|
468 | 468 | </style> |
469 | 469 | </head> |
470 | 470 | <body> |
| 471 | +<h2>0.97.1 Release</h2> |
| 472 | + |
| 473 | +<ul class="task-list"> |
| 474 | +<li>"Transforming JSON with filters" example fixed</li> |
| 475 | +<li>Added a class-specific in-place new to the json class that is implemented in terms of the global version</li> |
| 476 | +<li>Reorganized header files, removing unnecessary includes. </li> |
| 477 | +<li>Incorporates validation contributed by Alex Merry for ensuring that there is an object or array on parse head.</li> |
| 478 | +<li>Incorporates fix contributed by Milan Burda for “Switch case is in protected scope” clang build error</li> |
| 479 | +</ul> |
| 480 | + |
471 | 481 | <h2>0.97 Release</h2> |
472 | 482 |
|
473 | 483 | <ul class="task-list"> |
@@ -496,7 +506,7 @@ <h2>0.96 Release</h2> |
496 | 506 | <li><p>Only json arrays now support <code>operator[](size_t)</code> to loop over values, this is no longer supported for <code>json</code> objects. Use a json object iterator instead.</p></li> |
497 | 507 | <li><p>The virtual methods <code>do_double_value</code>, <code>do_longlong_value</code> and <code>do_ulonglong_value</code> of <code>json_input_handler</code> and <code>json_outputhandler</code> have been renamed to <code>do_float_value</code>, <code>do_integer_value</code>, and <code>do_unsigned_value</code>, <br> |
498 | 508 | and their parameters have been changed from <code>double</code>, <code>long long</code>, and <code>unsigned long long</code> to typedefs <code>float_type</code>, <code>integer_type</code>, and <code>unsigned_type</code>.<br> |
499 | | -The rationale for this change is to allow different configurations for internal number types.</p></li> |
| 509 | +The rationale for this change is to allow different configurations for internal number types (reversed in 0.97.)</p></li> |
500 | 510 | </ul> |
501 | 511 |
|
502 | 512 | <p>General changes</p> |
@@ -860,7 +870,7 @@ <h2>Release 0.71</h2> |
860 | 870 | <h2>Release 0.70</h2> |
861 | 871 |
|
862 | 872 | <ul class="task-list"> |
863 | | -<li><p>Since 0.50, jsoncons has used snprintf for default serialization of double values to string values. This can result in invalid json output when running on a locale like German or Spanish. The period character (‘.’) is now always used as the decimal point, non English locales are ignored.</p></li> |
| 873 | +<li><p>Since 0.50, jsoncons has used snprintf for default serialization of double values to string values. This can result in invalid json output when running on a locale like German or Spanish. The period character (‘.’) is now always used as the decimal point, non English locales are ignored.</p></li> |
864 | 874 | <li><p>The output_format methods that support alternative floating point formatting, e.g. fixed, have been deprecated.</p></li> |
865 | 875 | <li> |
866 | 876 | <p>Added a template method as_vector to the json class. If a json value is an array and conversion is possible to the template type, returns a std::vector of that type, otherwise throws an std::exception. Specializations are provided for std::string, bool, char, int, unsigned int, long, unsigned long, long long, unsigned long long, and double. For example</p> |
|
0 commit comments