## Motivation
Currently, `tracing-journald` prefixes event fields with the number
of parent spans, if present, in order to namespace field values. This
turned out to be unnecessary as journald preserves values for duplicated
field names.
## Solution
This branch removes event field prefixes and emits parent span names
and their field/value pairs as additional key/value pairs.
See #1698: Properly write large payloads to journal.
I'd appreciate a very careful review; this cmsg stuff is nasty, and
while it's well documented in `cmsg(3)` I had to fiddle a bit because
the corresponding functions in libc aren't const and thus don't permit a
direct allocation of the buffer as most `cmsg` C code around does.
Closes#1698
## Motivation
Linux limits the maximum amount of data permitted for a single Unix
datagram; sending large payloads directly will fail.
## Solution
Follow systemd.io/JOURNAL_NATIVE_PROTOCOL/ and check for `EMSGSIZE` from
`send()`; in this case write the payload to a memfd, seal it, and pass
it on to journald via a corresponding SCM_RIGHTS control message.
Per discussion in #1698 this adds no dependency on `nix`, and instead
implements fd forwarding directly with some bits of unsafe `libc` code.
See #1710: Do not write strings in Debug representation.
## Motivation
As discussed in #1710 writing strings literally makes tracing-journald
behave like other journal clients, and allows 3rd party journal readers
to extract the original value from the journal without having to
"un"-parse the Debug representation of Rust strings.
Fixes#1710.
Per discussion with @hawkw in #1698 I'm adding a few simple integration
tests for the journald subscriber, to have some safety net when
implementing the actual issue in #1698.
These tests send messages of various complexity to the journal, and then
use `journalctl`'s JSON output to get them back out, to check whether
the message arrives in the systemd journal as it was intended to.
## Motivation
Increase test coverage for the journald subscriber and codify a known
good state before approaching a fix for #1698.
