mathgl-8.0.1-bp156.2.43

- Update spec file to use openmpi-macros-devel and always use the
  default OpenMPI implementation.

- Also disable octave bindings for openSUSE >= 1699 (TW), as
  octave 8.x is not yet supported by mathgl.

- Disable octave bindings on Leap/SLE 15 due to missing SWIG 4.x.
- Use %cmake macros as far as possible.
- Fix underlinking of libmgl-mpi, requires libmgl.
- Spec file cleanup.

- Add mathgl-libharu2_4-compat.patch: Fix compilation against
  libharu 2.4.x [https://sourceforge.net/p/mathgl/bugs/48/].
- Build against wxWidgets >= 3.1 (wxGTK3) instead of
  wxWidgets=3.0.

- Update to version 8.0.1:
  * Increase accuracy at line segment skipping.
  * Changes in SOVERSION numbering.
  * Minor spelling fix.
- Changes from version 8.0:
  * Change version numbering according Debian rules.
  * Add accurate line segment and quadrangle/triangle face cutting
    at axis border crossing.
  * Add 'fastcut' for disable accurate but slow line/trig/quad
    primitives cutting at axis borders.
  * Restore back transparent background by default. You can set
    non-transparent background by using 'clf' or 'background'
    functions.
  * Compatibility changes for Qt5.15 and MSVC.
  * Change colorbar C-functions by adding last argument for
    options.
  * Add alpha-value in fill_background().
  * Disable Qt4 interface by default.
- Update mathgl-fix-python-module-path.patch to additionally
  correct the locations of installed python module files.
- Update shlib naming scheme in keeping with upstream so version
  changes.
- Build for multiple python versions (while ensuring stuff like
  docs, etc are only built once -- for the primary flavor).
- Add Requires: libmgl-mpi%{libversion} dependency on -devel
  package as it is linked to.
- Required version of swig is at least 4.0 to build octave
  bindings.
- Drop %{name}-rpmlintrc: No longer needed.

- Update to version 2.5:
  * Add functions ‘sum’, ‘prod’ at formula evaluation, like
    ‘sum(_i^2,5)’ will produce 30=0+1^2+2^2+3^2+4^2, and
    ‘prod(1+_i,5)’ will produce 120=5!. You can nest them for
    variables ‘_i’,‘_j’,...,‘_z’, like ‘sum(sum(_j+_i^2,5),5)’
    will give 200.
  * Add user-defined functions ‘fn1()’...‘fn9()’ at formula
    evaluation, like ‘fn1(3)\x^_1’ will produce x^3.
  * Add functions ‘value’, ‘spline’ at formula evaluation, like
    ‘value(a,5)’ will give value of 6-th element of a, i.e. a[5].
  * Add keep for keeping the data phase/sign or value along line i
    and j in given direction
  * Add mglODEs() (i.e. extend ode) for solving PDE, cascades and
    etc.
  * Add dcont to draw a set of curves, which are intersection of
    two isosurfaces at given level.
  * Add lines to draw a set of lines with arrows between points.
  * Add first, last to get array of first/last indexes of values
    larger specified one along given direction.
  * Add bernoulli, binomial, brownian, discrete, exponential,
    gaussian, shuffle, uniform, uniformint for data filling by
    random numbers with different distributions.
  * Add readbin to read data from binary files.
  * Add texparse to enable/disable TeX-like command parsing at
    text output.
  * Update ode: regularization is used if dt*tmax<0 (total time
    may differ, but number of steps still the same).
  * Add mgl_default_graph() – the global (default) pointer to HMGL
    for keeping plot settings, simplifying external scripts and
    capture it in GUI.
  * Add style ‘p’ in command smooth to use parabolic smoothing of
    arbitrary points.
  * Allow background for scaling image, centering image and
    tessellate image as mosaic.
  * Allow filling background by the specified color.
  * Allow masks for candle, area, region, bars, barh, chart, tape,
    face, polygon, ellipse, rhomb.
  * Allow single value specification for contour level in cont,
    contd, contp, contv, cont3, dcont, tricont.
  * Add mgl_dual_save_hdf(), mgl_real_save_hdf(),
    mgl_int_save_hdf() for saving single number in HDF5 file.
  * Add mgl_formula_calc(), mgl_formula_calc_c() for evaluating
    textual formulas of arbitrary set of data arrays.
  * Add mglDataList – global mglDataA* collection for
    viewing/changing in GUI data from external languages (like,
    Python, ...).
  * Add Armadillo interface for mglData and mglDataC.
  * If nsub parameter of command hist is negative (nsub<0) then
    the linear interpolation is used instead of spline one.
  * Add dark text colors for mgllab.
  * Minor bugfixes.
  * OBSOLETE: Qt4 interface will be disabled in the next release.
  * OBSOLETE: Functions colorbar will have extra argument opt for
    options, similarly to axis and other plotting functions. In
    C++ and SWIG-based interfaces this change "masked" by
    providing the default value for the argument.
- Minor rebase of mathgl-fix-python-module-path.patch to apply
  cleanly against updated version.
- Bump library so version appropriately.
- Enable python3 bindings: now supported by upstream; build only
  for default python3 flavour.
- Disable python bindings for Leap, since python >= 3.8 is
  required whereas Leap 15.2/3 only have python 3.6.
- Re-enable octave bindings.
- Rename python sub-package to have name consistent with
  openSUSE's python3 package naming guidelines.
- Update _constraints to require 8 GB disk space as required by
  updated version.
- Drop sed fix for location of numpy header file: no longer
  needed.

