Add the option to analyse inputs

[thehubbard]

Hello

It would be useful if the code also processed included sub files via the \input command

For example at the moment a master document such as

\documentclass[11pt]{article}
\begin{document}

\section{Introduction}
\input{Intro.tex}

\section{Conclusion}
\input{Conclusion.tex}

\end{document}

would only analyse section headers. The paragraphs etc which are contained in sub files Intro.tex and Conclusion.tex are ignored. At the moment textidote has to be run individually on each independent file which takes time. A feature enhancement would take sub files into account.

Many thanks

[sylvainhalle]

Hi! That's an interesting suggestion. However, it requires interpreting LaTeX markup, something that TeXtidote does not currently do. This is by design: once you start doing this, there is no end to it --TeXtidote will end up as a reimplementation of LaTeX if you wait long enough. (See my similar response to issue #31.)

I'd be reticent to add such a feature. Note however that you can run TeXtidote with multiple file names on the command line, and it will check each of these files in batch. You can even automate this by using a configuration file (see Readme). This is not exactly what you want, but close enough I think.

[hoehnp]

Hi!
Thank you for your great tool. I would be also interested in this feature. Would you accept a PR implementing the functionality. As you suggest, one could maybe implement it by doing preprocessing with an external script which adds the tex files in a folder recursively. At the end oen could maybe have links in the mail file going to the input files. What are your ideas?

[sylvainhalle]

Hi! Finding all tex files recursively from a root folder can already be done with find; you can also pass that list of files to TeXtidote relatively easily:

textidote $(find . -name "*.tex")

But I agree that this finds all tex files, not only the ones that appear in a non-commented \input or \include instruction inside a main file.

Maybe following the inputs would not be such a big deal after all. One could add a --follow command line argument that would add filenames encountered in a list, and process all files from the list in sequence. I'll flag this as a new feature for the next release (0.7).