Indexing of compressed archives / includes/excludes
I used recoll for indexing a huge directory tree with thousands of files of various types, including binaries, source code, PDF files, text files, various archive types (ZIP, 7zip, rar). I only wanted to index PDFs, inside and outside of archives. Thus, I added the config "onlyNames = *.pdf". It turned out, I needed to add "*.zip *.7z *.rar" as well, to make recoll process archives. However, this lead to indexing of the whole archives, including any other file types (source code, binaries, etc.) inside it.
Having a quick look at the code showed that:
- rclzip.py/rcl7z.py do not honor onlyNames
- rclzip.py/rcl7z.py only honor explicit excludes ("skippedNames" and/or "zipSkippedNames")
- other archive filters neither honor onlyNames nor skippedNames
I wonder, if the following ideas make sense, to improve that situation:
- Do not filter any includes/excludes inside filters but move all that code to recoll itself. This brings includes/excludes support to other filters as well, while not repeating code over and over again.
- Make recoll honor onlyNames/skippedNames for archives as well. (Maybe as a optional feature / checkbox in settings.)
What do you think?
Edited by Michael Niewöhner