- Do not compile with python2 on Leap 15.4

- Remove the unneeded WebKit build requirements.
  '-Denable-json-sample=off' disables the WebKit dependency.

- Disable octave bindings: no longer compatible with any
  suppported octave version.
- Disable linking to texconfig zypper posttrans configuration
  script; this script is no longer installed by texlive.

- Fix for /usr/lib -> /usr/libexec migration.

- Add 4G min disk space default in _constraints

- Update to version 2.4.4:
  * Add minmax for positions of local maximums and minimums.
  * Add ContGen() and ContFGen() to draw contour lines manually.
    Correspondingly cont and contf commands are extended.
  * Add conts for coordinate of contour lines.
  * Extend put for partial filling if data size(s) are smaller as
    destination.
  * Classes mglFormula and mglFormulaC now can read file with
    given sizes if string have the format
    ":fname:x1:x2:y1:y2:z1:z2". Here data is supposed in range
    [x1,x2]*[y1,y2]*[z1,z2] (by default [0,1]*[0,1]*[0,1]). Calc()
    in this case return the interpolated values.
  * mglData now have operator() which work as SubData() function.
  * Minor bugfixes.
- Disable python bindings where python2 is unavailable (still no
  python3 support from upstream unfortunately).
- Drop conditionals for old, unsupported openSUSE versions.
- Enable octave bindings for openSUSE versions > 1500 (doesn't
  work with octave version in Leap 15.1, 15.2).
- Run spec-cleaner for minor specfile cleanups.
- Drop qt4 package definitions.
- Drop Group tags from packages.

- openmpi has been renamed to openmpi1 for openSUSE >= 1550;
  adapt accordingly.

- Use -ffat-lto-objects when building static libraries.

- Update to version 2.4.3:
  * Add 'clabel' command -- draw labels for colorbar. Should be
    used *after* drawing colorbar!
  * Extend 'ctick' command
  * Add subpixel smoothing for masks
  * Boxes around text (style '@') now use actual height and
    position of the text.
  * Add mask to EPS export. Note, mask angles are reduced to
    45*(0,1,...7) degrees for decreasing pattern size in the EPS.
  * Update default masks: '*' become dot, '^' become bricks, 'd'
    become plus, 'D' become tacks, ';' and 'j' change lengths.
  * Note, you can use brush.ods to prepare user-defined masks.
  * Add styles '^' and '_' for command 'smooth' to find
    upper/lower bound of the data.
  * Improve FlowP() to draw both branches (in positive and
    negative time direction).
  * Improve CGI interface and update website.
  * Introduce struct mdual as interface for C and C++ complex
    numbers. It is implicitly converted to std::complex<> in C++.
    And need to call c2mdual() and mdual2c() in pure C. However
    mdual is binary compatible with C _Complex numbers.
  * Add flag in CMake to manually disable support of C99 complex
    numbers.
  * Bypass user-specified extension in base font family name.
  * Improve hints in mgllab and udav.
  * Add utility 'mgltask' for making output file with a set of
    copies of mask-file. It useful for making set of initial
    conditions with a few parameters varied in specified range.
  * Add example of OpenGL output.
  * Bugfix for approximate min and max position.
  * Bugfix for position of SVG output.
  * Compatibility fixes for new versions of CMake, compilers and
    libraries.
  * INCOMPATIBLE: Formally pure C interface handle complex numbers
    by new way due to strict following of C/C++ standards.
    However, the MathGL library is binary compatible with previous
    version(s).
- Drop mathgl-fltk-includedir.patch: incorporated upstream.
- Fix Group for %{name}-doc-pdf to Documentation/Other, as
  Documentation/PDF isn't a valid group.

