Re: problem finding mktemp in info file

[Stepan Kasal]

Hello Padraig,

On Wed, Dec 17, 2008 at 12:15:11AM +0000, Pádraig Brady wrote:
> Is there any way to markup example code so that
> 'blah' is not converted to ’blah’ ?

indeed, that is very confusing and the groff man pages (at least the
versions in recent Fedora) are not accurate.

Contrary to popular belief, several ASCII characters do not represent
themselves in the roff source.  These are not only escape characters
(backslash, dot in first column, apostrophe as parameter delimiter).
Such characters include ' (apostrophe) and ` (greve accent, often
used as left quote).  This list may or may not be complete, I don't
know.

Here is the markup table (result of my experiments with groff):

' \(cq          right quote
\' \(aa         accute accent
\(aq            apostrophe (ASCII 0x27)
` \(oq          left quote
\` \(ga         greve accent (ASCII 0x60)

The three without ASCII code listed are characters not present ion
ASCII, so in an Unicode locale, the code point (or what's that
Unicode-speak term) is >= 256.

If you are using a C locale, or an locale with ISO-8859-x encoding,
these Unicode characters may not be available, so a substitute
character is used, and the bug is masked.

In short, shell code snippets shall use \(aq for single quotes and \`
for the "backticks" (the official name of that ASCII character is
"greve accent"---it seems weird to have an accented space character
in ASCII, but that's how things stay).

See also http://bugzilla.redhat.com/471025

Have a nice day,
        Stepan Kasal