Btukfyl

I am trying to combine the answers of Remove extra curly braces and def'ed plain-text argument behaves differently from plain text, and I am failing. Here's an MWE:

documentclass{article}

makeatletter

usepackage{seqsplit}

newcommand{dosomething}[1]{expandafterseqsplitexpandafter{@firstofone#1}}



begin{document}

    defkong{looooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooong}

    defkongg{kong}

    defkonggg{{looooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooong}}



    % works

    dosomething{looooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooong}



    % works, too

    expandafterdosomething{expandafterkong}



    % works with extra {...}

    dosomething{{looooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooong}}



    % works with extra {...}, too

    expandafterexpandafterexpandafterdosomething{expandafterexpandafterexpandafterkongg}



    % does not work

    expandafterexpandafterexpandafterdosomething{expandafterexpandafterexpandafterkonggg}

end{document}

For the larger picture, this is what I am getting at:

documentclass{article}

usepackage[T1]{fontenc}

usepackage{amsmath,hyperref}



%1 usepackage{cleveref}



usepackage{showlabels,xpatch,seqsplit}

makeatletter

patchcmd{SL@margintext}{{SL@prlabelname{#1}}}{#1}{}{err}

patchcmd{maketag@@@}{{df@label}}{{expandafterdf@label}}{}{err}



patchcmd{showlabelsetlabel}{#1}{raisebox{0pt}[0pt][0pt]{parbox{1.5marginparwidth}{scriptsizevspace*{-0.5baselineskip}seqsplit{#1}}}}{}{err}



%2 renewcommand{showlabelsetlabel}[1]{meaningdf@label}



%4 newcommand{dosomethingelse}[1]{expandafterseqsplitexpandafter{@firstofone#1}}

%4 patchcmd{showlabelsetlabel}{seqsplit}{dosomethingelse}{}{err}



begin{document}



    %3 patchcmd{label@in@display@noarg}{{{#1}}}{{#1}}{}{err}



    section{A}

    label{sec:Loooooooooooooooooooooooooooooooooooooooooong}



    begin{equation}

    label{eqn:Loooooooooooooooooooooooooooooooooooooooooong}

        X = Y

    end{equation}

end{document}

This code behaves well until I load cleveref (uncomment %1). You will see why after uncommenting %2: df@label, set by amsmath's label@in@display, has an extra layer of {...}.

cleveref, on the other hand, seems to be the one introducing this:

deflabel@in@display@noarg#1{cref@old@label@in@display{{#1}}}%

deflabel@in@display@optarg[#1]#2{%

  cref@old@label@in@display{[#1]{#2}}}%

And it seems to depend on it, as you can see when you try to xpatch it out (uncomment %3). So instead of that, I tried to remove the curly braces as before (comment %2, uncomment %4), to no avail.

I think the problem with your code isn't the definition of dosomething but your test cases. Let's have a look at how the expansion goes in your test cases:

1. dosomething{looong}
   
   expandafterseqsplitexpandafter{@firstofone looong}
   
   seqsplit{looong}
   
   



2. 
   

   expandafterdosomething{expandafterkong}
   
   dosomething{expandafterkong}
   
   expandafterseqsplitexpandafter{@firstofoneexpandafterkong}
   
   seqsplit{expandafterkong}
   
   

   
   
   

   Here you can see that the expandafters are not properly applied, as one goes into the parameter of seqsplit which probably is not what you intended. It still seems to work accidentally because of how seqsplit works internally.

   
   



3. dosomething{{looong}}
   
   expandafterseqsplitexpandafter{@firstofone{looong}}
   
   seqsplit{looong}
   
   



4. 
   

   expandafterexpandafterexpandafterdosomething{expandafterexpandafterexpandafterkongg}
   
   expandafterdosomething{expandafterexpandafterexpandafterkongg}
   
   dosomething{expandafterexpandafterexpandafterkongg}
   
   expandafterseqsplitexpandafter{@firstofoneexpandafterexpandafterexpandafterkongg}
   
   seqsplit{expandafterexpandafterexpandafterkongg}
   
   

   
   
   

   Again, here expandafters find its way into seqsplit's parameter.

   
   


5. 
   

   expandafterexpandafterexpandafterdosomething{expandafterexpandafterexpandafterkonggg}
   
   expandafterdosomething{expandafterexpandafterexpandafterkonggg}
   
   dosomething{expandafterexpandafterexpandafterkonggg}
   
   

   
   
   

   The same problem here, which prevents the @firstofone trick to apply.

   
   

On the other hand, if you put expandafters in the right dose, the result is correct with all of your three kong macros:

expandafterdosomethingexpandafter{kong}



expandafterexpandafterexpandafterdosomethingexpandafterexpandafterexpandafter{kongg}



expandafterdosomethingexpandafter{konggg}