As per discussion in - https://github.com/open-source-parsers/jsoncpp/issues/404
Null should not be pass to memcmp, it may show undesired behaviour, so avoid doing that using assertion.
Also, changed one direct "assert" to JSON_ASSERT - it will be decided if exceptions are used or not.
See https://github.com/open-source-parsers/jsoncpp/issues/411#issuecomment-180974558
I was unable to produce a warning in Clang, so I am not certain. But based on a [SO answer](http://stackoverflow.com/questions/25480059/gcc-conversion-warning-when-assigning-to-a-bitfield), I think I've fixed the following:
```
/tmp/jsoncpp/src/lib_json/json_value.cpp: In copy constructor 'Json::Value::CZString::CZString(const Json::Value::CZString&)':
/tmp/jsoncpp/src/lib_json/json_value.cpp:235:18: error: conversion to 'unsigned char:2' from 'unsigned int' may alter its value [-Werror=conversion]
storage_.policy_ = (other.cstr_
~~~~~~~~~~~~
? (static_cast<DuplicationPolicy>(other.storage_.policy_) == noDuplication
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
? noDuplication : duplicate)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~
: static_cast<DuplicationPolicy>(other.storage_.policy_));
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
```
* Add move constructor to Value::CZString
* Add unit test for Value move constructor
* Allow includer to specify in advance the value for
JSON_HAS_RVALUE_REFERENCES
In value.h, ValueConstIterator can convert to ValueIterator, I think that is a bug. the correct way is ValueIterator can convert to ValueConstIterator.
Based on a patches to CMake by:
Ådne Hovda <ahovda@openit.com>:
commit 7b1cdb00279908cacabada92f8a53e4986465423
jsoncpp: Provide 'isfinite' implementation on older AIX and HP-UX
Newer AIX and HP-UX platforms provide 'isfinite' as a <math.h> macro.
Older versions do not, so add the definition if it is not provided.
Michael Scott <michael.scott@gbgplc.com>:
commit 9217b678b305d7df7471ba476a81bf28961fdfa3
jsoncpp: Provide 'isfinite' impl on more HP-UX versions (#15576)
Some versions of HP-UX do not define 'isfinite' or 'finite' in math.h
for Itanium when preprocessing with C++, so we have to add the
definition ourselves instead to map to the internal version.
Joerg Sonnenberger <joerg@bec.de>:
commit 75644dafe54c21902f14cfe58cb8338b553b69d8
jsoncpp: Fix compilation as C99 on Solaris
In C99 mode, Solaris variants may already define isfinite, so check for
the existence first.
At all 3 places isMultiLine is checked in for loop :
for (int index = 0; index < size && !isMultiLine; ++index) {
It means !isMultiLine is always true (otherwise do not enter loop), so || condition does not depend on isMultiLine, so removed that.
Introduce 'allowSpecialFloats' for readers and 'useSpecialFloats' for writers, use consistent macro snprintf definition for writers and readers, provide new unit tests for #209
* Clean up closing statements for if conditions, functions, macros,
and other entities. Newer versions of CMake do not require you to
redundantly respecify the parameters to the opening arguments.
We had already fixed Value to hold UTF-8 properly, but only the newer
StreamWriter was writing UTF-8 properly.
Old FasterWriter etc. were using asCString() instead of asString() in
Value::writeValue().
Hopefully this change does not break any existing code. Seems unlikely.
issue #240
This reverts commit 1c58876185.
std::snprintf() is only available in C++11, which is not provided by
all compilers. Since the C library snprintf() can easily be used as a
replacement on Linux systems, this patch changes jsoncpp to use the C
library snprintf() instead of C++11 std::snprintf(), fixing the build error
below:
src/lib_json/json_writer.cpp:33:18: error: 'snprintf' is not a member of 'std'
See #231, #224, and #218.
