https://bz.apache.org/bugzilla/show_bug.cgi?id=70367

--- Comment #1 from abernal <[email protected]> ---
Thanks for this, Pierre. A global switch for mod_http3 is a fair thing to want,
especially with a static build. A few things I noticed reading the patch
against trunk:

- h3_create_server_config() allocates the config with apr_pcalloc, and the
patch doesn't set a default for the new h3_enabled, so it starts at 0.
h3_is_enabled() then returns 0 for any server that never uses the directive, so
HTTP/3 would be off unless H3Engine On is written. The module already has a
tri-state helper for this (h3_tri_flag / H3_FLAG_UNSET), which would let it
default to on and merge correctly.

- APLOGNO(10500) is already used by mod_md (mod_md_drive.c). New messages can
go in with an empty APLOGNO() and get a number from the log-tag tooling.

- The attachment includes h3_config.h.orig/.rej files, so it may be worth
regenerating it against current trunk.

On the directive itself: HTTP/3 is already enabled per server only when "h3" is
in Protocols and a certificate is configured, so it would help to know what
H3Engine Off gives you that leaving h3 out of Protocols doesn't. The mod_http3
maintainers will have the final say on that.

-- 
You are receiving this mail because:
You are the assignee for the bug.
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to