Follow-up Comment #2, bug #67356 (group groff): For at least ps and pdf devices, font S (Symbol) is--as I deemed it in bug #63366 for lack of an official term--an extra-special font. (To summarize 63366: special fonts are searched for a glyph if one doesn't exist in the current font. The user can modify the list of special fonts. An extra-special font has the former property but not the latter.)
For the purposes of the following, I am defining semantic fallbacks as those
in tmac/tty-char.tmac that match this regular expression: <.+>
tty-char.tmac contains 77 semantic fallbacks. Of these, font S contains
glyphs for 69 of them. (I determined this by using the script generated by
this shell command:
egrep '<.+>' tmac/tty-char.tmac |
sed 's/.tty-char ../grep '\''^/;
s/\*/\\*/;
s@\] .*@\t'\'' font/devps/S || echo NOT FOUND@'
I make no guarantees about the portability of this command to non-GNU greps
and seds.)
So, for ps and pdf output, there are only 8 characters for which a user might
need a semantic fallback. Those characters, and their fallbacks as defined in
tty-char.tmac, are:
.tty-char \[sc] <section>
.tty-char \[dg] <*>
.tty-char \[dd] <**>
.tty-char \[nc] <not\~superset>
.tty-char \[coproduct] <coproduct>
.tty-char \[+e] <epsilon>
.tty-char \[%0] <permille>
.tty-char \[ps] <paragraph>
For typesetter devices, we can offer far better fallbacks than most of the
semantic ones above. \[+e] can fall back to \[*e], \[nc] can be constructed
by overstriking two other glyphs, etc.
So I see little benefit to making any semantic fallbacks available to ps or
pdf documents.
_______________________________________________________
Reply to this item at:
<https://savannah.gnu.org/bugs/?67356>
_______________________________________________
Message sent via Savannah
https://savannah.gnu.org/
signature.asc
Description: PGP signature
