It is sometime useful to write a script to perform a batch editing on several files in a source tree. Such a script is usually very specific and used only once.
Such scripts can be used to rework some identifiers in code, modify file header comments, edit the path of a resource or the like. I want to stress that I am concerned by the case where the script addresses an exceptional problem, so that it does not fit in the usual workflow of the project.
I think it is important to keep record of these scripts and reference the usage of the script in the RCS logbook. Where is the appropriate place to keep this record?
Should they be added to the commit description in the RCS logbook? Should they be added to the ticket description of the item being processed when they are used? Should they be saved as regular files in the project and added to the RCS, despite the fact their usage is deemed to be unique? Is there another way?