WG, Sorry for joining in the discussion late. I've only just found some time to reply.
My thoughts below... The new format looks great. <PRI>VERSION TIMESTAMP HOSTNAME APP-NAME PROCID MSGID [SD-ID]s MSG Replace all received null characters with either <00> or /0. My preference is <00>. Keep MSGID in the header as a required field SD-IDs should come before the MSG. Otherwise encoding issues and MSG delimiter will become a problem. Store all messages written to disk in UTF-8 format. This allows any received encoding to be stored safely without loss or corruption. My preference is to enforce UTF-8 for data encoding on the wire. This allows US-ASCII to be used for the first 127 characters and Unicode mappings into UTF-8 for all other international characters. Trying to switch encodings for each message based on the SD-ID language or local setting will be a parsing nightmare. As far as I know, all modern systems are now capable of sending in US-ASCII or mapping their own language into UTF-8. Can anyone think of a good reason not to enforce UTF-8? I believe the above format would be easy to implement in both a sender and receiver. Mandating that the disk storage format is UTF-8 would also help reporting and parsing of all languages and character sets. Mapping over UDP should be limited to a single message per packet. When mapping over plain TCP I believe we should limit the total message size to 65507 bytes (to keep it compatible with UDP) and delimit each message stream with an LF, or CRLF. Either delimiter would work for me. Rainer, keep up your good work and persistence on the drafts. I believe the new format will solve a lot of problems. Cheers Andrew _______________________________________________ Syslog mailing list Syslog@lists.ietf.org https://www1.ietf.org/mailman/listinfo/syslog
