nghttpx: Wildcard path matching

This commit adds wildcard path matching.  If path pattern given in
backend option ends with "*", it is considered as wildcard path.  "*"
must match at least one character.  All paths which include wildcard
path without last "*" as prefix, and are strictly longer than wildcard
path without last "*" are matched.

src/shrpx-unittest.cc

@@ -130,6 +130,8 @@ int main(int argc, char *argv[]) {
       !CU_add_test(pSuite, "http_create_via_header_value",
                    shrpx::test_shrpx_http_create_via_header_value) ||
       !CU_add_test(pSuite, "router_match", shrpx::test_shrpx_router_match) ||
+      !CU_add_test(pSuite, "router_match_wildcard",
+                   shrpx::test_shrpx_router_match_wildcard) ||
       !CU_add_test(pSuite, "router_match_prefix",
                    shrpx::test_shrpx_router_match_prefix) ||
       !CU_add_test(pSuite, "util_streq", shrpx::test_util_streq) ||

src/shrpx.cc

@@ -1688,6 +1688,18 @@ Connections:
               match  against  "nghttp2.org".   The exact  hosts  match
               takes precedence over the wildcard hosts match.
 
+              If path  part ends with  "*", it is treated  as wildcard
+              path.  The  wildcard path  behaves differently  from the
+              normal path.  For normal path,  match is made around the
+              boundary of path component  separator,"/".  On the other
+              hand, the wildcard  path does not take  into account the
+              path component  separator.  All paths which  include the
+              wildcard  path  without  last  "*" as  prefix,  and  are
+              strictly longer than wildcard  path without last "*" are
+              matched.  "*"  must match  at least one  character.  For
+              example,  the   pattern  "/foo*"  matches   "/foo/"  and
+              "/foobar".  But it does not match "/foo", or "/fo".
+
               If <PATTERN> is omitted or  empty string, "/" is used as
               pattern,  which  matches  all request  paths  (catch-all
               pattern).  The catch-all backend must be given.

src/shrpx_config.cc

@@ -967,6 +967,12 @@ int parse_mapping(Config *config, DownstreamAddrConfig &addr,
       auto host = StringRef{std::begin(g.pattern) + 1, path_first};
       auto path = StringRef{path_first, std::end(g.pattern)};
 
+      auto path_is_wildcard = false;
+      if (path[path.size() - 1] == '*') {
+        path = StringRef{std::begin(path), std::begin(path) + path.size() - 1};
+        path_is_wildcard = true;
+      }
+
       auto it = std::find_if(
           std::begin(wildcard_patterns), std::end(wildcard_patterns),
           [&host](const WildcardPattern &wp) { return wp.host == host; });
@@ -975,7 +981,7 @@ int parse_mapping(Config *config, DownstreamAddrConfig &addr,
         wildcard_patterns.emplace_back(host);
 
         auto &router = wildcard_patterns.back().router;
-        router.add_route(path, idx);
+        router.add_route(path, idx, path_is_wildcard);
 
         auto iov = make_byte_ref(downstreamconf.balloc, host.size() + 1);
         auto p = iov.base;
@@ -985,13 +991,20 @@ int parse_mapping(Config *config, DownstreamAddrConfig &addr,
 
         rw_router.add_route(rev_host, wildcard_patterns.size() - 1);
       } else {
-        (*it).router.add_route(path, idx);
+        (*it).router.add_route(path, idx, path_is_wildcard);
       }
 
       continue;
     }
 
-    router.add_route(g.pattern, idx);
+    auto path_is_wildcard = false;
+    if (pattern[pattern.size() - 1] == '*') {
+      pattern = StringRef{std::begin(pattern),
+                          std::begin(pattern) + pattern.size() - 1};
+      path_is_wildcard = true;
+    }
+
+    router.add_route(pattern, idx, path_is_wildcard);
   }
   return 0;
 }

src/shrpx_router.cc

@@ -31,10 +31,10 @@
 
 namespace shrpx {
 
-RNode::RNode() : s(nullptr), len(0), index(-1) {}
+RNode::RNode() : s(nullptr), len(0), index(-1), wildcard_index(-1) {}
 
-RNode::RNode(const char *s, size_t len, size_t index)
+RNode::RNode(const char *s, size_t len, ssize_t index, ssize_t wildcard_index)
-    : s(s), len(len), index(index) {}
+    : s(s), len(len), index(index), wildcard_index(wildcard_index) {}
 
 Router::Router() : balloc_(1024, 1024), root_{} {}
 
@@ -64,21 +64,30 @@ void add_next_node(RNode *node, std::unique_ptr<RNode> new_node) {
 } // namespace
 
 void Router::add_node(RNode *node, const char *pattern, size_t patlen,
-                      size_t index) {
+                      ssize_t index, ssize_t wildcard_index) {
   auto pat = make_string_ref(balloc_, StringRef{pattern, patlen});
-  auto new_node = make_unique<RNode>(pat.c_str(), pat.size(), index);
+  auto new_node =
+      make_unique<RNode>(pat.c_str(), pat.size(), index, wildcard_index);
   add_next_node(node, std::move(new_node));
 }
 
