* caddyfile: add `observe_catchall_hosts` option
Signed-off-by: Mohammed Al Sahaf <msaa1990@gmail.com>
* correct JSON field name and doc comment
Signed-off-by: Mohammed Al Sahaf <msaa1990@gmail.com>
---------
Signed-off-by: Mohammed Al Sahaf <msaa1990@gmail.com>
* WIP tracing span attributes
* better test
* only write attributes after other middleware (and request)
* Fix test to use header response placeholders
* packet_conn_wrappers: Initial changes
* packet_conn_wrappers: Unwrap a packet conn only if there are no wrappers
---------
Co-authored-by: Matt Holt <mholt@users.noreply.github.com>
* capture the buffered body once, then reset clonedReq.Body before each retry
* no copy
* keep receiver name
* set the buf to nil after extraction and only return it to pool if not nil
---------
Co-authored-by: WeidiDeng <weidi_deng@icloud.com>
* chore: ugh, lint fix...
Signed-off-by: Mohammed Al Sahaf <msaa1990@gmail.com>
* more lint fixes
Signed-off-by: Mohammed Al Sahaf <msaa1990@gmail.com>
---------
Signed-off-by: Mohammed Al Sahaf <msaa1990@gmail.com>
`pflag.GetStringSlice` treats commas as delimiters, which causes issues
when passing headers whose values contain commas (`X-Robots-Tag:
noindex, nofollow`). These are incorrectly split into multiple headers
and errors out:
- `X-Robots-Tag: noindex`
- ` nofollow`
Switch to `pflag.GetStringArray`, which does not split on commas[1].
Note that this changes behavior for cases where multiple headers were
provided in a single argument with commas (`--header-down "X-Foo:
Bar,X-Bar: Foo"`). Such cases will now be treated as a single header
value. If this breaking change is unacceptable, we will need a smarter
fallback mechanism.
[1] https://github.com/spf13/pflag/pull/90
* set Content-Length for precompressed files
* Update modules/caddyhttp/fileserver/staticfiles.go
Co-authored-by: Matt Holt <mholt@users.noreply.github.com>
---------
Co-authored-by: Matt Holt <mholt@users.noreply.github.com>
* use the new http.Protocols to handle h1, h2 and h2c requests
* fix lint
* keep ConnCtxKey for now
* fix handling for h2c
* check http version while reading the connection
* check if connection implements connectionStater when it should
* add comments about either h1 or h2 must be used in the listener
* fix if check
* return a net.Conn that implements connectionStater if applicable
* remove http/1.1 from alpn if h1 is disabled
* fix matching if only h1 is enabled
---------
Co-authored-by: Matt Holt <mholt@users.noreply.github.com>