Age | Commit message (Collapse) | Author |
|
* reverseproxy: Enable error logging for connection upgrades
* reverseproxy: Change some of the error levels, unsugar
* Use unsugared log in one spot
Co-authored-by: Matthew Holt <mholt@users.noreply.github.com>
|
|
* add CookieHashSelection for session affinity
* add CookieHashSelection for session affinity
* register module
* reverse_proxy: Add and fix cookie lb_policy
* reverse_proxy: Manage hmac.write error on cookie hash selection
* reverse_proxy: fix some comments
* reverse_proxy: variable `cookieValue` is inside the else block
* reverse_proxy: Abstract duplicate nuanced logic of reservoir sampling into a function
* reverse_proxy: Set a default secret is indeed useless
* reverse_proxy: add configuration syntax for cookie lb_policy
* reverse_proxy: doc typo and improvement
Co-authored-by: utick <123liuqingdong@163.com>
|
|
* implement default values for header directive
closes #3804
* remove `set_default` header op and rely on "require" handler instead
This has the following advantages over the previous attempt:
- It does not introduce a new operation for headers, but rather nicely
extends over an existing feature in the header handler.
- It removes the need to specify the header as "deferred" because it is
already implicitely deferred by the use of the require handler. This
should be less confusing to the user.
* add integration test for header directive in caddyfile
* bubble up errors when parsing caddyfile header directive
* don't export unnecessarily and don't canonicalize headers unnecessarily
* fix response headers not passed in blocks
* caddyfile: fix clash when using default header in block
Each header is now set in a separate handler so that it doesn't clash
with other headers set/added/deleted in the same block.
* caddyhttp: New idle_timeout default of 5m
* reverseproxy: fix random hangs on http/2 requests with server push (#3875)
see https://github.com/golang/go/issues/42534
* Refactor and cleanup with improvements
* More specific link
Co-authored-by: Matthew Holt <mholt@users.noreply.github.com>
Co-authored-by: Денис Телюх <telyukh.denis@gmail.com>
|
|
Before, if there was an error in the error handler, we would not write a
status code, which resulted in Go writing a 200 for us by default, which
does not make sense when there's an error. Now we write the second
error's status if available, otherwise 500.
|
|
|
|
Uncovered in #3807
|
|
|
|
* Add Caddyfile support for request_body:
```
request_body {
max_size 10000000
}
```
* Improve Caddyfile parser for request_body module
* Remove unnecessary `continue`
* Add sample for caddyfile_adapt_test
|
|
* caddytls: Support multiple issuers
Defaults are Let's Encrypt and ZeroSSL.
There are probably bugs.
* Commit updated integration tests, d'oh
* Update go.mod
|
|
* nitpicks and small improvements in basicauth module
1:
roll two if statements into one, since err will be nil in the second case anyhow
2:
unlock cache mutex after reading the key, as this happens by-value and reduces code complexity
3:
switch cache sync.Mutex to sync.RWMutex for better concurrency on cache fast track
* allocate the right kind of mutex
|
|
|
|
Co-authored-by: Matt Holt <mholt@users.noreply.github.com>
Co-authored-by: Francis Lavoie <lavofr@gmail.com>
|
|
Also, turns out that `Add` on headers will work even if there's nothing there yet, so we can remove the condition I introduced in #3832
|
|
|
|
* fastcgi: Add timeouts support to Caddyfile adapter
* fastcgi: Use tabs instead of spaces
|
|
|
|
* fileserver: Improve and clarify file hiding logic
* Oops, forgot to run integration tests
* Make this one integration test OS-agnostic
* See if this appeases the Windows gods
* D'oh
|
|
Always follow the code path of hashing and comparing a plaintext
password even if the account is not found by the given username; this
ensures that similar CPU cycles are spent for both valid and invalid
usernames.
Thanks to @tylerlm for helping and looking into this!
|
|
|
|
|
|
* check if the host is a placeholder
* Update modules/caddyhttp/reverseproxy/caddyfile.go
Co-authored-by: Matt Holt <mholt@users.noreply.github.com>
|
|
We have users that have site blocks like *.*.tld with on-demand TLS
enabled. While *.*.tld does not qualify for a publicly-trusted cert due
to its wildcards, On-Demand TLS does not actually obtain a cert with
those wildcards, since it uses the actual hostname on the handshake.
