[tex-live] \string in xetex

[Joseph Wright]

On 28/04/2015 17:08, David Carlisle wrote:
> I asked a similar question on the xetex list but I wonder if anyone here
> has a better grasp than me on the use of utf16 v utf 32 in xetex.web and
> can offer any insight into the use of \string in xetex.
> 
> The attached file produces a log of
> 
> A=B
> A=C
> A=D
> A=E
> B=C
> B=D
> B=E
> C=D
> C=E
> D=E
> 
> 
> with luatex, but with xetex it produces
> 
> ! Undefined control sequence.
> l.6 \edef\c{\string��
>                        }
> ?
> ! Undefined control sequence.
> l.10 \edef\e{\string��
>                         }
> ?
> A=B
> C=E
>  )
> 
> 
> which is clearly pretty wrong
> 
> (\string is producing two tokens with the UTF-16 representation,
> rather than one token)
> 
> 
> I don't think it's a show stopper for TL2015 as it's always been that way
> as far as I can see, but it would be nice to fix it, but my attempts to do
> so have taught me a lot about the web sources but not produced anything
> usable.
> 
> 
> David

An aside here. With

  \edef\c{\string��}

I'd expect an error as �� is a letter in UnicodeData.txt

    1D400;MATHEMATICAL BOLD CAPITAL A;Lu;0;L;<font> 0041;;;;N;;;;;

With LaTeX in TL'15 you see that with both XeTeX and LuaTeX as both are
now loading unicode-letters.def. However, for plain the formats are
build using unicode-letters.tex (XeTeX, up-to-date I think) and
luatex-unicode-letters.tex (based on an older unicode-letters.tex).
Probably the LuaTeX data file needs updating or alternatively they could
both use unicode-letters.def as this remains compatible with plain
although it's part of the LaTeX2e kernel distribution.
--
Joseph Wright