[ckip ci] Add quoting functionality to new parsing method

This commit is contained in:
ado 2023-08-13 11:36:56 +02:00
parent 0b3b719155
commit 4e4c3a6e02
2 changed files with 256 additions and 45 deletions

View File

@ -1,7 +1,4 @@
#pragma once
#include <cstdint>
#include <cstdio>
#include <cstdlib>
#include <vector>
namespace ss {

View File

@ -355,8 +355,8 @@ public:
template <typename U, typename... Us>
no_void_validator_tup_t<U, Us...> try_same() {
parser_.clear_error();
auto value =
parser_.reader_.converter_.template convert<U, Us...>();
auto value = parser_.reader_.converter_.template convert<U, Us...>(
parser_.reader_.split_data_);
if (!parser_.reader_.converter_.valid()) {
parser_.handle_error_invalid_conversion();
}
@ -692,12 +692,13 @@ private:
// TODO handle differently
if (buff_filled_ == 0 || (buff_filled_ == 1 && buff_[0] == '\n') ||
(buff_filled_ == 2 && buff_[0] == '\r' && buff_[1] == '\n')) {
free(file_);
fclose(file_);
file_ = nullptr;
}
begin_ = buff_;
curr_ = buff_;
shifted_curr_ = buff_;
end_ = buff_ + buff_filled_;
}
@ -735,6 +736,251 @@ private:
}
};
// TODO move
using quote = typename setup<Options...>::quote;
using trim_left = typename setup<Options...>::trim_left;
using trim_right = typename setup<Options...>::trim_right;
using escape = typename setup<Options...>::escape;
using multiline = typename setup<Options...>::multiline;
constexpr static auto is_const_line =
!quote::enabled && !escape::enabled;
using line_ptr_type =
std::conditional_t<is_const_line, const char*, char*>;
bool match(const char* const curr, char delim) {
return *curr == delim;
};
// todo test for huge delimiters
bool match(const char* const curr, const std::string& delim) {
return strncmp(curr, delim.c_str(), delim.size()) == 0;
};
size_t delimiter_size(char) {
return 1;
}
size_t delimiter_size(const std::string& delim) {
return delim.size();
}
void trim_left_if_enabled(line_ptr_type& curr) {
if constexpr (trim_left::enabled) {
while (trim_left::match(*curr)) {
++curr;
}
}
}
void trim_right_if_enabled(line_ptr_type& curr) {
if constexpr (trim_right::enabled) {
while (trim_right::match(*curr)) {
++curr;
}
}
}
// matches delimiter taking spacing into account
// returns {spacing + delimiter size, is delimiter}
template <typename Delim>
std::tuple<size_t, bool> match_delimiter(line_ptr_type begin,
const Delim& delim) {
line_ptr_type curr = begin;
trim_right_if_enabled(curr);
// just spacing
// TODO handle \r\n
if (*curr == '\n' || (*curr == '\r' && *(curr + 1) == '\n')) {
return {0, false};
}
// not a delimiter
if (!match(curr, delim)) {
shift_if_escaped(curr);
return {1 + curr - begin, false};
}
curr += delimiter_size(delim);
trim_left_if_enabled(curr);
// delimiter
return {curr - begin, true};
}
void shift_if_escaped(line_ptr_type& curr) {
if constexpr (escape::enabled) {
if (escape::match(*curr)) {
if (curr[1] == '\0') {
if constexpr (!multiline::enabled) {
// TODO handle
throw "unterminated escape";
// handle_error_unterminated_escape();
}
return;
}
shift_and_jump_escape();
}
}
}
void shift_and_set_current() {
if constexpr (!is_const_line) {
if (escaped_ > 0) {
// shift by number of escapes
std::copy_n(shifted_curr_ + escaped_,
curr_ - shifted_curr_ - escaped_,
shifted_curr_);
shifted_curr_ = curr_ - escaped_;
return;
}
}
shifted_curr_ = curr_;
}
void shift_and_jump_escape() {
shift_and_set_current();
