double-conversion-3.3.1-160000.2.2

- update to 3.3.1:
  * Add _ITERATOR_DEBUG_LEVEL=2 and _DEBUG defines
  * Build system and CI tweaks
  * Add some missing headers

- update to 3.3.0:
  * Fix some compile warnings in Visual Studio
  * Set permissions for github workflows
  * Add flags to control trailing decimal and zero in exponent
    form when input has one significant digit

- update to 3.2.1
  * Disable quiet nan test on windows
  * Test on all platforms.
  * Fix warnings on Windows
  * Run ctests first.
  * Give shared-lib option and test install
  * Install Windows debbuger (pdb) files.
  * Add a cast to silence a signedness conversion warning.
  * Issue #184 : Fixed all -Wzero-as-null-pointer-constant warnings

- update to 3.2.0:
  * Fix quiet NANs on MIPS* and PA-RISC architectures.

- update to 3.1.7:
  * Reintroduce macros, if DOUBLE_CONVERSION_NON_PREFIXED_MACROS is set
  * Also add support for Synopsys ARC64 architecture

- update to 3.1.6:
  * Features some code cleanups.
  * Adds the following new architectures: loongarch, xtensa, nios2, e2k.

- Use proper cmake calls
- Do not pull in py2 that is EOL but use python3 for build

- Update to version 3.1.5 jsc#ECO-1190, jsc#SLE-11474
  * Fix `0x` for string->double conversion when Hex Floats are allowed.
    Avoid integer overflow when exponents for hex floats were too big.
    Update version number.
  * Fixed warning in gcc4.9.
  * Merged changes to install libraries in the correct place when
    using 64-bit libraries.
  * Use relative includes in the library. This shouldn't have any visible effect
    for consumers of the library.
  * Fix typo in test.
  * Fix separator characters when they they don't fit into 8 bits.
  * Check correctly for _MSC_VER.
  * Allow the library to be compiled for Emscripten.

- Update to version 3.0.3
  * Support RISC-V
  * Bug fixes
- Switch to cmake
- Use %license

- Fix RPM groups.
  Slightly trim/compact descriptions.