LGTM

Please take a look.
This reenables libFuzzer on i386 on Linux (accidentally disabled here), and fixes a previously failing test.

I'm also thinking i386 libFuzzer is broken on iOS (though I have no way of confirming this) and thus caused the greendragon bot failure: http://green.lab.llvm.org/green/job/clang-stage2-coverage-R/3958/consoleFull#console-section-35

Please take a look.
I could be less conservative here and only disable i386 on Windows but I frankly don't think the risk of breaking other platforms is worth the benefit (nothing).
This is intended to fix https://bugs.chromium.org/p/chromium/issues/detail?id=957971#c5 I've verified that the "fuzzer" target is no longer possible to build on x86 (as opposed to possible but failing as it was before).
It's also worth noting that this breakage is in part caused by libFuzzer building with MSVC (__builtin_clzll isn't supported by MSVC but _BitScanReverse64 is supported, but not on i386) instead of indirectly with clang.

• fix comment

PTAL.

• add missing period

• use exit(1)

• improve message

• improve error message and look for it in tests

• add newline

@kcc I've changed things so that libFuzzer will fail if the argument to -seed_inputs is a non existent file or is empty? What do you think of this change?

• Add more tests to verify we catch empty lists

• Make LF fail if no seed list

• fix bug

• only use @ in argument

• test

• use new format

• remove extra newline

• ideal but test failing
• fix issue

• fmt
• fix nits
• fmt

• improve comments

• change name of test
• Get list argument working again
• combine code
• rename
• Use old method

• remove stale comment
• remove extra zeros
• fix broken test

Please take a look.

• change order

• fix comment

More context is here: https://github.com/google/clusterfuzz/issues/267

PTAL.
This change allows fork mode to be used reasonably on Win.

• undo accidental

• improve comment

Add test

I'm going to add a test for this.