Thoughts on makeinfo

[Randall Sawyer]

First of all, I acknowledge that re-writing makeinfo from C source to Perl was no small feat. Upon first reading that in the history of the app, I was impressed.

I am pleased that makeinfo is now a Perl product. Perl certainly lends itself well to writing a parser.

Makeinfo works. It works well.

I appreciate being able to learn how the tools I use actually work from the inside when I need to. Over the past few days, I have copied the contents of makeinfo and edited it to verbosely show me its progress and the names and values of variables, printed to stderr. I also re-arranged much of the source code to help me to follow along. I have found some redundant and unused portions; but, these are minimal.

One request I have, is that in future release of texinfo, that the Perl package YAML be optionally included via autotools. In my re-edit of makeinfo, by using YAML instead of Data::Dumper, the contents of the parser tree is much more comprehensible. As YAML reads and writes data using the same syntax - converting data in a consistent manner into and from scalars, arrays and hashes, it allows for  writing various scripts which can share the same files and/or pipes.

I did find two bugs which only show up when using makinfo in an atypical manner. I will go and post each of them on the texinfo-bug list unless I see that it is already in the archive.