Age | Commit message (Collapse) | Author |
|
|
|
fix a nil pointer dereference in AUpstreams.GetUpstreams when AUpstreams.Versions is not set (fixes caddyserver#5809)
Signed-off-by: Pascal Vorwerk <info@fossores.de>
|
|
These happen to be harmless memory aliasing
but I guess the linter can't know that and we
can't really prove it in general.
|
|
Co-authored-by: Mohammed Al Sahaf <msaa1990@gmail.com>
Co-authored-by: Francis Lavoie <lavofr@gmail.com>
|
|
* reverseproxy: Always return new upstreams (fix #5736)
* Fix healthcheck logger race
|
|
* use gofmput to format code
* use gci to format imports
* reconfigure gci
* linter autofixes
* rearrange imports a little
* export GOOS=windows golangci-lint run ./... --fix
|
|
|
|
replaceble parts (#5695)
* reverseproxy: do not parse upstream address too early if it contains replaceble parts
* remove unused method
* cleanup
* accommodate partially replaceable port
|
|
|
|
* caddyhttp: Make use of http.ResponseController
Also syncs the reverseproxy implementation with stdlib's which now uses ResponseController as well https://github.com/golang/go/commit/2449bbb5e614954ce9e99c8a481ea2ee73d72d61
* Enable full-duplex for HTTP/1.1
* Appease linter
* Add warning for builds with Go 1.20, so it's less surprising to users
* Improved godoc for EnableFullDuplex, copied text from stdlib
* Only wrap in encode if not already wrapped
|
|
|
|
allows AUpstreams to be instantiated externally
|
|
* update quic-go to v0.37.0
* Bump to Go 1.20
* Bump golangci-lint version, yml syntax consistency
* Use skip-pkg-cache workaround
* Workaround needed for both?
* Seeding weakrand is no longer necessary
---------
Co-authored-by: Matt Holt <mholt@users.noreply.github.com>
Co-authored-by: Francis Lavoie <lavofr@gmail.com>
|
|
* caddytls: Don't purge cert cache on config reload
* Update CertMagic
This actually avoids reloading managed certs from storage
when already in the cache, d'oh.
* Fix bug; re-implement HasCertificateForSubject
* Update go.mod: CertMagic tag
|
|
This avoids copying the Upstream, which has an atomically-accessed value
in it.
|
|
|
|
* added weighted round robin algorithm to load balancer
* added an adapt integration test for wrr and fixed a typo
* changed args format to Caddyfile args convention
* added provisioner and validator for wrr
* simplified the code and improved doc
|
|
* reverseproxy: WIP streaming timeouts
* More verbose logging by using the child logger
* reverseproxy: Implement streaming timeouts
* reverseproxy: Refactor cleanup
* reverseproxy: Avoid **time.Timer
---------
Co-authored-by: Francis Lavoie <lavofr@gmail.com>
|
|
|
|
(#5494)
* reverseproxy: Expand port ranges to multiple upstreams in CLI + Caddyfile
* Add clarifying comment
|
|
(#5487)
When only a single request has the least amount of requests, there's no need to compute a random number, because the modulo of 1 will always be 0 anyways.
|
|
|
|
|
|
|
|
* acmeserver: Configurable `resolvers`, fix smallstep deprecations
* Improve default net/port
* Update proxy resolvers parsing to use the new function
* Update listeners.go
Co-authored-by: itsxaos <33079230+itsxaos@users.noreply.github.com>
---------
Co-authored-by: itsxaos <33079230+itsxaos@users.noreply.github.com>
|
|
Co-authored-by: Francis Lavoie <lavofr@gmail.com>
|
|
Co-authored-by: Dávid Szabó <david.szabo97@gmail.com>
|
|
|
|
|
|
|
|
listener wrapper (#5424)
Co-authored-by: WeidiDeng <weidi_deng@icloud.com>
Co-authored-by: Francis Lavoie <lavofr@gmail.com>
|
|
Co-authored-by: Matt Holt <mholt@users.noreply.github.com>
|
|
* log: make `sink` encodable
* deduplicate logger fields
* extract common fields into `BaseLog` and embed it into `SinkLog`
* amend godoc on `BaseLog` and `SinkLog`
* minor style change
---------
Co-authored-by: Francis Lavoie <lavofr@gmail.com>
|
|
Also renamed the field to SubjectsRaw, which can be considered a breaking change but I don't expect this to affect much.
|
|
|
|
Fixes: https://github.com/caddyserver/caddy/issues/5420
|
|
Co-authored-by: Francis Lavoie <lavofr@gmail.com>
|
|
|
|
* cmd: Expand cobra support
* Convert commands to cobra, add short flags
* Fix version command typo
Co-authored-by: Emily Lange <git@indeednotjames.com>
* Apply suggestions from code review
Co-authored-by: Matt Holt <mholt@users.noreply.github.com>
---------
Co-authored-by: Emily Lange <git@indeednotjames.com>
Co-authored-by: Matt Holt <mholt@users.noreply.github.com>
|
|
Co-authored-by: Francis Lavoie <lavofr@gmail.com>
Co-authored-by: Weidi Deng <weidi_deng@icloud.com>
|
|
|
|
* reverseproxy: Don't buffer chunked requests (fix #5366)
Mostly reverts 845bc4d50b437995d574819850206e4b3db4040d (#5289)
Adds warning for unsafe config.
Deprecates unsafe properties in favor of simpler, safer designed ones.
* Update modules/caddyhttp/reverseproxy/caddyfile.go
Co-authored-by: Y.Horie <u5.horie@gmail.com>
* Update modules/caddyhttp/reverseproxy/reverseproxy.go
Co-authored-by: Y.Horie <u5.horie@gmail.com>
* Update modules/caddyhttp/reverseproxy/reverseproxy.go
Co-authored-by: Y.Horie <u5.horie@gmail.com>
* Remove unused code
---------
Co-authored-by: Y.Horie <u5.horie@gmail.com>
|
|
* added some tests for parseUpstreamDialAddress
Test 4 fails because it produces "[[::1]]:80" instead of "[::1]:80"
* support absolute windows path in unix reverse proxy address
* make IsUnixNetwork public, support +h2c and reuse it
* add new tests
|
|
* log response size for websocket request
* record size when using hijack bufio.Writer
|
|
Co-authored-by: Francis Lavoie <lavofr@gmail.com>
Fixes undefined
|
|
|
|
* Fixes #5236
* enable request body buffering in reverse proxy
when the request header has Transfer-Encoding: chunked
|
|
|
|
* reverseproxy: Set origreq in active health check
Fix #5281
* Oops; dereference Request
|
|
* If upstreams are all using same host but with different ports
ie:
foobar:4001
foobar:4002
foobar:4003
...
Because fnv-1a has not a good enough avalanche effect
Then the hostByHashing result is not well balanced over
all upstreams
As last byte FNV input tend to affect few bits, the idea is to change
the concatenation order between the key and the upstream strings
So the upstream last byte have more impact on hash diffusion
|