Jun, thanks so much for writing this, it was really useful for me recently!
It seems that it did catch almost all of the identifiers except one: content_item__get_latest_revision was not replaced with content_item__get_latest_revision in a few places, so I had to replace them manually. No idea why it missed them. There was no error in the output of dump_parser.pl. If you are interested in debugging it I can to send you the relevant portions of the dump file.