Read old records

[lbudai]

This PR supports read_old_records only for journal source.

[kira-syslogng]

Build SUCCESS, the tests were executed on test branch: master and test suite: functions

[kira-syslogng]

Build SUCCESS, the tests were executed on test branch: master and test suite: functions

[kira-syslogng]

Build SUCCESS, the tests were executed on test branch: master and test suite: functions

[mochrul]

(slight off topic)Hmm. I "received" similar complaint from the ubuntu, that sometimes the build is hung around the python tests (https://ubuntudiff.debian.net/q/package/syslog-ng). It may worth an issue.

[kira-syslogng]

Build SUCCESS, the tests were executed on test branch: master and test suite: functions

[lbudai]

Now, read-old-records implemented also for file-reader.
(this is not a port from pe, I've re-implemented this feature)

[kira-syslogng]

Build SUCCESS, the tests were executed on test branch: master and test suite: functions

[kira-syslogng]

Build SUCCESS, the tests were executed on test branch: master and test suite: functions

[furiel]

+1 Looks good. I have only one suggestion.
The new parameter: new_state(boolean) in (...)_apply_state might be confusing, at least in terms of variable naming.

options->read_old_records = TRUE;
Maybe a 3-state variable instead of this boolean would be better. Something like ENUM(TRUE,FALSE,OFF), when OFF would mean the original behaviour: when new_state is false. Thus we could maintain the original signature of log_proto_buffered_server_apply_state, and there is no need need of _apply_existing_state and _apply_new_state.

EDIT: Not relevant. It does not make sense to have 3state variable, because it only controls if we start at the beginning of a file or the end, and one of these should always be applied for new files.

[lbudai]

@furiel: thanks for the review note.
I wouldn't change the code because we would set read_old_records to OFF only when there is no existing persist state in the persist file: in this case we would need to override the read_old_records value in the options structure... and I think it is not as readable as the current version (it is possible to find a better name than new_state, but I don't have such a name at this time).

The key point here is that we applying an existing state or a new one: when existing one is applied then we continue to read logs from the position stored in the persist file, otherwise
the user can set the start position: head(read_old_records=yes) or tail(read_old_records=no).
When the user does not set this value, the default is yes.

[lbudai]

Hmm...the affile read-old-records will be handled in a separate PR as it needs some change in functionality.
But systemd-journal part of the PR can be merged.

[pzoleex]

@kira-syslogng test this please test branch=pzolee-journal-source-read-old-record;

[kira-syslogng]

Build SUCCESS, the tests were executed on test branch: master and test suite: functions

[kira-syslogng]

Build FAILURE, the tests were executed on test branch: pzolee-journal-source-read-old-record and test suite: functions

[pzoleex]

@kira-syslogng test this please test branch=pzolee-journal-source-read-old-record;

[kira-syslogng]

Build SUCCESS, the tests were executed on test branch: pzolee-journal-source-read-old-record and test suite: functions