leptonica-1.75.3-bp153.1.28

- Update to 1.75.3:
  * See changes in the version-notes.html file.
  * Fixed a stack based buffer overflows in gplotRead() and ptaReadStream()
    when parsing crafted files can lead to denial of service
    (CVE-2018-7186 bsc#1081576)
  * Fixed a buffer overflow in pixHtmlViewer in prog/htmlviewer.c
    (unsanitized input (rootname)), which could potentially lead to
    arbitrary code exeuction. (CVE-2018-7247 bsc#1081631)

- Fix RPM groups.

- Update to 1.75.1.
  * See changes in the version-notes.html file.
- Add leptonica-license.txt, version-notes.html, moller52.jpg to
  docs of the liblept package.

- Update to 1.74.4.
  * See changes in the version-notes.html file.
- Add BuildRequires: gnuplot.

- Update to version 1.74.1
  Version 1.74.1 4 Jan 17
  * Configuration changes to support the patch number in the
    version (major.minor.patch).
  * Removed all remaining pixDisplayWrite() calls in prog/.
  * Cleaned up and/or promoted about 15 programs to full regression
    tests.  There are now 95 tests in the regression set.
  * Over half the initial coverity scan warnings have been removed.
  Version 1.74.0 10 Dec 16
  * Leptonica development was moved to github.  The master is at:
    github.com/danbloomberg/leptonica
  * New modes for RGB --> gray conversion.
  * New functions added for displaying a pix from a pixa.
  * Split out sort/hash/set/map functions for dna, sarray and pta.
  * More robust horizontal deskew on multi-column page images.
  * Improve webpio_reg test.
  * Remove X11 display for gplot; it is no longer supported.
  * Remove most sleep calls, which were put in for gplot; no longer needed.
  * Removed use of gthumb in library.
  * Removed use of pixDisplayWrite() in the library; still in some progs.
  * Improved test for endianness in makefile.static; no longer requires
    any local files or building and running a program.
  * Modified all files for doxygen output (spearheaded by Jurgen Buchmuller)
  * Improved plotting of the boxes in a boxa.
  * Replaced the slow point hash function with a simple fast one.
  * Added pam (4 component) format writing to pnmio.c (Jurgen Buchmuller)
  * Improved rendering of pixa in side/by/sides.
  * Better utilities for pixa and pixacomp.
  * Add read/write serialization functions from/to memory for all
    the major data structures that do not already have them.
  * More serialized boot recognizers stored as self-generating code.
  * Cleaned up generating an adapted recognizer from the boot recognizer.
  * Simplified temp file naming; removed most instances of named
    temp files from non-debug code; use tmpfile() and a wrapper
    l_makeTempFilename().
  * Simplify and streamline multipage tiff reading (Jeff Breidenbach).
  * Improvement of Otsu thresholding.
  * Recognize outstanding contributors to leptonica over the years.
  * New gif mem read/write interface that avoids writing a temp file,
    contributed by Tobias Peirick.
  * Use double arrays (dna) instead of float (numa) for set ops.
  * Cleanup of gray quantization functions and tests.
  * Refactored connectivity-conserving operations, to make them more useful.
  * Provided methods for measuring and regularizing the width of strokes.
  * Removed viewfiles.c from library; code is now in prog htmlviewer.c.
  * Better debugging in page segmentation (pageseg.c)
  * Deprecated the pixDisplayWrite*() debugging methods.
  * Added about 15 regression tests to the framework in alltests_reg.c
  * Final mods for compatibility with tesseract 4.00.

- Update to 1.73.
  * All lept_* functions have been rewritten to avoid path rewrites
    for output to temp files, which were introduced in 1.72.
  * Added grayscale histogram functions that can be used to compare
    images.
  * Added functions to determine if an image region has horizontal
    text lines.
  * Added functions to compare photo regions of images to determine
    if they're essentially the same.
  * Added red-black tree utility functions to implement maps and
    sets.
  * Improved security of tiff and gif reading, to prevent memory
    corruption when reading bad data.
- Change major library version to 5.
- Change programs prefix from "leptonica" to "lept" (make the names
  simpler to use).

- Update to 1.72:
  * Better handling of 1 bpp colormap read/write with png so that
    they are losseless. The colormap is always removed on read and
    the conversion is to the simplest non-cmapped pix that can
    fully represent the input -- both with and without alpha.
  * Fixed overflow bug in pixCorrelationBinary().
  * Fixed orientation flags and handling of 16 bit RGB in tiff.
  * Also new wrappers to TIFFClientOpen(), so we no longer go
    through the file descriptor for memory operations.
  * Improvements in the dewarp functions.
  * New box sequence smoothings.
  * New antialiased painting through mask; previously it was only
    implemented for connected components in a mask.
  * Better error handling and debug output with jpeg2000
    read/write.
  * Implemented base64 encoding. This allows binary data to be
    represented as a C string that can be compiled. Used this in
    bmf utility.
  * Implemented automatic code generation for deserialization from
    compiled strings (stringcode.*)
  * Regression tests write to leptonica subdir of in windows; in
    unix it is optional.  This avoids spamming the directory.
  * Added new colorspace conversions (XYZ, LAB).
  * New source files: encoding.c, bmfdata.h, stringcode.c,
    stringcode.h, bootnumgen.c.
  * Removed source files: convolvelow.c, graymorphlow.c
  * New programs: genfonts_reg, colorize_reg, texturefill_reg,
    autogentest1, autogentest2.
  * alltests_reg now has 66 tests.
- Added new build requirement openjpeg2-devel for openSUSE > 13.1.

