Extractcode replaces : in file names with _

[JonoYang]

This replacement is causing an issue with how debian system package resources are found and associated in the scancode.io docker pipeline. Some debian .list files have : in their names to separate the architecture from the package name, e.g. libc6:amd64.list. However, extractcode extracts this file as libc6_amd64.list. The code run in the docker pipeline is trying to find the original, unmodified name (libc6:amd64.list), and such the pipeline does not find the declared resources of a package from the .list as it was extracted as libc6_amd64.list

[petergardfjall]

I ran into a similar issue with files like this one: AuthenticationHandler`1.cs being "translated" into AuthenticationHandler_1.cs. It seems to be a convention in certain C# projects to add backticks to filenames.

Would a possibility be to make it possible to opt-out on the filename conversion by introducing a --preserve-filenames flag to extractcode?