libjansson-2.14-160000.2.2

- Update to 2.14:
  * New Features:
    + Add `json_object_getn`, `json_object_setn`, `json_object_deln`, and the
    corresponding `nocheck` functions.
  * Fixes:
    + Handle `sprintf` corner cases.
  * Build:
    + Symbol versioning for all exported symbols (gh#akheron/jansson#523).
    + Fix compiler warnings.
  * Documentation:
    + Small fixes.
    + Sphinx 3 compatibility (gh#akheron/jansson#530).
- Use GitHub as source URLs: Release hasn't been uploaded to digip.org.
- Cleanup with spec-cleaner.
- Add check section.

- Update to 2.13.1
  * New Features:
  - Add jansson_version_str() and jansson_version_cmp() for runtime
    version checking
  - Add json_object_update_new(), json_object_update_existing_new()
    and json_object_update_missing_new() functions
  - Add json_object_update_recursive()
  * Bug fixes:
  - Add infinite loop check in json_deep_copy()
  - Enhance JANSSON_ATTRS macro to support earlier C standard(C89)
  - Update version detection for sphinx-build
  * Documentation improvements

- update to new upstream release 2.12
  * Bug fixes:
  - Fix error message in `json_pack()` for NULL object (#409).
  - Avoid invalid memory read in `json_pack()` (#421).
  - Call va_end after va_copy in `json_vsprintf()` (#427).
  - Improve handling of formats with '?' and '*' in `json_pack()`
    (#438).
  - Remove inappropriate `jsonp_free()` which caused
    segmentation fault in error handling (#444).

- update to new upstream release 2.11
  * New features:
  - Add `json_pack()` format specifiers s*, o* and O* for values
    that can be omitted if null (#339).
  - Add `json_error_code()` to retrieve numeric error codes
    (#365, #380, #381).
  - Enable thread safety for `json_dump()` on all systems.
    Enable thread safe `json_decref()` and `json_incref()` for
    modern compilers (#389).
  - Add `json_sprintf()` and `json_vsprintf()` (#393).
  * Bug Fixes:
  - Fix incorrect report of success from `json_dump_file()` when
    an error is returned by `fclose()` (#359).
  - Make json_equal() const-correct (#344).
  - Fix incomplete stealing of references by `json_pack()` (#374)
- changelog from 2.9
  * New features:
  - Add JSON_EMBED encoding flag allowing arrays and objects to
    be encoded into existing streams (#329).
  - Add `json_dumpb()` function for dumping to a pre-allocated
    buffer (#328).
  - Add `json_dumpfd()` and `json_loadfd()` functions for dumping
    to streaming file descriptors (#328).
  - Add support for parsing buffers larger than 2GB (#309).

- update to new upstream release 2.9
  * New features:
    Add json_auto_t to automatically decref a value that goes out of scope.
    Available only on GCC and Clang. (#301)
  * Small documentation fixes.
- changelog from 2.8
  * New features:
  - Always preserve insertion order of object items. json_object_iter()
    and friends, json_object_foreach() and json_dumps() and friends now
    always work in the insertion order of object items (#293).
  - Add json_object_foreach_safe() macro that allows json_object_del()
    calls during iteration (#230).
  - Add json_get_alloc_funcs() to allow reading the allocation functions
    set by json_set_alloc_funcs() (#262, #264).
  - Add json_pack() format specifiers s?, o? and O? for values that can
    be null (#261, #270).
  * Bug fixes:
  - Fix a crash when parsing inputs consisting of very deeply nested
    arrays or objects (#282, #284).
  - Never convert numbers to integers in the parser when
    JSON_DECODE_INT_AS_REAL is set.
  - Fix a use-after-free in json_pack() error handling.
  - Fix subnormal number parsing on mingw32.
  - Handle out-of-memory situations gracefully in the hashtable
    implementation

- Update to new upstream release 2.7
  * The json_pack et al functions recognize the specifiers "s%", "+%"
  and "+#"
  * Add length-aware string constructors "json_stringn()" and
  "json_stringn_nocheck()", length-aware string mutators
  "json_string_setn()" and "json_string_setn_nocheck()",
  and a function for getting string's length "json_string_length()"
  * Support ""\u0000"" escapes in the decoder. The support can be
  enabled by using the ""JSON_ALLOW_NUL"" decoding flag
  * Add "json_boolean_value()" as an alias for "json_is_true()"
  * Add JSON_REAL_PRECISION encoding flag/macro for controlling real
  number precision
  * Define the maximum indentation as JSON_MAX_INDENT.
  * Fix cases when some malformed ""\uNNNN"" escapes could crash
  the decoder with an assertion failure
  * Avoid integer overflows with very long strings in UTF-8 decoder
  and hashtable
  * Check for *NULL* key in "json_object_get()"
  and "json_object_del()"
  * "json_unpack()": Allow mixing JSON_STRICT with optional keys.
  * Fix int/int32 mismatch
  * Parse subnormal numbers correctly

- Add baselibs.conf

- Update to new upstream release 2.6
  * Fixed CVE-2013-6401: avoid bad performance in large JSON objects
  * The json_pack et al functions recognize the specifiers "s#", "+"
  and "+#"
  * Add the "JSON_DECODE_INT_AS_REAL" flag
  * Add "json_array_foreach()" function
  * Add the "json_boolean()" macro that returns the JSON true or
  false value based on its argument
  * Add "json_load_callback()" that calls a callback function
  repeatedly to read the JSON input.
  * Add JSON_ESCAPE_SLASH encoding flag to escape all occurences of
  '/' with '\/'.
- Drop defunct gpg-offline and use local source_validator instead

- Verify GPG signature.

- Update to new upstream release 2.3.1
  * Add support for optional object keys with the "{s?o}" syntax
  in json_unpack and the like
  * Add json_object_update_existing() and json_object_update_missing()
  for updating only existing keys or only adding missing keys to an
  object.
  * Add json_object_foreach() for more convenient iteration over
  objects.
  * When decoding JSON, write the number of bytes that were read from
  input to "error.position" also on success. This is handy with
  JSON_DISABLE_EOF_CHECK.
  * Add support for decoding any JSON value, not just arrays or
  objects. The support is enabled with the new "JSON_DECODE_ANY"
  flag.
  * Avoid problems with object's serial number growing too big.
  * Remove "+" and leading zeros from exponents in the encoder.
  * Decoding functions now return NULL if the first argument is NULL.