url_canon_etc.cc | Explore in Territory

// Copyright 2013 The Chromium Authors
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.

#ifdef UNSAFE_BUFFERS_BUILD
// TODO(crbug.com/350788890): Remove this and spanify to fix the errors.
#pragma allow_unsafe_buffers
#endif

// Canonicalizers for random bits that aren't big enough for their own files.

#include <string.h>

#include "url/url_canon.h"
#include "url/url_canon_internal.h"

namespace url {

namespace {

// Returns true if the given character should be removed from the middle of a
// URL.
inline bool IsRemovableURLWhitespace(int ch) { … }

// Backend for RemoveURLWhitespace (see declaration in url_canon.h).
// It sucks that we have to do this, since this takes about 13% of the total URL
// canonicalization time.
template <typename CHAR>
const CHAR* DoRemoveURLWhitespace(const CHAR* input,
                                  int input_len,
                                  CanonOutputT<CHAR>* buffer,
                                  int* output_len,
                                  bool* potentially_dangling_markup) { … }

// Contains the canonical version of each possible input letter in the scheme
// (basically, lower-cased). The corresponding entry will be 0 if the letter
// is not allowed in a scheme.
// clang-format off
const char kSchemeCanonical[0x80] = …;
// clang-format on

// This could be a table lookup as well by setting the high bit for each
// valid character, but it's only called once per URL, and it makes the lookup
// table easier to read not having extra stuff in it.
inline bool IsSchemeFirstChar(unsigned char c) { … }

template <typename CHAR, typename UCHAR>
bool DoScheme(const CHAR* spec,
              const Component& scheme,
              CanonOutput* output,
              Component* out_scheme) { … }

// The username and password components reference ranges in the corresponding
// *_spec strings. Typically, these specs will be the same (we're
// canonicalizing a single source string), but may be different when
// replacing components.
template <typename CHAR, typename UCHAR>
bool DoUserInfo(const CHAR* username_spec,
                const Component& username,
                const CHAR* password_spec,
                const Component& password,
                CanonOutput* output,
                Component* out_username,
                Component* out_password) { … }

// Helper functions for converting port integers to strings.
inline void WritePortInt(char* output, int output_len, int port) { … }

// This function will prepend the colon if there will be a port.
template <typename CHAR, typename UCHAR>
bool DoPort(const CHAR* spec,
            const Component& port,
            int default_port_for_scheme,
            CanonOutput* output,
            Component* out_port) { … }

// clang-format off
//   Percent-escape all characters from the fragment percent-encode set
//   https://url.spec.whatwg.org/#fragment-percent-encode-set
const bool kShouldEscapeCharInFragment[0x80] = …;
// clang-format on

template <typename CHAR, typename UCHAR>
void DoCanonicalizeRef(const CHAR* spec,
                       const Component& ref,
                       CanonOutput* output,
                       Component* out_ref) { … }

}  // namespace

const char* RemoveURLWhitespace(const char* input,
                                int input_len,
                                CanonOutputT<char>* buffer,
                                int* output_len,
                                bool* potentially_dangling_markup) { … }

const char16_t* RemoveURLWhitespace(const char16_t* input,
                                    int input_len,
                                    CanonOutputT<char16_t>* buffer,
                                    int* output_len,
                                    bool* potentially_dangling_markup) { … }

char CanonicalSchemeChar(char16_t ch) { … }

bool CanonicalizeScheme(const char* spec,
                        const Component& scheme,
                        CanonOutput* output,
                        Component* out_scheme) { … }

bool CanonicalizeScheme(const char16_t* spec,
                        const Component& scheme,
                        CanonOutput* output,
                        Component* out_scheme) { … }

bool CanonicalizeUserInfo(const char* username_source,
                          const Component& username,
                          const char* password_source,
                          const Component& password,
                          CanonOutput* output,
                          Component* out_username,
                          Component* out_password) { … }

bool CanonicalizeUserInfo(const char16_t* username_source,
                          const Component& username,
                          const char16_t* password_source,
                          const Component& password,
                          CanonOutput* output,
                          Component* out_username,
                          Component* out_password) { … }

bool CanonicalizePort(const char* spec,
                      const Component& port,
                      int default_port_for_scheme,
                      CanonOutput* output,
                      Component* out_port) { … }

bool CanonicalizePort(const char16_t* spec,
                      const Component& port,
                      int default_port_for_scheme,
                      CanonOutput* output,
                      Component* out_port) { … }

void CanonicalizeRef(const char* spec,
                     const Component& ref,
                     CanonOutput* output,
                     Component* out_ref) { … }

void CanonicalizeRef(const char16_t* spec,
                     const Component& ref,
                     CanonOutput* output,
                     Component* out_ref) { … }

}  // namespace url
chromium/url/url_canon_etc.cc
