`valueToString` takes an argument `unsigned int precision`, but it is used with `%d` rather than `%u` in the `snprintf` format string. Make the format string look for an unsigned value instead.
GCC 7, when compiling with -Wimplicit-fallthrough=1 or higher, issues a warning which can be suppressed using a comment that matches certain regular expressions. The comment change does just that: signal to GCC that the fall through is intentional.
Fixes#676
* Un-deprecate removeMember overloads, return void
Sometimes we just want to remove something we don't need anymore. Having
to supply a return buffer for the removeMember function to return something
we don't care about is a nuisance. There are removeMember overloads that
don't need a return buffer but they are deprecated. This commit un-deprecates
these overloads and modifies them to return nothing (void) instead of the
object that was removed.
Further discussion: https://github.com/open-source-parsers/jsoncpp/pull/689
WARNING: Changes the return type of the formerly deprecated removeMember
overloads from Value to void. May break existing client code.
* Minor stylistic fixes
Don't explicitly return a void value from a void function. Also, convert
size_t to unsigned in the CZString ctor to avoid a compiler warning.
Without value_type, Boost.Test version 1.65.0 throws a compiler error when a Json::Value object is compared to another with BOOST_TEST. Example and further discussion are in https://github.com/open-source-parsers/jsoncpp/issues/671.
Hi Christopher,
Thank you for reaching out and sorry to hear about the troubles.
Regarding the pip3 error, it was indeed caused by our image updates. We've cleaned-up the way we set-up the Python environment and now strictly enforce Python version use using pyenv. Which means that if you want to use a different Python version than the system one (which is 2.7.6), you have to explicitly specify it. Adding a "before_install: pyenv global 3.5" step to your travis.yml should switch the system version and make pip3 work without installing any additional packages.
It took up too much room at the top.
Note that we needed to remove it from 2 places, since the main
index.html seems not to use the same top-of-page as the rest uses.
