Forward allowlisted query params for remote providers (#146)
Remote proxying drops the query string, so paginated JSON APIs (e.g. Galaxy `versions/?limit=&offset=`) cannot be served. Forwarding the raw query would let arbitrary params mint unbounded cache entries and bypass path-anchored blocklists. - add `provider.QueryForwarder`, mapping each forwarded param for a path to a validator that canonicalises or rejects its value - forward only validated params upstream and into the cache key, single-valued and in sorted order - classify on the bare path so blocklist/allowlist/mutability patterns ignore the query - evict a path together with its `path?...` query variants - leave providers without `QueryForwarder` unchanged Reviewed-on: #146 Co-authored-by: unkin-agent <unkin-agent@unkin.net> Co-committed-by: unkin-agent <unkin-agent@unkin.net>
This commit was merged in pull request #146.
This commit is contained in:
@@ -0,0 +1,58 @@
|
||||
package v1
|
||||
|
||||
import (
|
||||
"net/url"
|
||||
"strings"
|
||||
"testing"
|
||||
|
||||
"git.unkin.net/unkin/artifactapi/internal/provider"
|
||||
)
|
||||
|
||||
type plainProvider struct{ provider.Provider }
|
||||
|
||||
type forwardingProvider struct{ plainProvider }
|
||||
|
||||
func digits(v string) (string, bool) {
|
||||
if v == "" || strings.Trim(v, "0123456789") != "" {
|
||||
return "", false
|
||||
}
|
||||
return strings.TrimLeft(v, "0"), true
|
||||
}
|
||||
|
||||
func (forwardingProvider) ForwardedQuery(path string) map[string]provider.QueryValidator {
|
||||
if strings.HasSuffix(path, ".tar.gz") {
|
||||
return nil
|
||||
}
|
||||
return map[string]provider.QueryValidator{"limit": digits, "offset": digits}
|
||||
}
|
||||
|
||||
func TestForwardQuery(t *testing.T) {
|
||||
tests := []struct {
|
||||
name string
|
||||
prov provider.Provider
|
||||
path string
|
||||
query string
|
||||
want string
|
||||
}{
|
||||
{"non-forwarder keeps path", plainProvider{}, "a/versions/", "limit=10&offset=5", "a/versions/"},
|
||||
{"no query", forwardingProvider{}, "a/versions/", "", "a/versions/"},
|
||||
{"allowlisted params sorted", forwardingProvider{}, "a/versions/", "offset=5&limit=10", "a/versions/?limit=10&offset=5"},
|
||||
{"unknown params dropped", forwardingProvider{}, "a/versions/", "x=1&limit=10&cb=zz", "a/versions/?limit=10"},
|
||||
{"only unknown params", forwardingProvider{}, "a/versions/", "x=1", "a/versions/"},
|
||||
{"first value wins", forwardingProvider{}, "a/versions/", "limit=10&limit=20", "a/versions/?limit=10"},
|
||||
{"invalid value dropped", forwardingProvider{}, "a/versions/", "limit=1%202&offset=5", "a/versions/?offset=5"},
|
||||
{"value canonicalised", forwardingProvider{}, "a/versions/", "limit=0010", "a/versions/?limit=10"},
|
||||
{"path without allowlist", forwardingProvider{}, "a/b-1.0.0.tar.gz", "limit=10", "a/b-1.0.0.tar.gz"},
|
||||
}
|
||||
for _, tt := range tests {
|
||||
t.Run(tt.name, func(t *testing.T) {
|
||||
q, err := url.ParseQuery(tt.query)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if got := forwardQuery(tt.prov, tt.path, q); got != tt.want {
|
||||
t.Errorf("forwardQuery = %q, want %q", got, tt.want)
|
||||
}
|
||||
})
|
||||
}
|
||||
}
|
||||
@@ -8,6 +8,7 @@ import (
|
||||
"log/slog"
|
||||
"net"
|
||||
"net/http"
|
||||
"net/url"
|
||||
"strconv"
|
||||
"strings"
|
||||
|
||||
@@ -115,7 +116,7 @@ func (h *ProxyHandler) handleProxy(w http.ResponseWriter, r *http.Request) {
|
||||
}
|
||||
|
||||
func (h *ProxyHandler) serveFetch(w http.ResponseWriter, r *http.Request, remote *models.Remote, prov provider.Provider, path string) {
|
||||
result, err := h.engine.Fetch(r.Context(), *remote, path, prov, r.Header)
|
||||
result, err := h.engine.Fetch(r.Context(), *remote, forwardQuery(prov, path, r.URL.Query()), prov, r.Header)
|
||||
if err != nil {
|
||||
var proxyErr *proxy.ProxyError
|
||||
if errors.As(err, &proxyErr) {
|
||||
@@ -325,6 +326,25 @@ func Scheme(r *http.Request) string {
|
||||
return "http"
|
||||
}
|
||||
|
||||
// forwardQuery appends the provider's validated query params to path in a
|
||||
// canonical (sorted, single-valued) form, since the result is the cache key.
|
||||
func forwardQuery(prov provider.Provider, path string, query url.Values) string {
|
||||
f, ok := prov.(provider.QueryForwarder)
|
||||
if !ok {
|
||||
return path
|
||||
}
|
||||
kept := url.Values{}
|
||||
for name, valid := range f.ForwardedQuery(path) {
|
||||
if v, ok := valid(query.Get(name)); ok {
|
||||
kept.Set(name, v)
|
||||
}
|
||||
}
|
||||
if len(kept) == 0 {
|
||||
return path
|
||||
}
|
||||
return path + "?" + kept.Encode()
|
||||
}
|
||||
|
||||
// rewriteServed rewrites a cached body for the requesting host and marks the
|
||||
// response uncacheable by shared caches, since it varies per Host.
|
||||
func rewriteServed(w http.ResponseWriter, r *http.Request, sr provider.ServeRewriter, remote models.Remote, src io.Reader) ([]byte, bool) {
|
||||
|
||||
Reference in New Issue
Block a user