-size_t Router::add_route(const StringRef &pattern, size_t index) {
+size_t Router::add_route(const StringRef &pattern, size_t idx, bool wildcard) {
+  ssize_t index = -1, wildcard_index = -1;
+  if (wildcard) {
+    wildcard_index = idx;
+  } else {
+    index = idx;
+  }
+
   auto node = &root_;
   size_t i = 0;
 
   for (;;) {
     auto next_node = find_next_node(node, pattern[i]);
     if (next_node == nullptr) {
-      add_node(node, pattern.c_str() + i, pattern.size() - i, index);
+      add_node(node, pattern.c_str() + i, pattern.size() - i, index,
-      return index;
+               wildcard_index);
+      return idx;
     }
 
     node = next_node;
@@ -93,12 +102,22 @@ size_t Router::add_route(const StringRef &pattern, size_t index) {
       // The common prefix was matched
       if (slen == node->len) {
         // Complete match
+        if (index != -1) {
           if (node->index != -1) {
             // Return the existing index for duplicates.
             return node->index;
           }
           node->index = index;
-        return index;
+          return idx;
+        }
+
+        assert(wildcard_index != -1);
+
+        if (node->wildcard_index != -1) {
+          return node->wildcard_index;
+        }
+        node->wildcard_index = wildcard_index;
+        return idx;
       }
 
       if (slen > node->len) {
@@ -112,27 +131,30 @@ size_t Router::add_route(const StringRef &pattern, size_t index) {
     if (node->len > j) {
       // node must be split into 2 nodes.  new_node is now the child
       // of node.
-      auto new_node =
+      auto new_node = make_unique<RNode>(&node->s[j], node->len - j,
-          make_unique<RNode>(&node->s[j], node->len - j, node->index);
+                                         node->index, node->wildcard_index);
       std::swap(node->next, new_node->next);
 
       node->len = j;
       node->index = -1;
+      node->wildcard_index = -1;
 
       add_next_node(node, std::move(new_node));
 
       if (slen == j) {
         node->index = index;
-        return index;
+        node->wildcard_index = wildcard_index;
+        return idx;
       }
     }
 
     i += j;
 
     assert(pattern.size() > i);
-    add_node(node, pattern.c_str() + i, pattern.size() - i, index);
+    add_node(node, pattern.c_str() + i, pattern.size() - i, index,
+             wildcard_index);
 
-    return index;
+    return idx;
   }
 }
 
@@ -169,8 +191,10 @@ const RNode *match_complete(size_t *offset, const RNode *node,
 } // namespace
 
 namespace {
-const RNode *match_partial(const RNode *node, size_t offset, const char *first,
+const RNode *match_partial(bool *pattern_is_wildcard, const RNode *node,
-                           const char *last) {
+                           size_t offset, const char *first, const char *last) {
+  *pattern_is_wildcard = false;
+
   if (first == last) {
     if (node->len == offset) {
       return node;
@@ -207,8 +231,12 @@ const RNode *match_partial(const RNode *node, size_t offset, const char *first,
       return nullptr;
     }
 
-    if (node->index != -1 && node->s[node->len - 1] == '/') {
+    if (node->wildcard_index != -1) {
       found_node = node;
+      *pattern_is_wildcard = true;
+    } else if (node->index != -1 && node->s[node->len - 1] == '/') {
+      found_node = node;
+      *pattern_is_wildcard = false;
     }
 
     assert(node->len == offset + n);
@@ -233,6 +261,7 @@ const RNode *match_partial(const RNode *node, size_t offset, const char *first,
       if (node->len == n) {
         // Complete match with this node
         if (node->index != -1) {
+          *pattern_is_wildcard = false;
           return node;
         }
 
@@ -246,16 +275,21 @@ const RNode *match_partial(const RNode *node, size_t offset, const char *first,
       // pattern is "/foo/" and path is "/foo", we consider they
       // match.
       if (node->index != -1 && n + 1 == node->len && node->s[n] == '/') {
+        *pattern_is_wildcard = false;
         return node;
       }
 
       return found_node;
     }
 
+    if (node->wildcard_index != -1) {
+      found_node = node;
+      *pattern_is_wildcard = true;
+    } else if (node->index != -1 && node->s[node->len - 1] == '/') {
       // This is the case when pattern which ends with "/" is included
       // in query.
-    if (node->index != -1 && node->s[node->len - 1] == '/') {
       found_node = node;
+      *pattern_is_wildcard = false;
     }
 
     assert(node->len == n);
@@ -272,12 +306,14 @@ ssize_t Router::match(const StringRef &host, const StringRef &path) const {
     return -1;
   }
 
-  node = match_partial(node, offset, std::begin(path), std::end(path));
+  bool pattern_is_wildcard;
+  node = match_partial(&pattern_is_wildcard, node, offset, std::begin(path),
+                       std::end(path));
   if (node == nullptr || node == &root_) {
     return -1;
   }
 
