On 06/11/14 13:57, Werner LEMBERG wrote: > groff uses different quotes depending on -Tascii or -Tutf8:
Naturally. UTF-8 is a multibyte character encoding; ASCII is strictly single byte (using only seven bits). > devascii: lq " Which is ASCII code 34 (0x22) > devlatin1: lq " Again, looks the same as ASCII code 34. > devutf8: lq u201C (via `glyphuni.cpp') And this is way outside the range of any 7-bit encoding, (it needs fourteen bits plus UTF-8 encoding overhead spreading it across three bytes), so has no ASCII representation. > Using \*[Lq] and \*[Rq] in Anthony's patch is definitely correct. I don't disagree. However, Anthony clearly *is* confused, for he claims to refer to ASCII output on UTF-8 devices. It may seem pedantic, but a confusingly expressed objective must perpetuate confusion. While ASCII may be a proper subset of UTF-8, it is a limited subset: it offers only two choices for representation of quotes, viz. code 34 (double quote) and code 39 (single quote or apostrophe).
