Digging through logs (that you may not have for various reasons) to find out who made a specific change, or who triggered a specific change while doing something else, can be problematic. Adding the user and workstation IDs to the auditor tables reduces that digging significantly for those tables you have added auditors to.
The branch below has three commits:
The first commit adds/changes auditor functions to allow for setting, getting, and clearing auditor information, as well as adding a couple of utility functions for updating auditors after changes to their origin columns. When all is said and done, calling auditor.update_auditors will ensure that all auditors are up to date and functional with lifecycle views, as well as having the new user and workstation id columns (and any other missing columns, for that matter).
The second commit is the upgrade script, unwrapped, for the first one. It includes calling auditor.update_auditors at the end.
The final commit calls the set and clear functions from various points in the Evergreen source. This involves adding set_audit_info methods to cstore, rstore, pcrud, and storage. Pcrud automatically sets audit info when fetching user information, the others require an explicit call from the code connecting to them. When a session disconnects from any of them they should clear the previously set audit information.
I added calls to the proper set_audit_info methods whenever I found points that needed them.
http://git.evergreen-ils.org/?p=working/Evergreen.git;a=shortlog;h=refs/heads/user/tsbere/auditor_boost
Cherry-picked and signed off in
http://
I've been using this in development for several weeks. It works and doesn't seem to cause any side effects.