Summary of bugs fixed for version 10.2.0 (2025-05-29):$$$Improvements and fixes$$$tensorprod: Fix error for certain tensor/vector combinations (bug #66950).$$$mkoctfile: Skip compiling object file with soversion for --mex -c -o (bug #66972).$$$Fix segmentation fault in octave-svgconvert when called with no inputs.$$$mkoctfile: Remove temporary C source files after compilation.$$$unicode2native: Fix conversion of short strings to UTF-16 or UTF-32.$$$Detect leading and trailing comments for all parse tree elements (bug #67004).$$$bar: Avoid listener error when replacing an existing bar plot (bug #67006).$$$Fix sparse matrix assignment with reverse range indexing (bug #66516).$$$ezplot: Enable polar and 2-D plots of constant functions (bug #66563).$$$Check for any kind of file not just regular files (bug #67018).$$$var/std: Issue an error for non-floating point inputs. Previous versions accepted integer type inputs but could produce erroneous outputs due to integer arithmetic and under/overflow errors (bug #67016).$$$annotation: Change verticalalignment default value (bug #67033).$$$Previously; symbols that are not actually callable by a user might have been added to the list of callable functions when loading .mex files. Remove these symbols from that list. This fixes unloading .mex files with clear.$$$Remove break points only from user functions (defined in .m file). This fixes an infinite recursion issue when re-loading updated .oct or .mex files.$$$For improved Matlab compatibility the strncmp and strncmpi functions now return true if the number of characters specified by N is equal to 0 (bug #57879).$$$Fix nargout in multi-output indexing expressions beginning with function call (bug #67096).$$$findobj: Do not match empty tags with [] (bug #67048).$$$Check for undefined output when indexing function output (bug #67111).$$$perms: Fix buffer overflow and crash; minor code cleanup (bug #67115).$$$mkoctfile: Support spaces in output path of linker step.$$$mkoctfile: Support spaces in path to binary when stripping debug info.$$$octave-svgconvert: Speed up printing to PDF (bug #66959).$$$error: Fix segmentation fault on missing fields in err.stack structure (bug #67143).$$$Avoid parse error for empty lines in debug mode (bug #67108).$$$Check if effective SOVERSION is exported from .mex file libraries (bug #67163).$$$GUI$$$Fix build error without QScintilla installed (bug #66962).$$$Avoid potential integer overflow on Windows.$$$Make dialogs shown by inputdlg; listdlg; questdlg; uigetdir; or uigetfile modal.$$$Build system / Tests$$$Add visibility attributes for the octave::base_fcn_handle class.$$$Add visibility attributes to parse tree symbols (bug #67056).$$$Add visibility attributes to Array<T> template class member functions.$$$Use correct preprocessor macro when instantiating MArray template class.$$$Disable visibility flags by default.$$$Support passing additional flags when linking octave* executables.$$$Single-quote path that might contain unescaped backslashes in the generation of the doc cache.$$$Ignore deprecation warning for std::wbuffer_convert in header.$$$error: Add more built-in self-tests.$$$Documentation$$$Recommend use of string function name as first argument to cellfun().$$$cellfun: Call out special accelerated functions.$$$Update documentation for feval; eval; evalin; evalc.$$$axis: Remove incorrect statement from documentation.$$$Do not put graphics object categories into Graphic Properties Index.$$$Adjust Perl regexp used to remove leading ‘:’ in HTML.$$$char: Document calling form with no inputs.$$$Fix missing hyperlinks in Table of Contents of Octave manual.$$$Fix overfull hbox warnings in Texinfo log files.$$$Replace Texinfo @dots macro with @enddots macro where appropriate.$$$Do not print ugly black boxes for overfull hboxes in PDF output.$$$Fix accidental doc text inclusions after @deftypefn.$$$Remove reference to Octave Forge in stats.txi.$$$Document use of : keyword as a range operator.$$$Correct and expand documentation on ranges.$$$Remove outdated t

