Re: Snippet 888 string, was Re: Lyric centre-on-word / ignore-punctuation bug

[David Wright]

On Sat 24 Feb 2018 at 16:59:37 (+0100), David Kastrup wrote:
> David Wright <address@hidden> writes:
> 
> > On Sat 24 Feb 2018 at 11:06:20 (+0100), David Kastrup wrote:
> >> Thomas Morley <address@hidden> writes:
> >> 
> >> > 2018-02-24 3:45 GMT+01:00 Kieren MacMillan <address@hidden>:
> >> >> Hi all,
> >> >>
> >> >> When a lyric syllable begins with a typographer’s single quote
> >> >> (e.g., ’cause, ’ll, etc.), the "ignore-punctuation" hack doesn't
> >> >> work. Any hints on how to fix this would be appreciated.
> >> >>
> >> >> Thanks,
> >> >> Kieren.
> >> >>
> >> >> %%%  SNIPPET BEGINS
> >> >> \version "2.19"
> >> >>
> >> >
> >> > Only characters from `space-set' are respected by `center-on-word'.
> >> > Extend the given string with "’"
> >> >
> >> >> #(define space-set
> >> >>   (list->char-set
> >> >>     (string->list "—.?-;,:“†‘’–— */()[]{}|<>!`~&…")))
> >> >
> >> > #(define space-set
> >> >   (list->char-set
> >> >     (string->list "’—.?-;,:“†‘’–— */()[]{}|<>!`~&…")))
> >> >
> >> > works for me.
> >> 
> >> Doesn't really look like dealing sensibly with utf-8.
> >
> > AFAICT the string in the original snippet is
> >
> > —.?-;,:“”‘’–— */()[]{}|<>!`~&…
> 
> I was talking about the code,

I wouldn't know about the code: I'm not familiar with scheme.
The three lines quoted don't exactly look complicated, but I
wouldn't like to judge.

> not the string (which is garbled, sure).

Yes, I've never seen an ungarbled version posted, so I thought
I'd just look at it carefully and see if it was obvious what
it was meant to be. And, as it happened, it was.

> > What I haven't figured out is why there are two long dashes;
> 
> en-dash and em-dash are different code points.

My apologies for being so sloppy as to call an em-dash a long
dash, as opposed to the shorter en-dash. There are still two
of them, as shown by the marked line:

 8211    2013    1       1      –
 8212    2014    2       1      —     ←
 8216    2018    1       1      ‘
 8217    2019    1       1      ’
 8220    201c    1       1      “
 8221    201d    1       1      ”
 8230    2026    1       1      …
Scanned 31 Unicode characters with 0 problems
$ 

I didn't bother to annotate the columns as I thought the 2 was
a bit obvious (apart from the fact that the recoded text has
two long dashes in it).

Cheers,
David.