Change the section name parsing to only remove trailing zeroes

[HoundThe]

Fixes #958

The removal of all NULL chars in the pelib was intentional as in the comment:

	// Historically, PELIB copies the name of the section WITHOUT zero chars,
	// even if the zero chars are in the middle. Aka ".text\0\0X" results in ".textX"

Not sure what was the reason, but based on the issue, I've changed the parsing to strip only the trailing zeroes.

[PeterMatula]

These tricky iterations again :-D
Corner cases:

1. I think it can handle even the case when an entire name is \0. But do we have this in regression tests? I think not. Can we add it? The easiest would probably be to hack some sample, possible the one already used if there are some more sections to play with, with some hex editor.
2. Can we handle if the name is not terminated at all - all the chars are printable? I think we decrease end once anyway, and then assign end + 1 as end. So it is probably ok. But again, can we have a test for this - if there already isn't one?

[HoundThe]

I think I tried all of these corner cases and I still have the samples I've created for this. I'll have a look into the regression suite if there is something like that, if not then I'll create a single sample that has these cases and update the tests.

[HoundThe]

I've added the test cases