Reading Lines Verbatim into a Sequence Variable

Question

I would like to create an “environment” (i.e. delimited parser) that reads lines verbatim and stores them in a sequence variable.

I have tried using an xparse +v type argument, then splitting the parsed token list on the endline character, like so:

\seq_set_split:Nxn \l__ks_verb_lines_seq { \char_generate:nn { 13 } { 12 } } { #1 }

where #1 is the verbatim tokens. Unfortunately, this approach strips the indentation from the lines, which makes them unsuitable for my purposes.

I do not require that xparse be used.

Answer 1

The description of \seq_set_split:Nnn mentions that it strips spaces. We can define a simplified version that does not. Here, I've not fully separated out from seq internals, which formally I should do (one could create a 'psuedo-seq' then set a real one from it):

\documentclass{article}
\usepackage{xparse}
\ExplSyntaxOn
\cs_new_protected:Npx \ks_set_split:Nn #1#2
  {
    \tl_set:Nn #1
      {
        \exp_not:N \__ks_set_split:w \exp_not:N \prg_do_nothing:
        #2
        \exp_not:N \__ks_set_split_end:
      }
    \tl_replace_all:Nnn #1 { \char_generate:nn { `\^^M } { 12 } }
      {
        \exp_not:N \__ks_set_split_end:
        \exp_not:N \__ks_set_split:w \exp_not:N \prg_do_nothing:
      }
    \tl_set:Nx #1 { \s__seq #1 }
  }
\cs_new:Npn \__ks_set_split:w #1 \__ks_set_split_end:
  { \exp_args:No \__seq_wrap_item:n {#1} }
\NewDocumentCommand \foo { +v }
  {
    \ks_set_split:Nn \l__ks_verb_lines_seq { #1 }
    \seq_show:N \l__ks_verb_lines_seq
  }
\ExplSyntaxOff

\begin{document}
\foo{some

   stuff
   
   which
is
 indented}

\end{document}

A feature request for a 'non-space stripping' version of \seq_set_split:Nnn is likely the best longer-term fix here: I suspect you are not the only user who wishes to do something like this.