if constexpr (!is_const_line) {
++escaped_;
}
++curr_;
}
void shift_push_and_start_next(size_t n) {
shift_and_push();
begin_ = curr_ + n;
}
void shift_and_push() {
shift_and_set_current();
split_data_.emplace_back(begin_, shifted_curr_);
}
void parse_next_line() {
escaped_ = 0;
auto check_buff_end = [&] {
if (curr_ == end_) {
auto old_buff = buff_;
if (last_read_) {
// TODO handle
throw "no new line at eof";
}
handle_buffer_end_reached();
end_ = buff_ + buff_filled_;
for (auto& [begin, end] : split_data_) {
begin = begin - old_buff + buff_;
end = end - old_buff + buff_;
}
begin_ = begin_ - old_buff + buff_;
curr_ = curr_ - old_buff + buff_;
}
};
while (true) {
// quoted string
if constexpr (quote::enabled) {
if (quote::match(*curr_)) {
begin_ = shifted_curr_ = ++curr_;
check_buff_end();
while (true) {
// quote not closed
if (!quote::match(*curr_)) {
// end of line
if constexpr (!multiline::enabled) {
// TODO test \r\n
if (*curr_ == '\n' || *curr_ == '\r') {
throw "unterminated quote";
}
}
++curr_;
check_buff_end();
continue;
}
auto [width, valid] =
match_delimiter(curr_ + 1, delim_char_);
// delimiter
if (valid) {
shift_push_and_start_next(width + 1);
curr_ += width + 1;
check_buff_end();
break;
}
// double quote
// eg: ...,"hel""lo",... -> hel"lo
if (quote::match(*(curr_ + 1))) {
shift_and_jump_escape();
++curr_;
check_buff_end();
continue;
}
if (width == 0) {
// eol
// eg: ...,"hello" \n -> hello
// eg no trim: ...,"hello"\n -> hello
shift_and_push();
++curr_;
// TODO handle differently
if (curr_[0] == '\r') {
++curr_;
}
check_buff_end();
return;
}
// mismatched quote
// eg: ...,"hel"lo,... -> error
// handle_error_mismatched_quote(curr_ - buff_);
throw "missmatched quote";
// split_data_.emplace_back(buff_, begin_);
}
}
}
// not quoted
while (true) {
if (*curr_ == '\n') {
split_data_.emplace_back(begin_, curr_);
return;
}
if (*curr_ == '\r' && *(curr_ + 1) == '\n') {
split_data_.emplace_back(begin_, curr_);
++curr_;
check_buff_end();
return;
}
if (*curr_ == delim_char_) {
split_data_.emplace_back(begin_, curr_);
begin_ = curr_ + 1;
++curr_;
check_buff_end();
break;
}
++curr_;
check_buff_end();
}
}
}
// read next line each time in order to set eof_
bool read_next() {
// TODO update division value
@ -754,45 +1000,7 @@ private:
split_data_.clear();
begin_ = curr_;
while (true) {
if (*curr_ == '\n') {
split_data_.emplace_back(begin_, curr_);
break;
}
if (*curr_ == '\r' && *(curr_ + 1) == '\n') {
split_data_.emplace_back(begin_, curr_);
++curr_;
break;
}
if (*curr_ == delim_char_) {
split_data_.emplace_back(begin_, curr_);
begin_ = curr_ + 1;
}
++curr_;
if (curr_ == end_) {
auto old_buff = buff_;
if (last_read_) {
// TODO handle
throw std::string{"no new line at eof"};
}
handle_buffer_end_reached();
end_ = buff_ + buff_filled_;
for (auto& [begin, end] : split_data_) {
begin = begin - old_buff + buff_;
end = end - old_buff + buff_;
}
begin_ = begin_ - old_buff + buff_;
curr_ = curr_ - old_buff + buff_;
}
}
parse_next_line();
++curr_;
buff_processed_ = curr_ - buff_;
@ -824,12 +1032,18 @@ private:
// TODO move to splitter
char* begin_{nullptr};
char* shifted_curr_{nullptr};
char* curr_{nullptr};
char* end_{nullptr};
bool last_read_{false};
split_data split_data_;
// TODO add to constructors
size_t escaped_{0};
bool unterminated_quote_{true};
bool unterminated_escape_{true};
};
////////////////