🚀 go-pugleaf

Leo Cacciari <leo.cacciari@univ-pau.fr> wrote:
>[...]
>\newcommand{\subscriptifnotempty}[2]{{#1\ifx#2\empty
>                                            \relax
>                                        \else
>                                            _{#2}
>                                        \fi}}
>\newcommand{\Set}[2][\empty]{\ensuremath{\subscriptifnotempty{\mathsf{#2}}{#1}}\xspace}
>
>\newcommand{\Mname}[2][\empty]{\ensuremath{\subscriptifnotempty{\mathbf{#2}}{#1}}\xspace}
>
>\newcommand{\INMSet}[2][\empty]{\ensuremath{{\Set{#1}{#2}}^{\text{in}}}\xspace}
>
>\begin{document}
>When I type \verb|\INMSet[\Mname{M}]{S}| I espect to obtain somethyng
>like ${\mathsf{S}}_{\mathbf{M}}^{\text{in}}$ but I get
>\INMSet[\Mname{M}]{S} (ie.\ ${\mathbf{M}}S^{\text{in}}$).
>
>Why?

(a) Main: You define \Set with an optional argument, but you use it
with two non-optional parameters in definition of \INMSet.
(b) In math non-argument-braces {...} create a subformula:
$X_a^b$ is different from ${X_a}^b$!
You should remove the braces in the definitions of
\subscriptifnotempty and \INMSet.
(c) In definition of \INMSet the ^{\text{in}} is not in \ensuremath.

Changes (a), (b) and (c):
\newcommand{\subscriptifnotempty}[2]{#1\ifx#2\empty
                                            \relax
                                        \else
                                            _{#2}
                                        \fi}
\newcommand{\Set}[2][\empty]{\ensuremath{\subscriptifnotempty{\mathsf{#2}}{#1}}\xspace}
\newcommand{\Mname}[2][\empty]{\ensuremath{\subscriptifnotempty{\mathbf{#2}}{#1}}\xspace}
\newcommand{\INMSet}[2][\empty]{\ensuremath{\Set[#1]{#2}^{\text{in}}}\xspace}

(d) You use a dangerous \empty-Test in \subscriptifnotempty:
If #2 is really empty and not the command \empty then the not wanted
\else-part is executed.

Empty-tests:
* In plainTeX the empty command is called \empty (\def\empty{})
* In LaTeX the empty command is called \@empty, but \empty works, too.
(i) simple test:
\newcommand{\commandname}[1]{%
  \def\PARAMETER{#1}%
  \ifx\PARAMETER\empty
    parameter empty
  \else
   parameter not empty 
  \fi}

Problem: the place for the command names are wasted.

(ii) test with local command:
\newcommand{\commandname}[1]{%
  \begingroup
    \def\0{#1}%
    \expandafter
  \endgroup
  \ifx\0\empty
    parameter empty
  \else
    parameter not empty
  \fi}

(iii) test with tricky \ifx whithout \empty and temporary command:
\newcommand{\commandname}[1]{%
  \ifx\\#1\\%
    parameter empty
  \else
    parameter not empty
  \fi}

Yours sincerely
  Heiko <oberdiek@ruf.uni-freiburg.de>

PS: A complicated solution with extensive use of optional arguments.
Nesting of arguments causes a problem:
  \commandA[\commandB[optB]]
The FIRST closing ] is the end of the optional argument of \commandA
UNLIKE curly braces!
==> {} needed:
  \commandA[{\commandB[optB]}]

\newcommand{\optsubscript}[2][]{#2%
  \ifx\\#1\\%
  \else_{#1}%
  \fi}
\newcommand{\protectopt}[1]{%
  \ifx\\#1\\%
  \else{#1}%
  \fi}
\renewcommand{\Set}[2][]{\ensuremath{%
  \optsubscript[\protectopt{#1}]{\mathsf{#2}}}\xspace}
\renewcommand{\Mname}[2][]{\ensuremath{%
  \optsubscript[\protectopt{#1}]{\mathbf{#2}}}\xspace}
\renewcommand{\INMSet}[2][]{\ensuremath{\Set[\protectopt{#1}]{#2}^{\text{in}}}\xspace}