libreoffice-online/common/Protocol.hpp
Ashod Nakashian e85854ed39 wsd: track modifying user commands
There is a race between the time of modifying
a document, receiving the modified flag, and
saving. This can happen when, for example,
the user modifies the document and closes
immediately. In that case, when uploading
we will not have the modified flag and will
not set the User-Modified attribute.

While this isn't 100% accurate, and it can
never be, it's still better to be conservative
and flag a version in storage as user-modified
than otherwise.

Change-Id: Ia504a7cddd4839bcbfeaaf9bf6c90ed8b68efa91
Signed-off-by: Ashod Nakashian <ashod.nakashian@collabora.co.uk>
2022-08-25 15:12:10 +01:00

314 lines
11 KiB
C++

/* -*- Mode: C++; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4; fill-column: 100 -*- */
/*
* This Source Code Form is subject to the terms of the Mozilla Public
* License, v. 2.0. If a copy of the MPL was not distributed with this
* file, You can obtain one at http://mozilla.org/MPL/2.0/.
*/
#pragma once
#include <cstdint>
#include <cstdlib>
#include <cstring>
#include <iomanip>
#include <map>
#include <regex>
#include <sstream>
#include <string>
#include <vector>
#include <StringVector.hpp>
#include <Util.hpp>
#define LOK_USE_UNSTABLE_API
#include <LibreOfficeKit/LibreOfficeKitEnums.h>
namespace COOLProtocol
{
// Protocol Version Number.
// See protocol.txt.
constexpr unsigned ProtocolMajorVersionNumber = 0;
constexpr unsigned ProtocolMinorVersionNumber = 1;
inline std::string GetProtocolVersion()
{
return std::to_string(ProtocolMajorVersionNumber) + '.'
+ std::to_string(ProtocolMinorVersionNumber);
}
// Parse a string into a version tuple.
// Negative numbers for error.
std::tuple<int, int, std::string> ParseVersion(const std::string& version);
inline bool stringToInteger(const std::string& input, int& value)
{
bool res;
std::tie(value, res) = Util::i32FromString(input);
return res;
}
inline bool stringToUInt32(const std::string& input, uint32_t& value)
{
bool res;
std::tie(value, res) = Util::i32FromString(input);
return res;
}
inline bool stringToUInt64(const std::string& input, uint64_t& value)
{
bool res;
std::tie(value, res) = Util::u64FromString(input);
return res;
}
inline
bool parseNameValuePair(const std::string& token, std::string& name, std::string& value, const char delim = '=')
{
const size_t mid = token.find_first_of(delim);
if (mid != std::string::npos)
{
name = token.substr(0, mid);
value = token.substr(mid + 1);
return true;
}
return false;
}
bool getTokenInteger(const std::string& token, const std::string& name, int& value);
bool getTokenUInt32(const std::string& token, const std::string& name, uint32_t& value);
bool getTokenUInt64(const std::string& token, const std::string& name, uint64_t& value);
bool getTokenString(const std::string& token, const std::string& name, std::string& value);
bool getTokenKeyword(const std::string& token, const std::string& name, const std::map<std::string, int>& map, int& value);
bool getTokenKeyword(const StringVector& tokens, const std::string& name, const std::map<std::string, int>& map, int& value);
bool getTokenInteger(const StringVector& tokens, const std::string& name, int& value);
/// Literal-string token names.
template <std::size_t N>
inline bool getTokenInteger(const std::string& token, const char (&name)[N], int& value)
{
// N includes null termination.
static_assert(N > 1, "Token name must be at least one character long.");
if (token.size() > N && token[N - 1] == '=' && token.compare(0, N - 1, name) == 0)
{
const char* str = token.data() + N;
char* endptr = nullptr;
value = std::strtol(str, &endptr, 10);
return (endptr > str);
}
return false;
}
/// Extracts a name and value from token. Returns true if value is a non-negative integer.
template <std::size_t N>
inline bool getNonNegTokenInteger(const std::string& token, const char (&name)[N], int& value)
{
return getTokenInteger(token, name, value) && value >= 0;
}
inline bool getTokenString(const StringVector& tokens,
const std::string& name,
std::string& value)
{
for (const auto& token : tokens)
{
if (getTokenString(tokens.getParam(token), name, value))
{
return true;
}
}
return false;
}
bool getTokenStringFromMessage(const std::string& message, const std::string& name, std::string& value);
bool getTokenKeywordFromMessage(const std::string& message, const std::string& name, const std::map<std::string, int>& map, int& value);
inline
std::vector<int> tokenizeInts(const char* data, const size_t size, const char delimiter = ',')
{
std::vector<int> tokens;
if (size == 0 || data == nullptr)
return tokens;
const char* start = data;
const char* end = data;
for (size_t i = 0; i < size && data[i] != '\n'; ++i, ++end)
{
if (data[i] == delimiter)
{
if (start != end && *start != delimiter)
tokens.emplace_back(std::atoi(start));
start = end;
}
else if (*start == delimiter)
++start;
}
if (start != end && *start != delimiter && *start != '\n')
tokens.emplace_back(std::atoi(start));
return tokens;
}
inline
std::vector<int> tokenizeInts(const std::string& s, const char delimiter = ',')
{
return tokenizeInts(s.data(), s.size(), delimiter);
}
inline bool getTokenIntegerFromMessage(const std::string& message, const std::string& name, int& value)
