Ref T3307. We sometimes need to truncate strings to a given length for display, storage, or for other reasons.
Currently, phutil_utf8_shorten() can do this for display, but because it handles combining characters it doesn't actually impose any storage limits on the result: if the input is o and 50 megabytes of combining umlauts, the output will be the same as the input, since this is only "1 display character" long.
Instead, allow truncation to specify maximum byte, codepoint (utf8 character) and glyph (display character) limits.