Move -Wmissing-prototypes from Makefile to .travis.yml

[tkelman]

since MSVC doesn't understand this flag, and the current
mechanism for building Julia with MSVC goes through the makefile

as suggested by @petercolberg at #55 (comment)

[tkelman]

I shouldn't spend too much time getting to the bottom of this right now, but there might be a serious issue with utf8proc 2 and/or flisp or julia's use of it with msvc, even 2013 which has at least been able to get through Julia's bootstrap without all that many patches. If I try to use this branch of utf8proc with any version of Julia since the upgrade to utf8proc 2 was merged, I get

 cd /c/projects/julia/base && /c/projects/julia/usr/bin/julia.exe -C x86-64 --output-ji /c/projects/julia/usr/lib/julia/inference.ji --startup-file=no coreimg.jl
make[1]: *** [/c/projects/julia/usr/lib/julia/inference.ji] Error 253

or similar when trying to start bootstrap. Reverting JuliaLang/julia@10359ec gets through bootstrap again, ref https://ci.appveyor.com/project/tkelman/julia/build/1.0.917 (failing to compile ccalltest.c at the end is expected). Anyone who's curious about this should be able to reproduce by working off of https://github.com/tkelman/julia/commits/tk/avmsvc-clean and enabling appveyor on your personal fork of Julia.

[stevengj]

Regarding the MSVC problem that's concerning, but I'm not sure what the problem could be. On Linux, I ran the utf8proc tests through valgrind with no problems.

[tkelman]

that won't exercise the msvc-only ifdefs, which might now be subtly wrong for the latest code - ref warnings noted in #63 (comment)

[stevengj]

I just looked at how flisp is calling utf8proc, and it seems wonky and antiquated. I'll put together a PR.

[stevengj]

Though I don't see why it should crash. I was worried about the low-level calls to utf8proc_decompose etcetera, but they seem fine; they still match what utf8proc_map does internally. The only change I want to make is that we should call utf8proc_category rather than directly accessing utf8proc_property_t, but I don't see why that should crash it.

[tkelman]

Hm. Dunno, guess I can try getting it going locally and see if I can remember how to work windbg, but I've got higher-priority things to work on right now. Mostly checking if the back-burner "in case we need it" msvc support might actually catch a real bug once in a while.

[stevengj]

Yeah, I totally agree. Porting to multiple compilers is a great way to catch bugs. (Unfortunately, sometimes you catch compiler bugs too.)