GNU Octave 4.2.2 Released
Octave Version 4.2.2 has been released and is now available for download. An official Windows binary installer is also available.
This version is a bug fixing release:
-
make leftdiv work for scalar \ int-matrix (bug #51682)
-
inputdlg.m: Avoid crash when prompt and defaults sizes differ (bug #53209)
-
tie
octave_classdef::numel
method to “numel” user override method (bug #46571) -
Fix performance of Sparse
fsolve
for complex sparse matrices (bug #53140) -
Fix performance of Sparse
fsolve
(bug #53140) -
octave.desktop.in: No repetition of Name in Comment field and start I10n (bug #53078)
-
Don’t create partially invalid graphic objects (bug #52904)
-
Test for incorrect regexprep on ARM platforms (bug #52810)
-
Fix incorrect regexprep on ARM platforms (bug #52810)
-
Correctly handle reading of characters >127 in
scanf
family (bug #52681) -
Fix addpath for UNC paths on Windows (bug #51268)
-
Protect being-deleted objects on figure list from second deletion (bug #52666)
-
dlmwrite.m: Close
fid
if filename is only one char long (bug #52679) -
Set gnuplot color data to half output range when autoscaling zero input range (bug #52624)
-
Add
polarplot()
to the list of unimplemented functions (bug #52643) -
configure.ac: Fix test for Java version (bug #52617)
-
For gnuplot toolkit, do not map TrueColor data to colormap size (bug #52599)
-
Make wheel scroll behave more consistently in pan mode (bug #52588)
-
Make gnuplot color have three components for interpolated edge color (bug #52595)
-
Simplify gnuplot toolkit scripts for image/non-image data plots (bug #52589)
-
Fix concatenation of empty char matrices with other strings (bug #52542)
-
build: Fix compiling
OCTAVE_ARPACK_OK_2
Fortran code (bug #52425) -
trisurf.m, trimesh.m: Fix input validation (bug #48109)
-
allow uncommenting in editor when line begins with whitespace (bug #52406)
-
do not extend selection when indenting/commenting in editor (bug #45610)
-
Remove all delimiters from whitespace list in
textscan
function (bug #52479) -
Calculate 1-norm of matrices to assess whether
NaN
orInf
are present (bug #39000) -
prevent extra ampersand under KDE in
cd-or-add-to-path
dialog (bug #52423) -
plotyy.m: Fix error when using
FUN2
argument (bug #48115) -
Check ARPACK library for buggy behavior in configure (bug #52425)
-
Fix printing integer type images (bug #51558)
-
Fix segfault in
delaunayn
when Qhull memory is not properly cleared (bug #52410) -
Fix segfault with CHOLMOD library and empty matrices (bug #52365)
-
tag global and persistent symbols as variables when parsing (bug #52363)
-
Properly restore the input stream pointer at end of
textscan
(bug #52116 et al.) -
Fix building with Qt4 for Windows (bug #52237)
-
Ensure numeric values are passed for the axes
"clim"
property (bug #52053) -
avoid abort on exit from GUI (bug #50664)
-
Correct auto limits on log axes with negative and zero values (bug #51861)
-
Fix warning in
quadgk
with zero size interval (bug #51867) -
sparse: correctly handle scalar column index (bug #51880)
-
Fix segfault in
ichol
under certain conditions (bug #51736) -
configure: ensure empty pkg-config results are actually empty (bug #51680)
-
Fix ‘legend hide’ for gnuplot (bug #50483)
-
qqplot.m: Fix typo in input validation (bug #51458)
-
Add possible
'\r'
to smartindent regex exprepression (bug #51279) -
make
strncmp
case sensitive again (bug #51384) -
Fix possible infinite loop in normest1.m (bug #51241)
-
also run unwind protect cleanup code on interrupt exceptions (bug #51209)
-
Fix crash when inverting complex matrices with
NaNs
(bug #51198) -
Improve accuracy of residue for inputs with very different magnitudes (bug #51148)
-
publish.m: Fix corruption of results for some code inputs (bug #51178)
-
residue.m: Remove code that filters out small return values (bug #34266, bug #49291)
-
avoid possible double free at interpreter exit (bug #51088)
-
show stack trace for errors in command line and startup files (bug #49346)
-
