/# Motivation
It turns out that some projects use capital .C to refer to C++.
This is true for the OpenFoam project which is a widely used open
source software for physics simulations.
Although I do not know of other projects that follow this convention
it is worth noting that both github and `cloc` recognize the .C
extension as a C++ file. This was showed in more detail in this Issue:
XAMPPRocky#1024
/# Implementation
The file `languages.json` contains a list of extensions for each
supported language. So adding ".C" should have been enough for this
feature. But the code was casting extensions to lowercase, which makes
.c and .C indistinguishable. Because of this the `.to_lower` was also
removed from the get_extension utility function.
The "to_lower" was probably added for a purpouse. In order to not break
functionality I added a second check. If the case sensitive test does
not return any result, a second test is done with the to-lower. This
will make the code a tiny bit slower, but only when the case-sensitive
check fails and will only repeat a rather fast check.
/# Testing
Tests are created by the `build.rs` script according to the files
present in `tests/data`. In order to test my solution I copied the
`tests/data/cpp.cpp` into a file called `tests/data/cpp_C.C`.
Adding this file without modifying the code will cause a failing test,
which is what we want. This is mostly a coincidence I think, for some
reason the C language summarize blank lines differently than the Cpp
summarize.
I think the more appropriate test would be that each of these files also
contain the name of the language that they are meant to be as part of
the top comment. This would require additional code changes which I
think is a bit outside of the scope of this change. I do intend to
create a different PR adding this to the tests, if that is ok.
Observation: Since this new file is a copy of an existing one, it does
not occupy extra space in the git file-store, since it stores files by
content hash.
/# Additional considerations
In Windows the file-system is case-insensitive but case-preserving.
I believe that this means that the current version of the code will
work correctly on windows to identify .C as cpp. But I do not have
a windows system to test this on.