Poul-Henning Kamp
Poul-Henning Kamp
Bugwash consensus: Yes, needs fixing, by defining what we want in ABI=vrt and the documenting. WS makes sense, but should probably grow its own vrt_ws.h include, to be sucked in...
bugwash: Some concern about the length if -j arguments. (Possible solution: allow multiple -j arguments, first must always select provider) Now that rewinding argc/argv works most places, we could remove...
Thinking about it, It may be a good idea to store bans (in the filesystem) separately from objects in stevedores. Something like: Adding a ban appends "{timestamp} {banspec}" to /var/run/varnishd.bans...
I'm pretty sure we implemented it the current way to cater to some bogus implementation of chunked, but that would be more than 14 years ago, so one would hope...
"Within varnishd (at the write end for statistics), volatile kind-of makes sense: Because we update values (e.g. increase/decrese), outdated values cause inaccurate statistics." That volatile in varnishd is to make...
This ticket is parked, pending VDD discussion about synth/error and body filters.
I fail to see benefits which warrant breaking working VCL/configs. I can see an argument for adding `*iB` units, and maybe even strongly recommend their use in docs, possibly even...
Uhm, this is quite a grab-bag of different stuff... Do you seriously propose that all this should be reviewed/approved as a single unit ?!
As I see it, the benefits of having a `.via` facility is that we can share: A) Health-probe of the intermediate B) Normal director services (fail-over, load-balancing) of the intermediate....