Walid Boudebouda

Regarding http_req_size parameter, currently it is not taken into account when using h2, instead the checked parameter is h2_max_frame_size + 9 (h2 header size), which defaults to 16k (the minimum)

- Added a callback in `struct cli_cmd_desc` that allows sensitive commands to provide their own logging function. If a sensitive command does not provide an implementation, then only the command...

Moved the command logging callback to `struct cli_cmd_desc` as suggested by dridi so that it can be overridden per command implementation. If the sensitive command is a cache process command,...

This PR should probably be discussed again in bugwash regarding error page customization brought up by Nils.

Bugwash: merge this with the existing `FEATURE_BIT(VALIDATE_HEADERS)` logic.

From last bugwash: > \ Along these lines, I would think it makes more sense to validate headers "on the way in and upon each change", rather than (again) at...

@delthas can you give https://github.com/varnishcache/varnish-cache/pull/4059 a try ? Note that this has only been tested locally, and I did not measure the performance impact. When working on this, I found...

We already error out when we get an unexpected 304 response from the backend : https://github.com/varnishcache/varnish-cache/blob/master/bin/varnishd/cache/cache_fetch.c#L375-L378 ``` if (bo->stale_oc != NULL && ObjCheckFlag(bo->wrk, bo->stale_oc, OF_IMSCAND)) { .... } else if...

> ``` > sub vcl_refresh_conditions { > if (!bereq.http.if-modified-since && > !bereq.http.if-none-match) { > call vcl_refresh_error; > } > } > ``` I think this should rather be `return (error);`...

Bugwash: > [15:09:38] \ I dont see any problems [15:09:49] \ but I guess we should give @nigoroll a chance