Summary of bugs fixed for version 10.1.0 (2025-03-28):$$$bug #66753: hist.m: Test for equal bin spacing using a numeric tolerance.$$$bug #66753: hist.m: Fix regression in determining equal bin spacing.$$$bug #66753: Fix thinko in cset be1d0c816788$$$bug #66647: Silence unexpected Octave:mixed-string-concat warning in num2str.m.$$$bug #66642: Code beautification for cset 2c2301104caf$$$bug #66642: Check for undefined outputs in cellfun and arrayfun.$$$bug #66642: Add cellfun BIST for function failing to return requested output.$$$bug #66642: Add arrayfun BIST for function failing to return requested output.$$$bug #66617: Avoid error for struct2cell with function without output arguments.$$$bug #66617: Add self-tests for structfun with function without output arguments.$$$bug #66558: Translate shortcuts and descriptions in settings dialog.$$$bug #66521: symbfact: Return row vectors for Matlab compatibility.$$$bug #66511: sparseqr: Support permutation output with CXSparse.$$$bug #66511: Emit error when qr() is called with permutation output and CXSparse library is used.$$$bug #66488: qr (sparse; 0) isn’t equal to qr (sparse; econ; vector).$$$bug #66477: Avoid ASAN error with path (_) manipulations.$$$bug #66466: legend and bar cause ASAN to crash Octave.$$$bug #66451: Ctrl-C kills Octave 10.0.0.$$$bug #66448: File encoding conversion errors on macOS 14.$$$bug #66399: fseek and ftell functions don’t always work correctly on Windows with filesizes >2GB.$$$bug #66315: menu() throws an error when Qt dialogs are not available.$$$bug #66256: movmad uses ‘mean absolute deviation’ while Matlab uses ‘median absolute deviation’.$$$bug #66156: Implement nanflag option for moving window functions.$$$bug #66025: movfun: Implement SamplePoints option.$$$bug #66010: Function glpk produces incorrect output.$$$bug #65964: nthargout does not propagate error ID.$$$bug #65928: movfun dimension constraints: Shouldn’t error for dim > ndims(x); or wlen > size(x; dim); or wlen=1.$$$bug #65876: Error retrieving data from struct values in containers.Map.$$$bug #65768: Segmentation fault on default branch.$$$bug #65753: Strip leading/trailing whitespace from “Function Index” search expression.$$$bug #65753: Documentation window: Function Index: Search box respects trailing spaces.$$$bug #65730: Some remaining cases of silent conversion of fractional inputs.$$$bug #65683: issorted: enable monotonic and strict... sort modes.$$$bug #65674: axes colormap property being set over figure colormap property.$$$bug #65665: Input validation for system().$$$bug #65645: Execute FIXME of perms.cc: Use constexpr instead of template specialisation.$$$bug #65641: view produces incorrect viewpoint when given a vector aligned with the primary axes.$$$bug #65637: short_disp doesn’t show ellipsis at end of long arrays.$$$bug #65617: clears editor setting Always show debug breakpoints and pointers....$$$bug #65577: Feature request: Editor right-click to run test without needing to clear %! first.$$$bug #65538: xint_value does not work the way it is intended.$$$bug #65531: iqr handling of empty inputs is not compatible.$$$bug #65499: Use separate GUI settings for Octave releases.$$$bug #65495: nchoosek error ‘gcd: all values must be integers’$$$bug #65459: uifigure (visible; off) temporarily creates a visible figure window.$$$bug #65447: jsonencode does not accept integer values larger than 999999.$$$bug #65441: nextpow2 incorrect for some inputs slightly larger than powers of two.$$$bug #65238: Improve nchoosek.m algorithm to prevent numerical issues.$$$bug #65221: movfun: Create inputParser only once.$$$bug #65176: unique.m: Enable third output with option stable.$$$bug #65134: griddata: Output size inconsistent for vector input interpolation points.$$$bug #65030: Column width in browser pane not retained between restarts of Octave.$$$bug #65010: Compatibility: Colormaps now default to 256 colors in Matlab.$$$bug #64995: Implement height and width as aliases for rows and columns for Matlab compatibility.$$$bug

Summary of bugs fixed for version 9.4.0 (2025-02-05):$$$Improvements and fixes$$$Fix error in calculation of file timestamp on Windows (bug #31080).$$$GUI$$$Update Spanish translation file (bug #66616).$$$Update Brazilian Portuguese translation file (bug #66554).$$$Build system / Tests$$$cellfun: Add test for function failing to return requested output (bug #66642).$$$arrayfun: Add test for function failing to return requested output (bug #66642).$$$Documentation$$$Clarify default editor difference between CLI and GUI (bug #66703).

