v2: Implement Caddyfile enhancements (breaking changes) (#2960)

* http: path matcher: exact match by default; substring matches (#2959)

This is a breaking change.

* caddyfile: Change "matcher" directive to "@matcher" syntax (#2959)

* cmd: Assume caddyfile adapter for config files named Caddyfile

* Sub-sort handlers by path matcher length (#2959)

Caddyfile-generated subroutes have handlers, which are sorted first by
directive order (this is unchanged), but within directives we now sort
by specificity of path matcher in descending order (longest path first,
assuming that longest path is most specific).

This only applies if there is only one matcher set, and the path
matcher in that set has only one path in it. Path matchers with two or
more paths are not sorted like this; and routes with more than one
matcher set are not sorted like this either, since specificity is
difficult or impossible to infer correctly.

This is a special case, but definitely a very common one, as a lot of
routing decisions are based on paths.

* caddyfile: New 'route' directive for appearance-order handling (#2959)

* caddyfile: Make rewrite directives mutually exclusive (#2959)

This applies only to rewrites in the top-level subroute created by the
HTTP caddyfile.

2 files changed, 61 insertions, 16 deletions

diff --git a/modules/caddyhttp/matchers.go b/modules/caddyhttp/matchers.go
index 40a767f..f63e48e 100644
--- a/modules/caddyhttp/matchers.go
+++ b/modules/caddyhttp/matchers.go
@@ -46,7 +46,17 @@ type (
 	// [customized or disabled](/docs/json/apps/http/servers/automatic_https/).
 	MatchHost []string
 
-	// MatchPath matches requests by the URI's path (case-insensitive).
+	// MatchPath matches requests by the URI's path (case-insensitive). Path
+	// matches are exact, but wildcards may be used:
+	//
+	// - At the end, for a prefix match (`/prefix/*`)
+	// - At the beginning, for a suffix match (`*.suffix`)
+	// - On both sides, for a substring match (`*/contains/*`)
+	// - In the middle, for a globular match (`/accounts/*/info`)
+	//
+	// This matcher is fast, so it does not support regular expressions or
+	// capture groups. For slower but more capable matching, use the path_regexp
+	// matcher.
 	MatchPath []string
 
 	// MatchPathRE matches requests by a regular expression on the URI's path.
@@ -197,11 +207,15 @@ func (m MatchPath) Match(r *http.Request) bool {
 	// being matched by *.php to be treated as PHP scripts
 	lowerPath = strings.TrimRight(lowerPath, ". ")
 
+	repl := r.Context().Value(caddy.ReplacerCtxKey).(*caddy.Replacer)
+
 	for _, matchPath := range m {
-		// special case: first character is equals sign,
-		// treat it as an exact match
-		if strings.HasPrefix(matchPath, "=") {
-			if lowerPath == matchPath[1:] {
+		matchPath = repl.ReplaceAll(matchPath, "")
+
+		// special case: first and last characters are wildcard,
+		// treat it as a fast substring match
+		if strings.HasPrefix(matchPath, "*") && strings.HasSuffix(matchPath, "*") {
+			if strings.Contains(lowerPath, matchPath[1:len(matchPath)-1]) {
 				return true
 			}
 			continue
@@ -216,15 +230,22 @@ func (m MatchPath) Match(r *http.Request) bool {
 			continue
 		}
 
+		// special case: last character is a wildcard,
+		// treat it as a fast prefix match
+		if strings.HasSuffix(matchPath, "*") {
+			if strings.HasPrefix(lowerPath, matchPath[:len(matchPath)-1]) {
+				return true
+			}
+			continue
+		}
+
+		// for everything else, try globular matching, which also
+		// is exact matching if there are no glob/wildcard chars;
 		// can ignore error here because we can't handle it anyway
 		matches, _ := filepath.Match(matchPath, lowerPath)
 		if matches {
 			return true
 		}
-
-		if strings.HasPrefix(lowerPath, matchPath) {
-			return true
-		}
 	}
 	return false
 }
diff --git a/modules/caddyhttp/matchers_test.go b/modules/caddyhttp/matchers_test.go
index 34a1647..8e06546 100644
--- a/modules/caddyhttp/matchers_test.go
+++ b/modules/caddyhttp/matchers_test.go
@@ -183,8 +183,18 @@ func TestPathMatcher(t *testing.T) {
 			expect: false,
 		},
 		{
+			match:  MatchPath{"/foo/bar/"},
+			input:  "/foo/bar/",
+			expect: true,
+		},
+		{
 			match:  MatchPath{"/foo/bar/", "/other"},
 			input:  "/other/",
+			expect: false,
+		},
+		{
+			match:  MatchPath{"/foo/bar/", "/other"},
+			input:  "/other",
 			expect: true,
 		},
 		{
@@ -213,19 +223,19 @@ func TestPathMatcher(t *testing.T) {
 			expect: false,
 		},
 		{
-			match:  MatchPath{"=/foo"},
-			input:  "/foo",
+			match:  MatchPath{"*substring*"},
+			input:  "/foo/substring/bar.txt",
 			expect: true,
 		},
 		{
-			match:  MatchPath{"=/foo"},
+			match:  MatchPath{"/foo"},
 			input:  "/foo/bar",
 			expect: false,
 		},
 		{
-			match:  MatchPath{"=/foo"},
-			input:  "/FOO",
-			expect: true,
+			match:  MatchPath{"/foo"},
+			input:  "/foo/bar",
+			expect: false,
 		},
 		{
 			match:  MatchPath{"/foo"},
@@ -233,12 +243,21 @@ func TestPathMatcher(t *testing.T) {
 			expect: true,
 		},
 		{
+			match:  MatchPath{"/foo*"},
+			input:  "/FOOOO",
+			expect: true,
+		},
+		{
 			match:  MatchPath{"/foo/bar.txt"},
 			input:  "/foo/BAR.txt",
 			expect: true,
 		},
 	} {
 		req := &http.Request{URL: &url.URL{Path: tc.input}}
+		repl := caddy.NewReplacer()
+		ctx := context.WithValue(req.Context(), caddy.ReplacerCtxKey, repl)
+		req = req.WithContext(ctx)
+
 		actual := tc.match.Match(req)
 		if actual != tc.expect {
 			t.Errorf("Test %d %v: Expected %t, got %t for '%s'", i, tc.match, tc.expect, actual, tc.input)
@@ -251,8 +270,13 @@ func TestPathMatcherWindows(t *testing.T) {
 	// only Windows has this bug where it will ignore
 	// trailing dots and spaces in a filename, but we
 	// test for it on all platforms to be more consistent
-	match := MatchPath{"*.php"}
+
 	req := &http.Request{URL: &url.URL{Path: "/index.php . . .."}}
+	repl := caddy.NewReplacer()
+	ctx := context.WithValue(req.Context(), caddy.ReplacerCtxKey, repl)
+	req = req.WithContext(ctx)
+
+	match := MatchPath{"*.php"}
 	matched := match.Match(req)
 	if !matched {
 		t.Errorf("Expected to match; should ignore trailing dots and spaces")