interp1.m: Return
NA
for all columns which are out of bounds (bug #51030) -
use
idx_type
for dimensions instead ofint
(bug #50934) -
show stack trace for wrong type arg errors (bug #50894)
-
Let mouse selection of Qt figures update
"currentfigure"
(bug #50666) -
disable qscintilla editor drag and drop so parent will handle it (bug #50559)
-
quadgk.m: Correct error messages which point to
quadv
(bug #50604) -
Set version on
AppUserModelId
(bug #50428) -
version-rcfile: Don’t try to execute startup directory, only startup.m (bug #50593)
-
dlmread: Return empty matrix when requested range is outside data (bug #50102)
-
Fix
eigs
for generalized nonsymmetric and shift-invert problems (bug #39573) -
Fix
eigs
for the generalized eigenvalue problem (bug #50546) -
datetick.m: Fix uneven range bugs (bug #50493)
-
datenum.m: Correct calculation for fractional leap years (bug #50508)
-
datenum.m: Allow horizontal vectors of dates with fractional months (bug #50508)
-
datenum.m: Accept legal input of vectors with fractional months (bug #50508)
-
fix the anchor position in the info text of the doc browser (bug #50422)
-
Fix order of legend labels with
plotyy
axes (bug #50497) -
Correct hggroup plot legends for gnuplot toolkit, add legend demo 17 items (bug #49341)
-
For gnuplot graphics toolkit, show only one key entry for errorbars (bug #49260)
-
Fix compilation of jit caused by cset d0562b3159c7 (bug #50398)
-
Remove inline keyword on file_stat destructor which breaks MacOS compilation (bug #50234)
Documentation bugs fixed:
-
playblocking.m: Correct documentation about start and limits inputs (bug #51217)
-
Fix eig output argument description (bug #50524)
-
Remove backslashes before double quotes in m-file docstrings (bug #52870)
-
Tweaks to use single quotes instead of double quotes (bug #52870)
-
Correct
fieldname
of returned struct inver
(bug #52845) -
Cleanup
@code
example in Appendix on test functions (bug #52852) -
Fixes for signal, image, audio, and OOP chapters (bug #52844)
-
Fix issues in geometry, polynomial, and interpolation chapters (bug #52835)
-
Fix TeX documentation for
qp
and clarify size of inputs (bug #52829) -
Correct errors in Diagonal matrix chapter of manual (bug #52814)
-
Replace
@math{1e^{XXX}}
sequences with raw1eXXX
(bug #52827) -
Use
'...'
rather than deprecated'\'
for line continuation (bug #52828) -
Make documentation Sec 26.1 more consistent and Sec 25.4 clearer (bug #52685)
-
Documentation fixes for linspace, logspace, lookup (bug #52785)
-
atan2d.m: Correct documentation to match
atan
docstring (bug #52786) -
Small tweaks to
fplot
and surfnorm docstrings (bug #52761) -
Rewrite documentation for Advanced Indexing (bug #52723)
-
Delete extra
']'
inscanf
docstring (bug #52742) -
Fix mistaken use of space between function and
'('
in documentation (bug #52723) -
Fix various inconsistencies in manual (bug #52712)
-
Fix typo in cset 8354b505ad6b (bug #52702)
-
Fix inconsistencies with
char
,strvcat
,strread
docstrings (bug #52702. -
Explain Matlab compatibility of
fopen
modes (bug #52644) -
Update documentation for keywords to include
classdef
statements (bug #52591) -
Fix documentation of third input to
lsode()
(bug #52664) -
Clarify quiver/quiver3 documentation when a linestyle is given (bug #52608)
-
new section about classdef classes with example (bug #44590)
-
Correct surface plot explanation of meshgridded results of 1 input (bug #52536)
-
Fix definition of Delaunay triangulation in docstrings (bug #52416)
-
accumarray.m: Add
'@'
to function handles in docstring (bug #52418) -
Update manual to explain
\deg
and\circ
symbols (bug #52287) -
Correct documentation for
randg
(bug #52118) -
Add documentation about PCRE library regexp stack overflow (bug #51589)
-
play.m: Correct documentation about start and limits inputs (bug #51217)
-
Redo docstring for
qz
(bug #50846) -
describe optional install dependencies PortAudio and SUNDIALS (bug #50513)
-
update
CITATION
date, version, and permalink to manual (bug #47058)