179) (@Joel-Dahne)
Closed issues:
- Arblib.setprecision broken for AcbRefMatrix (#177)
v1.0.0
Arblib v1.0.0
This is a major release, with a new version of Flint, some new features and a couple of breaking changes.
- The package now uses the newly released Flint 3. With this the, previously standalone, Arb library has now been merged into Flint. This package is however still mainly focused on wrapping the part of Flint that corresponds to what previously was Arb. There are very few breaking changes from this update, see the release notes for Flint 3 for more details.
- This release reduces the usage of
XLike
types, such asArbLike
, and instead uses e.g.ArbOrRef
. This could lead to breaking changes in a few cases. See #168 for motivation and the changes. - We no longer overload
Base.union
,Base.intersect
andBase.contains
, instead useArblib.union
,Arblib.intersection
andArblib.contains
(#172). Arblib.midpoint
now has support forAcb
andArblib.union
andArblib.intersection
now has support for polynomials (#173)- Multiargument versions of
+
,*
,min
andmax
has been added, which should improve performance slightly (#164, #166). - The constructors for vectors and matrices have been slightly updated. In general this should not lead to any breaking changes unless one uses the lower level constructors (#167).
Merged pull requests:
- Documentation (#159) (@Joel-Dahne)
- Remove constructor of Int from
ArfLike
and add fromArfOrRef
(#160) (@Joel-Dahne) - Preparation for Flint 3 (#162) (@Joel-Dahne)
- Make setindex! always use set! (#163) (@Joel-Dahne)
- Arithmetic cleanup and additions (#164) (@Joel-Dahne)
- Poly multiarg methods an optimizations (#166) (@Joel-Dahne)
- Update vector and matrix constructors (#167) (@Joel-Dahne)
- Avoid use of
XLike
(#168) (@Joel-Dahne) - Update precision and setprecision (#169) (@Joel-Dahne)
- Avoid overloading
Base.union
,Base.intersect
andBase.contains
(#172) (@Joel-Dahne) - Add midpoint for Acb and union and intersection for polynomials (#173) (@Joel-Dahne)
- Release 1.0 (#174) (@Joel-Dahne)
- Fix missing compat for LinearAlgebra (#175) (@Joel-Dahne)
Closed issues:
v0.8.1
v0.8.0
Arblib v0.8.0
Merged pull requests:
- Some new methods (#155) (@Joel-Dahne)
- Fix minimum and maximum (#156) (@Joel-Dahne)
- Update struct for Series (#157) (@Joel-Dahne)
v0.7.3
Arblib v0.7.3
Merged pull requests:
v0.7.2
v0.7.1
Arblib v0.7.1
Closed issues:
- Why Acb is not a subtype of Complex? (#151)
Merged pull requests:
- 1.6 cleanup (#152) (@Joel-Dahne)