Three potential issues in jsoncons

[yimingming]

Hello experts,
Based on the static coding analysis for jsoncons, it seems that the following issues need your double check.

1. in jsoncons/include/jsoncons/detail/write_number.hpp ,
   line 366: char buffer[100]; it is not initialized.
   line 433: char buffer[100]; it is not initialized.
2. in jsoncons/include/jsoncons/bigint.hpp,
   line 211: The constructor allocates field data_ of jsoncons::basic_bigint<std::allocator >::dynamic_storage but there is no destructor. Resource leak.
   Could you please check if it has the potential risk. Thanks for your feedback in advance.

[danielaparker]

1. in jsoncons/include/jsoncons/detail/write_number.hpp ,
   line 366: char buffer[100]; it is not initialized.
   line 433: char buffer[100]; it is not initialized.
2. in jsoncons/include/jsoncons/bigint.hpp,
   line 211: The constructor allocates field data_ of jsoncons::basic_bigint<std::allocator >::dynamic_storage but there is no destructor. Resource leak.
   Could you please check if it has the potential risk. Thanks for your feedback in advance.

Regarding (1), it's just a buffer of 100 chars, passed to another function that needs a temporary work space, initialization would serve no purpose, one initial value is as good as any other, they're never examined. We could initialize buffer[0] to 0, if that would make the analyzer happy.

Regarding (2), there is no resource leak in a bigint object, all memory is allocated and deallocated, but there is an asymmetry, the memory is allocated in the constructor of dynamic_storage (which is a data member of bigint), and deallocated in the destructor of bigint (via a function call on the dynamic_storage member.) While that's not a risk, it's not good design either, we should change it, thanks for identifying this issue.

[yimingming]

Danielaparker,
Thanks for your quick feedback.