{
return getTokenInteger(StringVector::tokenize(message), name, value);
}
/// Returns the first token of a message.
inline
std::string getFirstToken(const char *message, const int length, const char delim = ' ')
{
return Util::getDelimitedInitialSubstring(message, length, delim);
}
template <typename T>
std::string getFirstToken(const T& message, const char delim = ' ')
{
return getFirstToken(message.data(), message.size(), delim);
}
inline
bool matchPrefix(const std::string& prefix, const std::string& message)
{
return (message.size() >= prefix.size() &&
message.compare(0, prefix.size(), prefix) == 0);
}
inline
bool matchPrefix(const std::string& prefix, const std::vector<char>& message)
{
return (message.size() >= prefix.size() &&
prefix.compare(0, prefix.size(), message.data(), prefix.size()) == 0);
}
inline
bool matchPrefix(const std::string& prefix, const std::string& message, const bool ignoreWhitespace)
{
if (ignoreWhitespace)
{
const size_t posPre = prefix.find_first_not_of(' ');
const size_t posMsg = message.find_first_not_of(' ');
return matchPrefix(posPre == std::string::npos ? prefix : prefix.substr(posPre),
posMsg == std::string::npos ? message : message.substr(posMsg));
}
else
{
return matchPrefix(prefix, message);
}
}
/// Returns true if the token is a user-interaction token.
/// Currently this excludes commands sent automatically.
/// Notice that this doesn't guarantee editing activity,
/// rather just user interaction with the UI.
inline
bool tokenIndicatesUserInteraction(const std::string& token)
{
// Exclude tokens that include these keywords, such as canceltiles statusindicator.
// FIXME: This is wrong. That the token happens to contain (or not) a certain substring is
// no guarantee that it "indicates user interaction". It might be like that at the moment,
// but that is coincidental. We should check what the actual whole token is, at least, not
// look for a substring.
return (token.find("tile") == std::string::npos &&
token.find("status") == std::string::npos &&
token.find("state") == std::string::npos &&
token != "userinactive");
}
/// Returns true if the token is a likely document modifying command.
/// This is never 100% accurate, but it is needed to filter out tokens
/// that certainly do not modify the document, such as 'load' and 'save'
/// commands. Some commands are certainly modifying, e.g. 'key', others
/// can only potentially be modifying, e.g. 'mouse' while dragging.
/// Note: this is only used when we don't have the modified flag from
/// Core so we flag the document as user-modified more accurately.
inline bool tokenIndicatesDocumentModification(const std::string& token)
{
// These keywords are chosen to cover the largest set of
// commands that may potentially modify the document.
// We need to assume modification rather than not.
return (
token.find("mouse") != std::string::npos || token.find("key") != std::string::npos ||
token.find("command") != std::string::npos ||
token.find("select") != std::string::npos || token.find("set") != std::string::npos ||
token.find("uno") != std::string::npos || token.find("input") != std::string::npos ||
token.find("move") != std::string::npos || token.find("paste") != std::string::npos ||
token.find("insert") != std::string::npos ||
token.find("resize") != std::string::npos ||
token.find("remove") != std::string::npos || token.find("sign") != std::string::npos);
}
/// Returns the first line of a message.
inline
std::string getFirstLine(const char *message, const int length)
{
return Util::getDelimitedInitialSubstring(message, length, '\n');
}
/// Returns the first line of any data which payload char*.
template <typename T>
std::string getFirstLine(const T& message)
{
return getFirstLine(message.data(), message.size());
}
/// Returns an abbreviation of the message (the first line, indicating truncation). We assume
/// that it adhers to the COOL protocol, i.e. that there is always a first (or only) line that
/// is in printable UTF-8. I.e. no encoding of binary bytes is done. The format of the result is
/// not guaranteed to be stable. It is to be used for logging purposes only, not for decoding
/// protocol frames.
inline
std::string getAbbreviatedMessage(const char *message, const int length)
{
if (message == nullptr || length <= 0)
{
return std::string();
}
const std::string firstLine = getFirstLine(message, std::min(length, 500));
// If first line is less than the length (minus newline), add ellipsis.
if (firstLine.size() < static_cast<std::string::size_type>(length) - 1)
{
return firstLine + "...";
}
return firstLine;
}
inline std::string getAbbreviatedMessage(const std::string& message)
{
const size_t pos = Util::getDelimiterPosition(message.data(), std::min<size_t>(message.size(), 501), '\n');
// If first line is less than the length (minus newline), add ellipsis.
if (pos < static_cast<std::string::size_type>(message.size()) - 1)
{
return message.substr(0, pos) + "...";
}
return message;
}
template <typename T>
std::string getAbbreviatedMessage(const T& message)
{
return getAbbreviatedMessage(message.data(), message.size());
}
};
/* vim:set shiftwidth=4 softtabstop=4 expandtab: */