Split email archives downloaded from Google Takeout (Download Your Data) service into individual emails. Based on experimentation it looks like Google uses the mboxrd dialect of mbox format with CRLF lines as discussed at the Wikipedia mbox article

The project is licensed under the BSD 3-Clause License - see the LICENSE.txt file included with the package.

The package provides both libraries and a buildable executable. See the code documentation on using the libraries.

The executable takes the following parameters:

-dir  <name>     : A directory to put the resulting messages to.
                   The directory must exist before running the program.

-mbox <name>     : An mbox file to process and split into messages.

-email <address> : An email which correspondence to be captured. Only
                   the actual address should be provided.

The program does not preserve unfinished last line of the last message in the archive. In the resulting files all message lines end with CRLF after the processing.

During the processing it creates temporary message files and then moves them into the UTC-timestamped .eml file. If the destination filename is already taken by another message, then the later message does not override it. It is left in the temporary file and the error is printed to stderr.

Also a message stays in a temporary file if the program fails to construct a name for the message file. Some forwarded messages, for example, lack the Date: header.