It adds statements in the manual entry that the OmitHeaders, OversignHeaders, SenderHeaders, SignHeaders parameters are comma-separated. This was not explicitly stated before.
It adds a statement in the manual entry explicitly listing which headers are included in the default value for the SignHeaders parameter. Previously, the manual entry incorrectly referred the user to RFC6376 and a list of headers that "SHOULD" be signed, although RFC6376 contains no such list. It only has a list of commonly-signed headers, and the default in OpenDKIM contains one extra header that is not in that list. Now, the manual entry informs the user exactly what the default is without requiring them to refer to RFC6376.
The config sample file has had fixes to the default values of some of these parameters, and it now has examples showing the use of the *,+header1,-header2 notation to modify the existing default.
Note: I asked if the above usage was actually valid, and haven't received an answer yet, but I think it must be valid usage, and so I have assumed for the moment that it is, and have created this pull request now rather than waiting for the answer. If it turns out not to be the case, I can modify this pull request by removing the examples that show that usage.
This pull request addresses issue #131.
It adds statements in the manual entry that the OmitHeaders, OversignHeaders, SenderHeaders, SignHeaders parameters are comma-separated. This was not explicitly stated before.
It adds a statement in the manual entry explicitly listing which headers are included in the default value for the SignHeaders parameter. Previously, the manual entry incorrectly referred the user to RFC6376 and a list of headers that "SHOULD" be signed, although RFC6376 contains no such list. It only has a list of commonly-signed headers, and the default in OpenDKIM contains one extra header that is not in that list. Now, the manual entry informs the user exactly what the default is without requiring them to refer to RFC6376.
The config sample file has had fixes to the default values of some of these parameters, and it now has examples showing the use of the *,+header1,-header2 notation to modify the existing default.
Note: I asked if the above usage was actually valid, and haven't received an answer yet, but I think it must be valid usage, and so I have assumed for the moment that it is, and have created this pull request now rather than waiting for the answer. If it turns out not to be the case, I can modify this pull request by removing the examples that show that usage.