Hi Branden,
G. Branden Robinson wrote on Tue, Aug 02, 2022 at 10:42:45AM -0500:
> At 2022-08-02T15:44:21+0200, Ingo Schwarze wrote:
>> In groff, this works for me:
>>
>> $ printf "a\\\\N'9'b" | groff -T ascii | hexdump -C | head -n 1
>> 00000000 61 09 62 0a 0a 0a 0a 0a 0a 0a 0a 0a 0a 0a 0a 0a |a.b.............|
>>
>> Mandoc behaves differently and treats \N'9' exactly like a literal HT:
>>
>> $ printf "a\\\\N'9'b" | mandoc | hexdump -C | grep 61
>> 00000050 61 20 20 20 20 62 0a 0a 20 20 20 20 20 20 20 20 |a b.. |
>>
>> In general, mandoc lets fewer control characters sneak through into
>> output than groff because i worry that control characters in output
>> might occasionally cause reliability or security issues.
> I don't predict high reliability from this technique
Heh. :-)
I'm used to this statement in the mandoc_char(7) manual page:
NUMBERED CHARACTERS
For backward compatibility with existing manuals, mandoc(1)
also supports the
\N'<number>' and \[char<number>]
escape sequences, inserting the character <number> from the
current character set into the output. Of course, this is
inherently non-portable and is already marked as deprecated
in the Heirloom roff manual; on top of that, the second form is
a GNU extension. For example, do not use \N'34' or \[char34],
use \(dq, or even the plain ‘"’ character where possible.
So i assumed it is well-known to not be the pinnacle of portability.
But since
https://www.gnu.org/software/groff/manual/html_node/Using-Symbols.html
documents \N without explicitely calling out its inherent portability
problems, maybe i should have mentioned the trap.
Then again, the wording
"Typeset the glyph with code n in the current font ..."
does provide an *implicit* hint that this can hardly be expected to
be device-independent.
> when attempting it on a platform that uses IBM code page 1047
> as its input encoding. ;-)
I would have expected the *output* font numbering to cause even
more serious trouble than the *input* encoding. Besides, not being
a masochist to that degree you appear to assume, i prefer this
counter-example:
printf "a\\\\N'9'b" | groff -T pdf > tmp.pdf
After that, the file tmp.pdf displays three characters:
The letter "a", the Euro-sign (oops!?), and the letter "b".
I expect we are soon going to dissuade Alejandro from his plan. :)
Yours,
Ingo
