"Anthony J. Bentley" writes: > Ingo Schwarze writes: > > Hi Carsten, > > > > Carsten Kunze wrote on Fri, Mar 20, 2015 at 01:07:55PM +0100: > > > > > in groff_char.7 it is specified that U+0303 is output for \(a~ but > > > actually U+007E is output. Why does the output (of nroff -Tutf8) > > > differ from the specification? > > > > It seems to me that you misread the documentation. It reads: > > > > Accents > > ------- > > > > The composite request is used to map most of the accents to non-spacing > > glyph names; the values given in parentheses are the original (spacing) > > ones. > > > > Output Input PostScript Unicode Notes > > ------------------------------------------------------------------------- > - > > ~ \[a~] tilde u0303 (u007E) tilde accent > > > > That means you get U+007E when using the character directly - > > which certainly makes sense. > > > > The combining (non-spacing) variant shown before the parenthesis > > only applies in the context of the .composite request. > > When looking at this, I learned something new: U+007E is *not* the > spacing equivalent to U+0303; U+02DC is. This effect is visible in > groff's PDF output. In fact, an ASCII tilde gets replaced with a U+0303
s/0303/02DC/ on this last line. > in PDF output, and must be escaped as \(ti to output a straight ASCII > tilde. However, UTF-8 nroff(1) output seems to output U+007E for both > \(a~ and ~ inputs. > > -- > Anthony J. Bentley
