Fail to scan directory

[adambhere]

Description

This issue is related to #1748.

A directory named "rcs" was ignored by Scancode 3.0.2 (and 3.1.1), but not its contents. Scancode threw a KeyError exception and the results did not cover the complete target directory tree. The reason is that Scancode takes "rcs" to be identical to "RCS", which is on its built-in ignore list.

The latest versions of Scancode remove "RCS" from the ignore list, but this seems a partial solution, it would be better to go further and make the checks of directory/file names against items on the ignore list case sensitive (except on Windows). Specifically, get_matches() in fileste.py changes the case of the item to be ignored to lower and I suggest it shouldn't.

How To Reproduce

Run a scan on a directory containing a subdirectory named "rcs" (with further subdirectories and files in it)

Command-line options used:
--copyright
--license
--ignore
--info
--strip-root
--timeout 300
--processes 15
{DIRECTORY_TO_SCAN}

System configuration

• OS: Ubuntu 18.04
• scancode-toolkit: 3.0.2 (also tested 3.1.1)

[pombredanne]

Thank you ++...

For reference if someone does not know about it, the RCS we are talking about is this (ancient) system: https://en.wikipedia.org/wiki/Revision_Control_System

[mjherzog]

I remember RCS!