Nils Goroll

pow wowed again: * `SessClose` docs in `include/tbl/vsl_tags.h` need TLC regarding the _Why_ (`sess_close_2str()` that is) * We might want to add something similar to `strerror()`

clarified with @bsdphk : The fetch should be agnostic to changes from VCL, I shall have another look and find out if there is any problem in that regard.

The vsmw code makes the assertion common in varnish that there is enough memory available, but also that we have enough disk space and file descriptors available. I think these...

(edit: update to better parameter names, clarification) from discussion with @Dridi and @bsdphk : https://pubs.opengroup.org/onlinepubs/9699919799/functions/V2_chap02.html#tag_15_14 > All functions that open one or more file descriptors shall, unless specified otherwise, atomically...

Except for the question if we should impose a delay before we get close to the limit, this ticket is ready for implementation. In the first iteration, the acceptor would...

👍🏽 in general, but I would like to ask about some divergence from our previous collaboration: > * `between_bytes_timeout` renamed to `beresp_idle_timeout` I think _between bytes_ is a very good,...

It's also showing up in [vtest](http://varnish-cache.org/vtest/) together with f00001 failures

I am going to revert it for now because the cause is not immediately obvious to me, I need to focus on other tasks and Dridis availability this week is...

VCL is what makes Varnish so powerful, so IMHO we should try to implement the "profiles" in VCL.

I do not care much about prune/merge. Technically it is actually both: For adjacent networks, It is a merge, for subnets, prune. I need to report some numbers.