This improves on that logic, but I am still not 100% satisfied with the
result since I think we need to also check if another site block is more
specific, like foo.example.tld, which might not have on-demand TLS
enabled, and make sure an automation policy gets created before the
more general policy with on-demand...
|
|
* caddyhttp: Restore original request params before error handlers
Fixes #3717
* Add comment
|
|
* reverseproxy: Fix dial placeholders, SRV, active health checks
Supercedes #3776
Partially reverts or updates #3756, #3693, and #3695
* reverseproxy: add integration tests
Co-authored-by: Mohammed Al Sahaf <msaa1990@gmail.com>
|
|
|
|
|
|
|
|
* reverseproxy: fix breakage in handling SRV lookup introduced by 3695
* reverseproxy: validate against incompatible config options with lookup_srv
* reverseproxy: add integration test cases for validations involving lookup_srv
* reverseproxy: clarify the reason for skipping an iteration
* grammar.. Oxford comma
Co-authored-by: Francis Lavoie <lavofr@gmail.com>
Co-authored-by: Francis Lavoie <lavofr@gmail.com>
Fixes #3753
|
|
Fixes #3763
|
|
* reverseproxy: default to port 80 for port-less upstream dial addresses
* reverseproxy: replace integration test with an adapter test
Fixes #3761
|
|
Without this change, a Content-Type header like "text/event-stream;charset=utf-8"
would not trigger the immediate flushing.
Fixes #3765
|
|
Signed-off-by: Dave Henderson <dhenderson@gmail.com>
|
|
* caddyfile: support vars and vars_regexp matchers in the caddyfile
* caddyfile: matchers: Brian Kernighan said printf is good debugging tool but didn't say keep them around
|
|
* metrics: Always track method label in uppercase
Signed-off-by: Dave Henderson <dhenderson@gmail.com>
* Just use strings.ToUpper for clarity
Signed-off-by: Dave Henderson <dhenderson@gmail.com>
|
|
Signed-off-by: Dave Henderson <dhenderson@gmail.com>
|
|
* metrics: Fixing panic while observing with bad exemplars
Signed-off-by: Dave Henderson <dhenderson@gmail.com>
* Minor cleanup
The server is already added to the context. So, we can simply use that
to get the server name, which is a field on the server.
* Add integration test for auto HTTP->HTTPS redirects
A test like this would have caught the problem in the first place
Co-authored-by: Matthew Holt <mholt@users.noreply.github.com>
|
|
For some reason this breaks automatic HTTP->HTTPS redirects. I am not
sure why yet, but as a hotfix remove this until we understand it better.
|
|
Signed-off-by: Dave Henderson <dhenderson@gmail.com>
|
|
* reverseproxy: construct active health-check transport from scratch (Fixes #3691)
* reverseproxy: do upstream health-check on the correct alternative port
* reverseproxy: add integration test for health-check on alternative port
* reverseproxy: put back the custom transport for health-check http client
* reverseproxy: cleanup health-check integration test
* reverseproxy: fix health-check of unix socket upstreams
* reverseproxy: skip unix socket tests on Windows
* tabs > spaces
Co-authored-by: Francis Lavoie <lavofr@gmail.com>
* make the linter (and @francislavoie) happy
Co-authored-by: Francis Lavoie <lavofr@gmail.com>
* One more lint fix
Co-authored-by: Francis Lavoie <lavofr@gmail.com>
Co-authored-by: Francis Lavoie <lavofr@gmail.com>
|
|
* reverse_proxy: ensure upstream address has port range of only 1
* reverse_proxy: don't log the error if upstream range size is more than 1
|
|
* fileserver: Fix try_files for directories, windows fix
* fileserver: Add new file type placeholder, refactoring, tests
* fileserver: Review cleanup
* fileserver: Flip the return args order
|
|
* Fix-3585: added placeholder for a PEM encoded value of the certificate
* Update modules/caddyhttp/replacer.go
Change type of block and empty headers removed
Co-authored-by: Matt Holt <mholt@users.noreply.github.com>
* fixed tests
Co-authored-by: Matt Holt <mholt@users.noreply.github.com>
|
|
* logging: Implement Caddyfile support for filter encoder
* logging: Add support for parsing IP masks from strings
wip
* logging: Implement Caddyfile support for ip_mask
* logging: Get rid of unnecessary logic to allow strings, not that useful
* logging: Add adapt test
|
|
|
|
Now, a filename to hide that is specified without a path separator will
count as hidden if it appears in any component of the file path (not
only the last component); semantically, this means hiding a file by only
its name (without any part of a path) will hide both files and folders,
e.g. hiding ".git" will hide "/.git" and also "/.git/foo".
We also do prefix matching so that hiding "/.git" will hide "/.git"
and "/.git/foo" but not "/.gitignore".
The remaining logic is a globular match like before.
|
|
Reported in:
https://caddy.community/t/dns-challenge-with-namecheap-and-split-horizon-dns/9611/17?u=matt
|
|
|
|
|
|
|
|
|