summaryrefslogtreecommitdiffstats
path: root/src/shrpx_router.cc
diff options
context:
space:
mode:
Diffstat (limited to '')
-rw-r--r--src/shrpx_router.cc420
1 files changed, 420 insertions, 0 deletions
diff --git a/src/shrpx_router.cc b/src/shrpx_router.cc
new file mode 100644
index 0000000..d3565db
--- /dev/null
+++ b/src/shrpx_router.cc
@@ -0,0 +1,420 @@
+/*
+ * nghttp2 - HTTP/2 C Library
+ *
+ * Copyright (c) 2015 Tatsuhiro Tsujikawa
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining
+ * a copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sublicense, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be
+ * included in all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
+ * LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ * OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ * WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+#include "shrpx_router.h"
+
+#include <algorithm>
+
+#include "shrpx_config.h"
+#include "shrpx_log.h"
+
+namespace shrpx {
+
+RNode::RNode() : s(nullptr), len(0), index(-1), wildcard_index(-1) {}
+
+RNode::RNode(const char *s, size_t len, ssize_t index, ssize_t wildcard_index)
+ : s(s), len(len), index(index), wildcard_index(wildcard_index) {}
+
+Router::Router() : balloc_(1024, 1024), root_{} {}
+
+Router::~Router() {}
+
+namespace {
+RNode *find_next_node(const RNode *node, char c) {
+ auto itr = std::lower_bound(std::begin(node->next), std::end(node->next), c,
+ [](const std::unique_ptr<RNode> &lhs,
+ const char c) { return lhs->s[0] < c; });
+ if (itr == std::end(node->next) || (*itr)->s[0] != c) {
+ return nullptr;
+ }
+
+ return (*itr).get();
+}
+} // namespace
+
+namespace {
+void add_next_node(RNode *node, std::unique_ptr<RNode> new_node) {
+ auto itr = std::lower_bound(std::begin(node->next), std::end(node->next),
+ new_node->s[0],
+ [](const std::unique_ptr<RNode> &lhs,
+ const char c) { return lhs->s[0] < c; });
+ node->next.insert(itr, std::move(new_node));
+}
+} // namespace
+
+void Router::add_node(RNode *node, const char *pattern, size_t patlen,
+ ssize_t index, ssize_t wildcard_index) {
+ auto pat = make_string_ref(balloc_, StringRef{pattern, patlen});
+ auto new_node =
+ std::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 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,
+ wildcard_index);
+ return idx;
+ }
+
+ node = next_node;
+
+ auto slen = pattern.size() - i;
+ auto s = pattern.c_str() + i;
+ auto n = std::min(node->len, slen);
+ size_t j;
+ for (j = 0; j < n && node->s[j] == s[j]; ++j)
+ ;
+ if (j == n) {
+ // 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 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) {
+ // We still have pattern to add
+ i += j;
+
+ continue;
+ }
+ }
+
+ if (node->len > j) {
+ // node must be split into 2 nodes. new_node is now the child
+ // of node.
+ auto new_node = std::make_unique<RNode>(
+ &node->s[j], node->len - j, 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;
+ node->wildcard_index = wildcard_index;
+ return idx;
+ }
+ }
+
+ i += j;
+
+ assert(pattern.size() > i);
+ add_node(node, pattern.c_str() + i, pattern.size() - i, index,
+ wildcard_index);
+
+ return idx;
+ }
+}
+
+namespace {
+const RNode *match_complete(size_t *offset, const RNode *node,
+ const char *first, const char *last) {
+ *offset = 0;
+
+ if (first == last) {
+ return node;
+ }
+
+ auto p = first;
+
+ for (;;) {
+ auto next_node = find_next_node(node, *p);
+ if (next_node == nullptr) {
+ return nullptr;
+ }
+
+ node = next_node;
+
+ auto n = std::min(node->len, static_cast<size_t>(last - p));
+ if (memcmp(node->s, p, n) != 0) {
+ return nullptr;
+ }
+ p += n;
+ if (p == last) {
+ *offset = n;
+ return node;
+ }
+ }
+}
+} // namespace
+
+namespace {
+const RNode *match_partial(bool *pattern_is_wildcard, const RNode *node,
+ size_t offset, const char *first, const char *last) {
+ *pattern_is_wildcard = false;
+
+ if (first == last) {
+ if (node->len == offset) {
+ return node;
+ }
+ return nullptr;
+ }
+
+ auto p = first;
+
+ const RNode *found_node = nullptr;
+
+ if (offset > 0) {
+ auto n = std::min(node->len - offset, static_cast<size_t>(last - first));
+ if (memcmp(node->s + offset, first, n) != 0) {
+ return nullptr;
+ }
+
+ p += n;
+
+ if (p == last) {
+ if (node->len == offset + n) {
+ if (node->index != -1) {
+ return node;
+ }
+
+ // The last '/' handling, see below.
