Review use of \xspace in macro definitions

[jensmaurer]

See http://tex.stackexchange.com/questions/86565/drawbacks-of-xspace with the first answer from the package author of \xspace. Quote:

So, if you find it useful, fine, it's there. But personally I wouldn't recommend it.

\xspace in \grammarterm caused spacing issues while addressing #1268.

[godbyk]

Since \grammarterm takes a mandatory argument, there's no need for \xspace there. \xspace is used primarily in macros that take no argument and expand to text. For example, consider the following:

\newcommand{\myname}{Kevin}% no \xspace
My name is \myname. \myname is my name.
% \myname swallows and following white space
% such that the following word is adjoined to the
% expanded 'Kevin' text producing,
% 'Kevinis my name.'
%
% You work around this by using doing one of the
% following:
\myname\ is my name.
\myname{} is my name.

% ... or use \xspace:
\newcommand{\mynamex}{Kevin\xspace}
My name is \mynamex. \mynamex is my name.

It looks like folks have been working around this issue by wrapping the s in braces to prevent \xspace from inserting space. Compare:

\grammarterm{blah}s
\grammarterm{blah}{s}

In this case, it'd be better to just remove the \xspace macro from \grammarterm.

[godbyk]

The following uses of \xspace look questionable to me:

\newcommand{\techterm}[1]{\textit{#1}\xspace}
\newcommand{\defnx}[2]{\indexdefn{#2}\textit{#1}\xspace}
\newcommand{\term}[1]{\textit{#1}\xspace}
\newcommand{\grammarterm}[1]{\textit{#1}\xspace}
\newcommand{\Fundescx}[1]{\textit{#1}\xspace}
\newcommand{\NTS}[1]{\textsc{#1}\xspace}
\newcommand{\Range}[4]{\tcode{#1#3,\penalty2000{} #4#2}\xspace}
\newcommand{\diffdef}[1]{\hfill\break\textbf{#1:}\xspace}
\newcommand{\stage}[1]{\item{\textbf{Stage #1:}}\xspace}
\newcommand{\doccite}[1]{\textit{#1}\xspace}
\renewcommand{\emph}[1]{\textit{#1}\xspace}
\newcommand{\numconst}[1]{\textsl{#1}\xspace}
\newcommand{\logop}[1]{{\footnotesize #1}\xspace}
\newcommand{\terminal}[1]{{\BnfTermshape ##1}\xspace}

These uses are probably fine:

\newcommand{\Cpp}{\texorpdfstring{C\kern-0.05em\protect\raisebox{.35ex}{\textsmaller[2]{+\kern-0.05em+}}}{C++}\xspace}
\newcommand{\CppIII}{\Cpp 2003\xspace}
\newcommand{\CppXI}{\Cpp 2011\xspace}
\newcommand{\CppXIV}{\Cpp 2014\xspace}
\newcommand{\opt}{{\ensuremath{_\mathit{opt}}}\xspace}
\newcommand{\xref}{\textsc{See also:}\xspace}

Slightly iffy depending on the expected behavior of the environments:

\newenvironment{note}[1][Note]{\noteintro{#1}}{\noteoutro{note}\xspace}
\newenvironment{example}[1][Example]{\noteintro{#1}}{\noteoutro{example}\xspace}

[zygoloid]

I'd like to move away from xspace entirely. Over-reliance on the \xspace in \opt has caused formatting issues in grammar snippets in the past. I think we just need to get into the habit of writing it \opt{} -- or even \opt{optional thing}.

[godbyk]

@zygoloid That's probably the better option. I would document the meaning of the empty braces, though, so that future authors don't remove them out of ignorance.

In my lists above, the 'questionable' uses of \xspace should be safe to remove as the \xspace won't have much appreciable impact (aside from kerning changes). The 'probably fine' uses require each use to be evaluated individually (some will need to be rewritten with {} or \space). I don't recall if \xspace can see past the end of an environment well enough to function consistently or not, so I would have to check and see what impact removing \xspace has there.

[jensmaurer]

I like \opt{thing}, it's much more content-focused than \opt{}.