Dridi Boukelmoune

That's because we start the leaking varnishlog as a daemon, so it exits the test's process group. Workaround, lower the leak window: ```diff --- bin/varnishtest/tests/u00006.vtc +++ bin/varnishtest/tests/u00006.vtc @@ -10,11 +10,6...

``` **** top shell_out| 0 Debug - "sockopt: Not setting unmodified SO_LINGER for a0=127.0.0.1:62816%00" **** top shell_out| 0 Debug - "sockopt: Not setting unmodified SO_KEEPALIVE for a0=127.0.0.1:62816%00" **** top shell_out|...

This was addressed in d497ec099 and later discussed on the commit mailing list. Summary: - the `-A` coverage becomes a no-op without the `-w` - should we fail a `-A`...

The varnishlog changes can be discussed in https://github.com/varnishcache/varnish-cache/pull/3744.

Bugwash note: @bsdphk was against the idea of keeping track of daemons with platform-specific facilities like linux namespaces.

I force-pushed a split of the transit buffer patch into 4 commits. Doing so, I tried to polish the commit log, C code and test cases. I couldn't make c112...

Registering the idea here: we should try to improve RFC coverage in the code as we make progress in this area. By that I mean comments like: https://github.com/varnishcache/varnish-cache/blob/e896529982027a4325dc3bf19f97c8ebbff4ef53/bin/varnishd/cache/cache_req_fsm.c#L414-L420

I'm wondering whether [VIP24](https://github.com/varnishcache/varnish-cache/wiki/VIP-24%3A-Hitpass-turning-into-hitmiss-after-ttl) also falls under the compliance umbrella.

I added a list of parsing mistakes in the issue description, two items I'm aware of so far.

Happy new year @slimhazard! I think I understand what's happening here and the assertion is here on the wrong assumption that transactions are only ever added where `vsl_candidate()` is called....