- Define MathGL_INSTALL_CMAKE_DIR to correct location

- Update to 2.4.2:
  * Add custom dialog for FLTK widgets. This can be constructed from
    C/C++ code, or from MGL script using Special commentaries.
  * Add progress for displaying the progress of execution (by ’#’
    symbols in console, or by progress-bar widget).
  * Add contp for contour lines on parametrically specified surface.
  * Add flow3 for flow threads of 3D vector field, which start from
    the plain.
  * Add variant of flow for drawing a flow thread, starting from
    specified point.
  * Add scaletext for disabling text size scaling in relative
    inplot-s (like columnplot, ...).
  * Add setup for setting all low-level flags.
  * Add coil for projecting periodical data.
  * Add modulo operation ‘%’ (x%y=fmod(x,y)) to the list of known
    operations in real-valued formulas.
  * Add style ‘.’ for cont to draw contour lines from saddle points.
  * Add style ‘.’ for flow to draw flow thread from saddle points.
  * Add option value in tube for setting number of edges manually.
  * Crop output of ode if NAN or INFINITY value occurs.
  * Remove style ‘~’ for plot and tens. Now, all curves try to omit
    points.
  * Add setup flag MGL_FULL_CURV which disable omitting points of curve.
  * Shift axis labels closer to axis.
  * Change size of LaTeX-like commands \sup, \sub to be the same as
    \overset, \underset.
  * Change ticks drawing in colorbar.
  * Add new section All samples of documentation with large set of
    samples for most of MathGL features.
  * Samples in documentation are generated automatically now
    (from sample.cpp).
  * INCOMPATIBLE: Make obsolete MGL command setsizescl. At this, the
    similar C/C++/Fortran/... functions still working.
  * Bugfixes.
- Rebase patch:
  * mathgl-examples-install.patch

- Have the -tex-doc subpackage conflict with texlive-mgltex-doc
  since they share several files

- Do not build with Qt4 on TW/15 as Qt WebKit 4 is not available
  (boo#1072033)
- Explicitly disable marking either Qt version as "qt" plugin:
  * mathgl-no-default-qt.patch

- Add Provides: mathgl for the libmathgl package, to make the
  mathgl-lang package installable.

- Adapt to latest openmpi packaging changes: setup the openmpi
  environment before calling cmake.

- Trim some irrelevant words from descriptions.