- Update to 1.71.
  * This version supports tesseract 3.0.4. In particular, 3.0.4 has
    automatic conversion of a set of scanned images, either in a
    directory or coming directly from a scanner, into pdf with
    injected text. This is something we've wanted to do for several
    years!
  * Improved jp2k header reading, including resolution.
  * Removed src files: rotateorthlow.c, pdfio.c, pdfiostub.c.
  * Renamed jp2kio.c, jp2kiostub.c ==> jp2kheader.c,
    jp2kheaderstub.c. These header reading functions parse the jp2k
    files, and don't require a jpeg2000 library.
  * New jp2kio.c, jp2kiostub.c, that uses openjpeg-2.X to read and
    write jp2k. We now support I/O from these formats: png, tiff,
    jpeg, bmp, pnm, webp, gif and jp2k as well as writing to
    PostScript and pdf.
  * New pdfio1.c, pdfio1stub.c, pdfio2.c pdfio2stub.c, where we've
    split functions into high and low level.
  * Fixed memory bug in bilateral.c.
  * Improved reading/write of binary data from file. For example,
    l_binaryReadStream() can now be used to capture data piped in
    via stdin.
  * Font directory now arg passed in everywhere (not hardcoded).
  * Don't write temporary files to /tmp; only to a small number of
    subdirectories, to avoid spamming the /tmp directory. E.g., for
    regression tests, the current output is now to /tmp/regout/.
  * For jpeg reading modify pixReadJpeg() to take as a hint a bit
    flag that allows extraction of only the luminance channel.
  * Allow wrapping of pdf objects from png images without
    transcoding.
  * Better support for alpha on read/write with png, including 1
    bpp with colormap, alpha (supported in png with transparency
    array).

- Update to 1.70.
  * New bilateral filtering.
  * New simple character recognition utility.
  * Improved dewarping functionality, in model building and
    rendering. More flexible use of ref models.
  * Better and more consistent handling of alpha layer in RGBA,
    though use of the spp field.  Ability to handle more png files
    with alpha, including palette with alpha.
  * New fast converters from jpeg and jpeg2000 to pdf, without
    transcoding.
  * Made bmp reader (and pix reading in general) more robust; avoid
    size overflow errors.
  * New text labelling operations; depth conversion of a set of
    images.
- Major version changed to 4.
- Rename liblept-devel to leptonica-devel.
- Drop liblept-1.69-no_return and leptonica-1.69-giflib5 patches
  (fixed upstream).
- BuildRequires: pkgconfig(libwebp) >= 0.2.0
- Correct URL and a source URL.

- fix build with giflib-5.x, leptonica-1.69-giflib5.patch

- Add liblept3-1.69-no_return.patch: fix
  no-return-in-nonvoid-function error. The bug reported upstream:
  http://code.google.com/p/leptonica/issues/detail?id=69.

- Update to 1.69.
  * Major version changed to 3.
  * New copyright (BSD, 2 clause) on src files.
  * Fixed bug in PDF generation for large files, using a new double
    array (dnabasic.c).  Added several new modes for PDF generation
    from sets of images.
  * Dewarp based on image content now aligns to left and right
    margins; works at book level; is more robust to bad disparity
    models; version 2 serialization.
  * Improved sorting efficiency of numas in cases where binning,
    which is order N, makes sense.
  * Fixed fpix serialization (now version 2).
  * New version (5) of xtractprotos, allows putting prototypes
    in-line in allheaders.h.  Having them separately in
    leptprotos.h still an option.
  * New src files: boxfunc4.c coloring.c, dnabasic.c.
- Build requires libjpeg-devel.

- Add "leptonica-" prefix to tools.
- Don't install test programs.
- Correct leptonica-tools summary and description.

- Add libtool as buildrequire to avoid implicit dependency.
- Use make_install macro.

- Dropped libgif patch (no need).
- Build against libwebp in openSUSE >= 12.1.
- Removed useless provides liblept-devel (no need anymore).
- Added license file to leptonica-tools package.

- Don't build versioned devel and tools packages.
- Use full URL for Source.
- Corrected License tag.

- Fix dependencies.
- Run autoreconf before configure.

- Devel package renamed to liblept2.

- Update to 1.68.
- Library major version changed to 2.
- Dropped no_return patch (fixed upstream).
- Added baselibs.conf as a source.
- libgif patch refresh.

- Don't build static library.