+ node = find_next_node(node, '/');
+ if (node != nullptr && node->index != -1 && node->len == 1) {
+ return node;
+ }
+
+ return nullptr;
+ }
+
+ // The last '/' handling, see below.
+ if (node->index != -1 && offset + n + 1 == node->len &&
+ node->s[node->len - 1] == '/') {
+ return node;
+ }
+
+ return nullptr;
+ }
+
+ 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);
+ }
+
+ for (;;) {
+ auto next_node = find_next_node(node, *p);
+ if (next_node == nullptr) {
+ return found_node;
+ }
+
+ node = next_node;
+
+ auto n = std::min(node->len, static_cast<size_t>(last - p));
+ if (memcmp(node->s, p, n) != 0) {
+ return found_node;
+ }
+
+ p += n;
+
+ if (p == last) {
+ if (node->len == n) {
+ // Complete match with this node
+ if (node->index != -1) {
+ *pattern_is_wildcard = false;
+ return node;
+ }
+
+ // The last '/' handling, see below.
+ node = find_next_node(node, '/');
+ if (node != nullptr && node->index != -1 && node->len == 1) {
+ *pattern_is_wildcard = false;
+ return node;
+ }
+
+ return found_node;
+ }
+
+ // We allow match without trailing "/" at the end of pattern.
+ // So, if pattern ends with '/', and pattern and path matches
+ // without that slash, we consider they match to deal with
+ // request to the directory without trailing slash. That is if
+ // 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.
+ found_node = node;
+ *pattern_is_wildcard = false;
+ }
+
+ assert(node->len == n);
+ }
+}
+} // namespace
+
+ssize_t Router::match(const StringRef &host, const StringRef &path) const {
+ const RNode *node;
+ size_t offset;
+
+ node = match_complete(&offset, &root_, std::begin(host), std::end(host));
+ if (node == nullptr) {
+ return -1;
+ }
+
+ 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 pattern_is_wildcard ? node->wildcard_index : node->index;
+}
+
+ssize_t Router::match(const StringRef &s) const {
+ const RNode *node;
+ size_t offset;
+
+ node = match_complete(&offset, &root_, std::begin(s), std::end(s));
+ if (node == nullptr) {
+ return -1;
+ }
+
+ if (node->len != offset) {
+ return -1;
+ }
+
+ return node->index;
+}
+
+namespace {
+const RNode *match_prefix(size_t *nread, const RNode *node, const char *first,
+ const char *last) {
+ if (first == last) {
+ return nullptr;
+ }
+
+ auto p = first;
+
+ for (;;) {
+ auto next_node = find_next_node(node, *p);
+ if (next_node == nullptr) {
+ return nullptr;
+ }
+
+ node = next_node;
+
+ auto n = std::min(node->len, static_cast<size_t>(last - p));
+ if (memcmp(node->s, p, n) != 0) {
+ return nullptr;
+ }
+
+ p += n;
+
+ if (p != last) {
+ if (node->index != -1) {
+ *nread = p - first;
+ return node;
+ }
+ continue;
+ }
+
+ if (node->len == n) {
+ *nread = p - first;
+ return node;
+ }
+
+ return nullptr;
+ }
+}
+} // namespace
+
+ssize_t Router::match_prefix(size_t *nread, const RNode **last_node,
+ const StringRef &s) const {
+ if (*last_node == nullptr) {
+ *last_node = &root_;
+ }
+
+ auto node =
+ ::shrpx::match_prefix(nread, *last_node, std::begin(s), std::end(s));
+ if (node == nullptr) {
+ return -1;
+ }
+
+ *last_node = node;
+
+ return node->index;
+}
+
+namespace {
+void dump_node(const RNode *node, int depth) {
+ fprintf(stderr, "%*ss='%.*s', len=%zu, index=%zd\n", depth, "",
+ (int)node->len, node->s, node->len, node->index);
+ for (auto &nd : node->next) {
+ dump_node(nd.get(), depth + 4);
+ }
+}
+} // namespace
+
+void Router::dump() const { dump_node(&root_, 0); }
+
+} // namespace shrpx