Mini-indices in the margins

[raghnysh]

@nrnrnr Many thanks for this system for literate programming, and even more for your wonderful book Build, Prove, Compare. Can you please describe how to get mini-indices of identifiers as in the book?

[nrnrnr]

The mini-indices use a fair amount of custom engineering. To start at the back end, noweb generates an "MDB file" that contains all the marginal indices in one place. That file includes a LaTeX environment for each mini-index, something like this:

\begin{marginalindex}{25}
  \inchunk{mod}{NW3lf9Qy-1I7sVC-3}% 27f
\end{marginalindex}
\begin{marginalindex}{45}
  \autogen{\textrm{type} Def}%
  \inchunk{\textrm{type} Echo}{NW3PGINC-qYINM-3}% S293j
  \autogen{\textrm{type} Exp}%
  \autogen{\textrm{type} Func}%
  \inchunk{\textrm{type} Name}{NW3lf9Qy-qYINM-1}% 43d
  \inchunk{\textrm{type} Namelist}{NW3lf9Qy-qYINM-1}% 43d
  \autogen{\textrm{type} Userfun}%
  \inchunk{\textrm{type} Value}{NW3lf9Qy-32mcxO-1}% 43f
  \inchunk{\textrm{type} Valuelist}{NW3lf9Qy-32mcxO-1}% 43f
\end{marginalindex}

The environment is defined in mini-idx.sty.

The MDB file is generated by a Lua script that crawls three kinds of information:

• The main .aux file for the book
• A list of the primitives defined in each chapter, obtained by crawling the source code
• Multiple lists of definition-use pairs for each chapter

Definition-use pairs can be obtained from the MLton compiler (for ML code), the rtags package (for C code), or a modified interpreter (for bridge-language code).

[raghnysh]

On 27 June 2023 22:36:48 GMT+05:30, Norman Ramsey @.***> wrote:

The mini-indices use a fair amount of custom engineering. To start at the back end, noweb generates an "MDB file" that contains all the marginal indices in one place. That file includes a LaTeX environment for each mini-index, something like this:

\begin{marginalindex}{25}
 \inchunk{mod}{NW3lf9Qy-1I7sVC-3}% 27f
\end{marginalindex}
\begin{marginalindex}{45}
 \autogen{\textrm{type} Def}%
 \inchunk{\textrm{type} Echo}{NW3PGINC-qYINM-3}% S293j
 \autogen{\textrm{type} Exp}%
 \autogen{\textrm{type} Func}%
 \inchunk{\textrm{type} Name}{NW3lf9Qy-qYINM-1}% 43d
 \inchunk{\textrm{type} Namelist}{NW3lf9Qy-qYINM-1}% 43d
 \autogen{\textrm{type} Userfun}%
 \inchunk{\textrm{type} Value}{NW3lf9Qy-32mcxO-1}% 43f
 \inchunk{\textrm{type} Valuelist}{NW3lf9Qy-32mcxO-1}% 43f
\end{marginalindex}

The environment is defined in mini-idx.sty.

The MDB file is generated by a Lua script that crawls three kinds of information:

• The main .aux file for the book
• A list of the primitives defined in each chapter, obtained by crawling the source code
• Multiple lists of definition-use pairs for each chapter

Definition-use pairs can be obtained from the MLton compiler (for ML code), the rtags package (for C code), or a modified interpreter (for bridge-language code).

Thanks for the explanation. I'll see if I can implement all these steps, or alternatively try for a more LaTeX -oriented solution.

[nrnrnr]

If you want scripts, let me know, although they are a royal mess.

[raghnysh]

Yes, the scripts would be useful. I'll send you a request by email.

On 27 June 2023 11:03:12 pm IST, Norman Ramsey @.***> wrote:

If you want scripts, let me know, although they are a royal mess.

[dbosk]

I'm also interested in this. Please CC me at dbosk at kth dot se. Or just upload the scripts here in this thread, that would also work.

[nrnrnr]

Attached is a stack of scripts plus some sample .prim and .du files.

A partial list of how book.mdb is built is as follows:

   ./import-du -absolute -methods $BUILD/src/usmalltalk/newtests.bdu $BUILD/src/usmalltalk/transcript.bdu $BUILD/src/usmalltalk/exercises.bdu $BUILD/src/usmalltalk/tests.bdu $BUILD/src/usmalltalk/simscript.bdu $BUILD/src/usmalltalk/bignum.bdu > usmalltalk-book.du
   ./mlprim ./typesys.nw ./typesysa.nw > typesys.prim
   ./mlprim ./upr.nw ./upra.nw > upr.prim
   ./cprim ./gcs.nw ./gcsa.nw > gcs.prim
   ./mlprim ./adt.nw ./adta.nw > adt.prim
   ./mlprim ./tuscheme.nw ./tuschemea.nw > tuscheme.prim
   ./mlprim ./mcl.nw ./mcla.nw > mcl.prim
   ./mlprim ./ml.nw ./mla.nw > ml.prim
   ./mlprim ./haskell.nw ./haskella.nw > haskell.prim
   ./mlprim ./mlscheme.nw ./mlschemea.nw > mlscheme.prim
   ./cprim ./schemes.nw ./schemesa.nw > schemes.prim
   ./cprim ./scheme.nw ./schemea.nw > scheme.prim
   ./cprim ./impcore.nw ./impcorea.nw > impcore.prim
   ./mlprim ./usm.nw ./usma.nw > usm.prim
       # was dependend o $TEXFILES $AUXFILES 
./mini5 book.aux -prim typesys.prim upr.prim gcs.prim adt.prim tuscheme.prim mcl.prim ml.prim tuscheme.prim haskell.prim mlscheme.prim schemes.prim scheme.prim impcore.prim adt.prim tuscheme.prim gcs.prim usm.prim -du timpcore-mlton.du uprolog-mlton.du uscheme-copy-rtags.du uml-mlton.du tuscheme-mlton.du molecule-mlton.du nml-mlton.du uhaskell-mlton.du uscheme-ml-mlton.du uschemeplus-rtags.du uscheme-rtags.du impcore-rtags.du umlx-mlton.du uscheme-ms-rtags.du usmalltalk-mlton.du timpcore-book.du uscheme-copy-book.du uml-book.du nml-book.du uscheme-ml-book.du uscheme-book.du impcore-book.du uscheme-ms-book.du usmalltalk-book.du -used indexed.dus > book.mdb

[raghnysh]

Many thanks. I'll try these scripts.