* support zeroes in string_
* support zeroes in writer; provide getString(char**, unsigned*)
* valueToQuotedStringN(), isCC0(), etc
* allow zeroes for cpptl ConstString
* allocated => non-static
By not calling validate(), we can add
non-invasive features which will be simply ignored when user-code
is compiled against an old version. That way, we can often
avoid a minor version-bump.
The user can call validate() himself if he prefers that behavior.
This is an improper solution. If multiple Readers exist,
then the effect stackLimit is reduced because of side-effects.
But our options are limited. We need to address the security
hole without breaking binary-compatibility.
However, this is not likely to cause any practical problems because:
* Anyone using `operator>>(istream, Json::Value)` will be using the
new code already
* Multiple Readers are uncommon.
* The stackLimit is quite high.
* Deeply nested JSON probably would have hit the system limits anyway.
Add indented_ as a bitfield. (Verified that sizeof(StyledStreamWriter)
remains 96 for binary compatibility. But the new symbol requires a minor
version-bump.)
Tests are currently failing when git cloning on Windows with autocrlf = true. In
that setup multiline comments contain \r\n EOLs. The test code assumes that
comments contain \n EOLs and opens the .actual files (etc.) with "wt" which
converts \n to \r\n. Thus we end up with \r\r\n EOLs in the output, which
triggers a test failure.
Instead we should cannonicalize comments while reading so that they contain only
\n EOLs. This approach simplifies other parts of the reader and writer logic,
and requires no changes to the test. It is a breaking change, but probably the
Right Thing going forward.
This change also fixes dereferencing past the end of the comment string in
StyledWriter::writeCommentBeforeValue.
Tests should be added with appropriate .gitattributes for the input files to
ensure that we run tests for DOS, Mac, and Unix EOL files on all platforms. For
now this change is enough to unblock Windows builds.
issue #116
The TARGET_INCLUDE_DIRECTORIES from inside the
IF(JSONCPP_WITH_CMAKE_PACKAGE) block was removed, since it only needs to
be set once.
In addition the CMAKE_VERSION check was simplified.
Fixed two build issues:
- JsonCPP currently doesn’t compile for Solaris due to platform
differences with ‘isfinite’ function. Fixed by adding proper include
and define for Solaris.
- JsonCPP currently doesn’t compile for GCC version 4.1.2 and earlier
due to use of ‘-Werror=*’ compile flag, which was introduced in a later
version. Fixed by adding version check to only add this flag on
supported versions of GCC.
Fixed two build issues:
- JsonCPP currently doesn’t compile for Solaris due to platform
differences with ‘isfinite’ function. Fixed by adding proper include
and define for Solaris.
- JsonCPP currently doesn’t compile for GCC version 4.1.2 and earlier
due to use of ‘-Werror=*’ compile flag, which was introduced in a later
version. Fixed by adding version check to only add this flag on
supported versions of GCC.
BUILD_SHARED_LIBS is a standard CMake argument that serves the purpose
of the custom JSONCPP_LIB_BUILD_SHARED. For now we force JSONCPP_LIB_BUILD_SHARED
to true if BUILD_SHARED_LIBS was defined.
Workaround for #51.
Newlines from comments separated by lines are retained when comments
are appended, so adding a newline between separate comments for a
node is not needed.
This is a patch that we have utilized at IDEXX Labs for the the bug described above.
We have tested and verified this on x86 32 and 64 bit linux and 32 bit arm.
`snprintf()` will use the current `LC_NUMERIC` locale
for converting a double to a string,
which will use a `,` instead of a `.` in some locales (e.g. de_DE).
`std::stringstream` allows setting the locale to `"C"` to always get a `.`.
This occurs only for that `stringstream` instance; no global is
altered.
for (int index = 0; index < size && !isMultiLine; ++index)
In addition to dead code, in the above if condition checking to !isMultiLine is of no use as it will be always true and hence "for" depends only on condition [index < size.]
The mentioned test case works fine in this case also.
