Interpret input as UTF-8

[jackvreeken]

Using the pragma open option, we can interpret both stdin and input
files as UTF-8 without having to explicitly open them first.

[jackvreeken]

I ran into UTF-8 interpretation issues. After searching around a bit, I implemented the solution from this post. This should solve issue 86.

I assumed we want to open every file as UTF-8. I do not know perl well enough to also support other encodings by e.g. adding a commandline argument.

[brendangregg]

Thanks, this works. Was hard to reproduce, but I got it in the end: the problem was with unicode characters in the elide code which turns a long function name, eg, "abcdef" into "ab...".

[jackvreeken]

Ah, sorry about that. I probably should have included an example. It was indeed that part of the Perl code that did not work correctly, as it was slicing on code points instead of code units (not knowing any better). Interestingly enough, the JavaScript code for the ellipses worked correctly, as zooming in and out again fixed it. I guess this is because it was interpreting everything as UTF-8 all along.