Summary of bugs fixed for version 9.3.0 (2024-12-12):$$$Improvements and fixes$$$jupyter-notebook: Avoid error if evaluated code closes an existing figure window. Use unwind_protect block to ensure figure and tempfile cleanup actions occur (bug #65644).$$$uitable: Fix updating value with combo box selection (bug #65873).$$$movfun: Correct output shape of n-D arrays when operating on dimensions higher than 2. This will affect all moving window functions (movmad; movmax; movmean; movmedian; movmin; movprod; movstd; movsum; and movvar) (bug #65927).$$$movvar and movstd: Use correct default value of 0 if an empty array is specified for opt to maintain consistent behavior with var and std. Improve input validation to ensure opt is one of 0; 1; or []. (bug #66021).$$$fopen: Convert identifier for locale charset to lower-case characters (bug #65963).$$$regexp: Avoid integer underflow with unsigned integers.$$$interp2: Provide a meaningful error message if the input reference array is a vector (bug #66086).$$$struct2hdl.m: Do not delete previously added axes children (bug #66221).$$$print: Fix regression in pdf fonts (bug #66306).$$$legend icons now inherit the “facelalpha” property to match appearance of patch and surface graphics objects (bug #66314).$$$barh properties now better match equivalent bar plot. Changing horizontal property for a bar or barh plot now consistently updates properties to match redrawn plot (bug #65671).$$$Tick mark labels are now correctly updated after changing x or y data for bar and barh plots (bug #65734).$$$hdl2struct.m: Fix missing legend markers (bug #65904).$$$Enable FLTK graphics toolkit on macOS if not multi-threaded (bug #66482).$$$Do not execute PKG_ADD script in current directory (bug #66477).$$$GUI$$$Update Brazilian Portuguese translation file (bug #66337).$$$Fix some missing white spaces in documentation viewer (bug #66459).$$$Build system / Tests$$$jupyter-notebook test no longer fails if a figure window is open (bug #65644).$$$movmad; movmax; movmean; movmedian; movmin; movprod; movslice; movstd; movsum; movvar: Add BISTs for general function operation.$$$interp2; interp3; interpn: Add more tests for non-numeric inputs (bug #66086).$$$Update formatting of movXXX functions.$$$Documentation$$$load: Correct documentation of -v7.3 and -hdf5 options (bug #42530).$$$interp2; interp3; interpn: Specify that inputs need to be numeric (bug #66086).$$$error: Document the ID argument (bug #66015).$$$sqp: Clarify size requirements of input arguments lb and ub (bug #66421).$$$Eliminate double period at end of “mode” graphics property documentation.$$$Fix makeinfo errors for comment block keywords (bug #66428).

GNU Octave 9.2.0 Released$$$Jun 7; 2024$$$$$$GNU Octave version 9.2.0 has been released and is now available for download. An official Windows binary installer is available. For macOS see the installation instructions in the wiki.$$$$$$Fix for an overlinking issue for built .oct files. If you are concerned by overlinking; re-built binaries that you built with mkoctfile in version 9.1.0 with the new version.$$$Better support for HiDPI scaling.$$$Fixes for newer versions of Qt6.$$$Additionally; the Windows binaries are now built with Qt6.

GNU Octave 8.4.0 Released$$$Nov 5; 2023$$$$$$Octave Version 8.4.0 has been released and is now available for download. An official Windows binary installer is also available. For macOS see the installation instructions in the wiki.$$$$$$This version is a bug fixing release:$$$$$$Improvements and fixes$$$mkoctfile: Skip parsing for less flags (bug #64590).$$$Fix getting number of non-zero elements in SuiteSparse matrices.$$$delaunayn: Avoid erroneous simplex removal for integer inputs (bug #64658).$$$delaunayn: Add precision loss warning for large int inputs (bug #64658).$$$delaunayn: Also warn for negative integers with large magnitude (bug #64658).$$$Correctly load all-zero sparse matrices from text files (bug #64696).$$$Allocate correct storage for interleaved complex mxArray objects (bug #64687).$$$mkoctfile: Prefer parsing over passing on argument after unknown argument (bug #64725).$$$fft: Avoid segmentation fault with ND-arrays (bug #64729).$$$fft: Avoid overwriting input for inplace operations (bug #64733).$$$intersect.m: Correct third output with stable flag (bug #60347).$$$Avoid use-after-free issue in parser (bug #60882).$$$betainc.m: Improve integer input logic and validation (bug #64726).$$$webread: Apply timeout also for initial connection (bug #64826).$$$Build system / Tests$$$mkpp.m: Remove stray text causing error with demo code.$$$Ensure that new test for saving sparse matrices writes a text file.$$$Search for tests in and install octave-value template sources (bug #64696).$$$Documentation$$$Correct variable name typo in documentation of slice().$$$Expand documentation for gcd() and lcm().$$$datevec: Clarify that heuristics are used to detect the format of a date string (bug #64563).$$$datenum: Add help text that inputs can be arrays; add examples (bug #54679).$$$datevec: Add help text that inputs can be arrays (bug #54679).$$$Describe various pitfalls with floating point ranges (bug #64692).$$$Add range example and some copyedits.