-  return node->index;
+  return pattern_is_wildcard ? node->wildcard_index : node->index;
 }
 
 ssize_t Router::match(const StringRef &s) const {

src/shrpx_router.h

@@ -38,7 +38,7 @@ namespace shrpx {
 
 struct RNode {
   RNode();
-  RNode(const char *s, size_t len, size_t index);
+  RNode(const char *s, size_t len, ssize_t index, ssize_t wildcard_index);
   RNode(RNode &&) = default;
   RNode(const RNode &) = delete;
   RNode &operator=(RNode &&) = default;
@@ -54,6 +54,10 @@ struct RNode {
   // Index of pattern if match ends in this node.  Note that we don't
   // store duplicated pattern.
   ssize_t index;
+  // Index of wildcard pattern if query includes this node as prefix
+  // and it still has suffix to match.  Note that we don't store
+  // duplicated pattern.
+  ssize_t wildcard_index;
 };
 
 class Router {
@@ -66,8 +70,13 @@ public:
   Router &operator=(const Router &) = delete;
 
   // Adds route |pattern| with its |index|.  If same pattern has
-  // already been added, the existing index is returned.
+  // already been added, the existing index is returned.  If
-  size_t add_route(const StringRef &pattern, size_t index);
+  // |wildcard| is true, |pattern| is considered as wildcard pattern,
+  // and all paths which have the |pattern| as prefix and are strictly
+  // longer than |pattern| match.  The wildcard pattern only works
+  // with match(const StringRef&, const StringRef&).
+  size_t add_route(const StringRef &pattern, size_t index,
+                   bool wildcard = false);
   // Returns the matched index of pattern.  -1 if there is no match.
   ssize_t match(const StringRef &host, const StringRef &path) const;
   // Returns the matched index of pattern |s|.  -1 if there is no
@@ -84,7 +93,8 @@ public:
   ssize_t match_prefix(size_t *nread, const RNode **last_node,
                        const StringRef &s) const;
 
-  void add_node(RNode *node, const char *pattern, size_t patlen, size_t index);
+  void add_node(RNode *node, const char *pattern, size_t patlen, ssize_t index,
+                ssize_t wildcard_index);
 
   void dump() const;
 

src/shrpx_router_test.cc

@@ -33,6 +33,7 @@ namespace shrpx {
 struct Pattern {
   StringRef pattern;
   size_t idx;
+  bool wildcard;
 };
 
 void test_shrpx_router_match(void) {
@@ -88,6 +89,59 @@ void test_shrpx_router_match(void) {
   CU_ASSERT(5 == idx);
 }
 
+void test_shrpx_router_match_wildcard(void) {
+  constexpr auto patterns = std::array<Pattern, 6>{{
+      {StringRef::from_lit("nghttp2.org/"), 0},
+      {StringRef::from_lit("nghttp2.org/"), 1, true},
+      {StringRef::from_lit("nghttp2.org/alpha/"), 2},
+      {StringRef::from_lit("nghttp2.org/alpha/"), 3, true},
+      {StringRef::from_lit("nghttp2.org/bravo"), 4},
+      {StringRef::from_lit("nghttp2.org/bravo"), 5, true},
+  }};
+
+  Router router;
+
+  for (auto &p : patterns) {
+    router.add_route(p.pattern, p.idx, p.wildcard);
+  }
+
+  CU_ASSERT(0 ==
+            router.match(StringRef::from_lit("nghttp2.org"),
+                         StringRef::from_lit("/")));
+
+  CU_ASSERT(1 ==
+            router.match(StringRef::from_lit("nghttp2.org"),
+                         StringRef::from_lit("/a")));
+
+  CU_ASSERT(1 ==
+            router.match(StringRef::from_lit("nghttp2.org"),
+                         StringRef::from_lit("/charlie")));
+
+  CU_ASSERT(2 ==
+            router.match(StringRef::from_lit("nghttp2.org"),
+                         StringRef::from_lit("/alpha")));
+
+  CU_ASSERT(2 ==
+            router.match(StringRef::from_lit("nghttp2.org"),
+                         StringRef::from_lit("/alpha/")));
+
+  CU_ASSERT(3 ==
+            router.match(StringRef::from_lit("nghttp2.org"),
+                         StringRef::from_lit("/alpha/b")));
+
+  CU_ASSERT(4 ==
+            router.match(StringRef::from_lit("nghttp2.org"),
+                         StringRef::from_lit("/bravo")));
+
+  CU_ASSERT(5 ==
+            router.match(StringRef::from_lit("nghttp2.org"),
+                         StringRef::from_lit("/bravocharlie")));
+
+  CU_ASSERT(5 ==
+            router.match(StringRef::from_lit("nghttp2.org"),
+                         StringRef::from_lit("/bravo/")));
+}
+
 void test_shrpx_router_match_prefix(void) {
   auto patterns = std::vector<Pattern>{
       {StringRef::from_lit("gro.2ptthgn."), 0},

src/shrpx_router_test.h

@@ -32,6 +32,7 @@
 namespace shrpx {
 
 void test_shrpx_router_match(void);
+void test_shrpx_router_match_wildcard(void);
 void test_shrpx_router_match_prefix(void);
 
 } // namespace shrpx