common/url_escape: add simple url_[un]escape() methods

Signed-off-by: Sage Weil <sage@redhat.com>

diff --git a/src/CMakeLists.txt b/src/CMakeLists.txt
index 39ce986a644e8e72b9a50034bb0e2d995c282426..910420914ae98d4bbb5db9b94c9dca46efdcb90b 100644 (file)
--- a/src/CMakeLists.txt
+++ b/src/CMakeLists.txt
@@ -399,6 +399,7 @@ set(libcommon_files
   ${crush_srcs}
   common/cmdparse.cc
   common/escape.c
+  common/url_escape.cc
   common/io_priority.cc
   common/Clock.cc
   common/ceph_time.cc

diff --git a/src/common/url_escape.cc b/src/common/url_escape.cc
new file mode 100644 (file)
index 0000000..6580d28
--- /dev/null
+++ b/src/common/url_escape.cc
@@ -0,0 +1,64 @@
+// -*- mode:C++; tab-width:8; c-basic-offset:2; indent-tabs-mode:t -*-
+// vim: ts=8 sw=2 smarttab
+
+#include "url_escape.h"
+
+#include <stdexcept>
+#include <sstream>
+
+std::string url_escape(const std::string& s)
+{
+  std::string out;
+  for (auto c : s) {
+    if (std::isalnum(c) || c == '-' || c == '.' || c == '_' || c == '~' ||
+       c == '/') {
+      out.push_back(c);
+    } else {
+      char t[4];
+      snprintf(t, sizeof(t), "%%%02x", (int)(unsigned char)c);
+      out.append(t);
+    }
+  }
+  return out;
+}
+
+std::string url_unescape(const std::string& s)
+{
+  std::string out;
+  const char *end = s.c_str() + s.size();
+  for (const char *c = s.c_str(); c < end; ++c) {
+    switch (*c) {
+    case '%':
+      {
+       unsigned char v = 0;
+       for (unsigned i=0; i<2; ++i) {
+         ++c;
+         if (c >= end) {
+           std::ostringstream ss;
+           ss << "invalid escaped string at pos " << (c - s.c_str()) << " of '"
+              << s << "'";
+           throw std::runtime_error(ss.str());
+         }
+         v <<= 4;
+         if (*c >= '0' && *c <= '9') {
+           v += *c - '0';
+         } else if (*c >= 'a' && *c <= 'f') {
+           v += *c - 'a' + 10;
+         } else if (*c >= 'A' && *c <= 'F') {
+           v += *c - 'A' + 10;
+         } else {
+           std::ostringstream ss;
+           ss << "invalid escaped string at pos " << (c - s.c_str()) << " of '"
+              << s << "'";
+           throw std::runtime_error(ss.str());
+         }
+       }
+       out.push_back(v);
+      }
+      break;
+    default:
+      out.push_back(*c);
+    }
+  }
+  return out;
+}

diff --git a/src/common/url_escape.h b/src/common/url_escape.h
new file mode 100644 (file)
index 0000000..3cb539b
--- /dev/null
+++ b/src/common/url_escape.h
@@ -0,0 +1,9 @@
+// -*- mode:C++; tab-width:8; c-basic-offset:2; indent-tabs-mode:t -*-
+// vim: ts=8 sw=2 smarttab
+
+#pragma once
+
+#include <string>
+
+extern std::string url_escape(const std::string& s);
+extern std::string url_unescape(const std::string& s);

diff --git a/src/test/common/CMakeLists.txt b/src/test/common/CMakeLists.txt
index 693a2bee66af8ed1fd5cc2540bd1a8b0a2013085..3a34fd05c1bde4c4839c44fd0e5452d9184990ee 100644 (file)
--- a/src/test/common/CMakeLists.txt
+++ b/src/test/common/CMakeLists.txt
@@ -139,6 +139,13 @@ add_executable(unittest_safe_io
 add_ceph_unittest(unittest_safe_io ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/unittest_safe_io)
 target_link_libraries(unittest_safe_io global)
 
+# unittest_url_escape
+add_executable(unittest_url_escape
+  test_url_escape.cc
+  )
+add_ceph_unittest(unittest_url_escape ${CMAKE_RUNTIME_OUTPUT_DIRECTORY}/unittest_url_escape)
+target_link_libraries(unittest_url_escape ceph-common)
+
 # unittest_readahead
 add_executable(unittest_readahead
   Readahead.cc

diff --git a/src/test/common/test_url_escape.cc b/src/test/common/test_url_escape.cc
new file mode 100644 (file)
index 0000000..6c27b64
--- /dev/null
+++ b/src/test/common/test_url_escape.cc
@@ -0,0 +1,36 @@
+// -*- mode:C++; tab-width:8; c-basic-offset:2; indent-tabs-mode:t -*-
+// vim: ts=8 sw=2 smarttab
+
+#include "common/url_escape.h"
+
+#include "gtest/gtest.h"
+
+TEST(url_escape, escape) {
+  ASSERT_EQ(url_escape("foo bar"), std::string("foo%20bar"));
+  ASSERT_EQ(url_escape("foo\nbar"), std::string("foo%0abar"));
+}
+
+TEST(url_escape, unescape) {
+  ASSERT_EQ(url_unescape("foo%20bar"), std::string("foo bar"));
+  ASSERT_EQ(url_unescape("foo%0abar"), std::string("foo\nbar"));
+  ASSERT_EQ(url_unescape("%20"), std::string(" "));
+  ASSERT_EQ(url_unescape("\0%20"), std::string("\0 "));
+  ASSERT_EQ(url_unescape("\x01%20"), std::string("\x01 "));
+}
+
+TEST(url_escape, all_chars) {
+  std::string a;
+  for (unsigned j=0; j<256; ++j) {
+    a.push_back((char)j);
+  }
+  std::string b = url_escape(a);
+  std::cout << "escaped: " << b << std::endl;
+  ASSERT_EQ(a, url_unescape(b));
+}
+
+TEST(url_escape, invalid) {
+  ASSERT_THROW(url_unescape("foo%xx"), std::runtime_error);
+  ASSERT_THROW(url_unescape("foo%%"), std::runtime_error);
+  ASSERT_THROW(url_unescape("foo%"), std::runtime_error);
+  ASSERT_THROW(url_unescape("foo%0"), std::runtime_error);
+}