update readpassphrase

[vgough]

@benrubson This updates readpassphrase to the latest version I've found in a quick search. I made a couple changes over the original:

• build as C++
• adjust include, add path.h include
• change NULL -> nullptr
• reformat with clang-format

[benrubson]

Clang-tidy complains a lot regarding this new version :)
And I think it should also complain (once some warnings will have been corrected ?) about the 2 (void)write(...).

[vgough]

There are still lots of clang-tidy complaints throughout encfs. Looks like most of the complaints in readpassphrase are readability-implicit-bool-cast and readability-braces-around-statements, neither of which bother me much for imported C code.

I don't see any clang-tidy or compiler complaints about (void)write, so not sure what you're expecting.

[benrubson]

Sorry, these (void)write warnings are not triggered by clang but by gcc... My mistake...
Taken from log given in #407 :

/home/makerpm/rpmbuild/BUILD/encfs-master/encfs/readpassphrase.cpp:128:46: warning: ignoring return value of 'ssize_t write(int, const void*, size_t)', declared with attribute warn_unused_result [-Wunused-result]
   (void)write(output, prompt, strlen(prompt));
                                              ^
/home/makerpm/rpmbuild/BUILD/encfs-master/encfs/readpassphrase.cpp:149:33: warning: ignoring return value of 'ssize_t write(int, const void*, size_t)', declared with attribute warn_unused_result [-Wunused-result]
     (void)write(output, "\n", 1);
                                 ^

I then would like to correct this, so that these warnings do not bother users anymore.
Thus I opened PR #424 (which I just updated with the few log lines above to clarify).

For sure clang-tidy complaints regarding the new readpassphrase version are not so important, but I think it would be interesting to have clang-tidy to warn as less as possible, to easily detect future code mistakes. This is why I opened #427 (as well as #426), to try to make an effort to correct clang-tidy & compilation warnings.

Many thanks 👍

[samrocketman]

@vgough this PR would be more appropriate as two commits because you're copying code. 1 commit for the original and 1 commit for your explicit changes. This way your changes are more obviously documented because I can't tell what is your changes and what is the original copy at first glance.