- Update to version 2.4.1:
  * Add beltc plot, which is belt with manual coloring.
  * Add style '~' for plot and tens to omit some points at output.
  * Add style ':' for axis to draw lines through point (0,0,0).
  * Miscellaneous bugfixes.
- Changes from version 2.4.0:
  * Add mgllab executable, which is FLTK based version of UDAV.
  * Add string manipulation in MGL language:
  - 'str'[n] - get string of only n-th symbol;
  - 'str'+v - add value v to the last character of the string;
  - 'str',val or 'str',!val - append numeric value to the string
  * Add time value to MGL language in format:
    'hh-mm-ss_DD.MM.YYYY' or 'hh-mm-ss' or 'DD.MM.YYYY'.
  * Add iris plot to show cross-dependencies of several data
    arrays.
  * Add flame2d to draw flame fractals.
  * Add bbox to set bounding box for 2D output.
  * Add section to get part of data between specified values.
  * Add detect to found curves along data maximums.
  * Add dilate to extend the region of 1-th value.
  * Add erode to narrow the region of 1-th value.
  * Add apde to solve PDE equation in inhomogeneous nonlinear
    media with spatial dispersion.
  * Add symbol to draw the glyphs defined by addsymbol.
  * Add addsymbol to declare user-defined symbol (or glyph), which
    can be used as markers for plot (with style '&') or drawn
    itself by symbol command.
  * Add openhdf to open all data arrays from HDF file.
  * Extend crop to cut to optimal size for FFT (i.e. to closest of
    2^n*3^m*5^l).
  * Add function mgl_data_max_first() and data suffixes
    .m[xyz][fl] to find first/last maximum along direction (see
    Data information).
  * Add function mgl_datac_diff_par() to parametric derivative of
    complex data (see diff).
  * Add style 'F' for bars and barh to set fixed bar widths.
  * Add style 'a' for plot, tens, area and region to accurate
    drawing of data, which partially lie out of axis range.
  * Add style '#' for region to draw wired plot.
  * Add possibility of manual shift in multiplot.
  * Add parsing arguments of options for MGL commands.
  * MGL command correl now can perform 2d and 3d correlations.
  * Option meshnum now change the number of drawn markers for
    plot, tens, step, mark and textmark.
  * Function step handle data with x.nx>y.nx similarly to bars.
  * Extend tile and tiles by adding manual coloring and face
    orientation.
  * Add variant of MGL command copy to copy data with "incorrect"
    names.
  * Improve tick labels drawing.
  * Improve time-ticks (add weeks) and add subticks to its.
  * Improve fplot to handle several singularities.
  * Add LaTeX command \dfrac{}{}. This is full-sized version of
    \frac{}{}. Unfortunately, there is no support of nesting.
    Contrary, \frac allow nesting.
  * Add mglODEc() - ODE solver for complex variables (see ode).
  * Add cmplx(a,b)=a+i*b to the list of known functions in formula
    parsing
  * Update CMake find_package to find MathGL.
  * Increase line width for wired text.
  * Update documentation: add description of new commands, add
    hint Mixing bitmap and vector output.
  * Add translation to Russian for most of labels and messages.
  * Name all are reserved in MGL scripts now as synonym of -1.
  * INCOMPATIBLE: Replace libmgl-qt by libmgl-qt4 and libmgl-qt5
    for resolving conflicts of simultaneously installed both Qt4
    and Qt5.
  * Minor bugfixes.
- Drop mathgl-oldtexinfo.patch: fixed upstream.
- Add mathgl-fltk-includedir.patch to fix include directives for
  fltk headers.
- Split out a -lang package.
- Package the binary and desktop file for  mgllab as part of the
  - udav package; their desktop files share the same icon
  unfortunately.
- Disable doc for openSUSE <= 1320; no longer builds with old
  texi2html there.
- Move AUTHORS, ChangeLog.txt, README, COPYING to -devel package
  for all openSUSE versions; package it using %%doc.
- Add _constraints file to require at least 2 GB of physical
  memory on obs; otherwise builds on 64-bit openSUSE > 1320 fail.

- Update to version 2.3.5.1:
  + Library libmgl-qt is removed. You need to use libmgl-qt4 or
    libmgl-qt5 explicitly now.
  + Compatibility changes for latest MS VisualStudio.
  + Bugfixes.
- Changes from version 2.3.5:
  + See Changelog.txt for list of changes.
- Rebase mathgl-examples-install.patch for updated version.
- Define a macro for deciding whether to enable octave bindings,
  but disable octave bindings for all openSUSE versions for now
  as mathgl fails to build with updated octave. This is a
  temporary workaround, and once the build issue is fixed, we
  should revert back to enabling octave for recent openSUSE
  versions. The "enable_octave" macro is defined with this in
  mind.
- Add rpmlintrc file to suppress warning of missing man pages for
  the example binaries. These just clutter up the rpmlint log and
  have no practical use.

- Drop dependency on swig3 as swig in leap got updated to 3.x series

- Update to version 2.3.4
  * Minor fixes and improvements, see included ChangeLog.txt for details
  * Update mathgl-examples-install.patch
  * Update mathgl-texmf-dir.patch
  * Drop obsolete getopt.patch
  * Drop obsolete mathgl-cmake-3.2.patch

- Fix octave-mathgl package dependencies

- Disable json examples
- Enable Octave module and Qt5 support for openSUSE Leap 42.1
- Use wxWidgets 3 for openSUSE Leap 42.1

- Enable Octave module for openSUSE > 13.2

- getopt.patch: Fix use of getopt
- Compile with %{optflags}
- Fix %postun -n %{libname}-wx%{libversion}

- Update to version 2.3.3
  * Minor fixes and improvements, see included ChangeLog.txt for details
  * Update mathgl-examples-install.patch
  * Update mathgl-texmf-dir.patch
- Fix build with old texinfo
  * mathgl-oldtexinfo.patch

- Disable Octave module due to swig incompatibility with Octave 4.0

- Add mathgl-cmake-3.2.patch: Fix build with cmake 3.2; X11 is no
  longer implicit linked.

- Build with MPI support

- Use texinfo for documentation generation on openSUSE > 13.2