GNU Octave 8.3.0 Released$$$Aug 8; 2023$$$$$$Octave Version 8.3.0 has been released and is now available for download. An official Windows binary installer is also available. For macOS see the installation instructions in the wiki.$$$$$$This version is a bug fixing release:$$$$$$Improvements and fixes$$$undo_string_escapes: Return \0 for NUL byte input (bug #64051).$$$Stop excessive memory usage for minor grid ticks (bug #64017).$$$Include needed headers in installed .cc files (bug #63922).$$$Fix floating point exception when sparse array reshaped to 0x0 (bug #64080).$$$unicode2native: Fix conversion to UTF-16 (bug #64139).$$$memory: Fix execution on Cygwin.$$$uniconv-wrappers: Avoid freeing uninitialized pointer (bug #64182).$$$lo-regexp.cc: Only include PCRE2/PCRE headers if feature test succeeded (bug #64172).$$$Fix segfault when saveobj() used with old-style class files (bug #64220).$$$Fix restoring breakpoints from dbstatus breakpoint structure (bug #64271).$$$colorbar.m: Fix undefined variable d (bug #64287).$$$Reset all axes properties except Units and Position with cla (reset).$$$unicode2native.m: Restore support for one input parameter (bug #64331).$$$native2unicode; unicode2native: Support empty second argument (bug #64331).$$$fullfile.m: Return empty cell for input with empty cell (bug #64377).$$$audioplayer.m: Emit error if input data Y is empty (bug #64379).$$$Remove obsolete callback constructor code for @audioplayer and @audiorecorder (bug #64379).$$$Avoid false detection of “non-real pivot” in ichol() (bug #63785).$$$Fix building against SuiteSparse 7.1.0 when SPQR is disabled (bug #64446).$$$ellipsoid.m: Avoid returning geometry data unless it is explicitly requested to avoid overwriting ans.$$$subsasgn: Avoid panic on invalid field names (bug #64213).$$$lex.ll: Remove $ as a valid character for identifiers.$$$fileparts.m: Allow for relative path on different drive on Windows (bug #64462).$$$datevec.m: Add some ISO 8601 conform formats (patch #9964).$$$Build system / Tests$$$betaincinv.m: Slightly increase test tolerance (for macOS 13).$$$colorbar.m: Add tests (bug #64287).$$$sparse-qr.cc: Fix compatibility with SuiteSparse 7.1.0 / CXSparse 4.0.4.$$$pcg.m: Use eps tolerance in test for ishermitian (bug #63787).$$$fileparts.m: Overhaul BISTs.$$$Documentation$$$Return documentation for help properties (bug #64067).$$$Add note to isprime.m on inputs over 2^64.

GNU Octave 8.2.0 Released$$$Apr 13; 2023$$$$$$Octave Version 8.2.0 has been released and is now available for download. An official Windows binary installer is also available. For macOS see the installation instructions in the wiki.$$$$$$This version is a bug fixing release:$$$$$$Improvements and fixes$$$fopen: Use “UTF-8” as default encoding for fopen (bug #63930).$$$fopen; unicode2native: Fix converting the encoding of short char arrays with invalid UTF-8 (bug #63930).$$$fopen: Try to gather complete UTF-8 surrogates when converting encoding (bug #63930).$$$fopen: Do not convert encoding for file streams with libc++ (bug #63930).$$$pr-output.cc: Fix output for format native-bit (bug #63940).$$$Fix evaluation of & and | expressions in conditional contexts (bug #63935).$$$Avoid clang warning about very unlikely buffer overflow.$$$mpoles.m: Overhaul function and use absolute tolerance for zero poles (bug #63937).$$$perms.m: Change unique output order to reverse lexicographic to match non-unique order (bug #63962).$$$Remove trailing \r from curl dir list (bug #63851).$$$Fix display of scalar complex variables with mixed Inf/NaN and floating point values (bug #63961).$$$Don’t use encoding facet when writing bytes to stream (bug #63931).$$$GUI$$$Speedup loading and saving preferences dialog (bug #63909).$$$Build system / Tests$$$inpolygon.m: Fix demo code (bug #63865).$$$if.tst: New test for (bug #63935).$$$acinclude.m4: Correct typo in #define PCRE2_CODE_UNIT_WIDTH.$$$lu: Add self-test with complex valued input.$$$Disable visibility flags by default (bug #61855; bug #63916).$$$Check whether using STL from LLVM or Apple (bug #63930).$$$Documentation$$$Improve documentation for linspace and logspace functions.$$$Correct and improve documentation for sparse() function.