kakoune/src/selectors.cc

642 lines
19 KiB
C++
Raw Normal View History

#include "selectors.hh"
#include "string.hh"
2012-10-08 19:12:09 +02:00
#include "utf8_iterator.hh"
#include <algorithm>
#include <boost/optional.hpp>
namespace Kakoune
{
using Utf8Iterator = utf8::utf8_iterator<BufferIterator, utf8::InvalidBytePolicy::Pass>;
2012-10-08 19:12:09 +02:00
2012-03-07 20:20:32 +01:00
namespace
{
template<bool punctuation_is_word = false>
2012-10-08 19:12:09 +02:00
bool is_word(Codepoint c)
{
2012-09-30 16:23:18 +02:00
return Kakoune::is_word(c);
}
2012-03-07 20:20:32 +01:00
template<>
2012-10-08 19:12:09 +02:00
bool is_word<true>(Codepoint c)
2012-03-07 20:20:32 +01:00
{
return !is_blank(c) and !is_eol(c);
}
2012-10-08 19:12:09 +02:00
static bool is_punctuation(Codepoint c)
{
return not (is_word(c) or is_blank(c) or is_eol(c));
}
enum class CharCategories
{
Blank,
EndOfLine,
Word,
Punctuation,
};
2012-03-07 20:20:32 +01:00
template<bool punctuation_is_word = false>
2012-10-08 19:12:09 +02:00
CharCategories categorize(Codepoint c)
{
if (is_word(c))
return CharCategories::Word;
if (is_eol(c))
return CharCategories::EndOfLine;
if (is_blank(c))
return CharCategories::Blank;
2012-03-07 20:20:32 +01:00
return punctuation_is_word ? CharCategories::Word
: CharCategories::Punctuation;
}
2012-10-08 19:12:09 +02:00
bool is_begin(const BufferIterator& it) { return it.is_begin(); }
bool is_end(const BufferIterator& it) { return it.is_end(); }
bool is_begin(const Utf8Iterator& it) { return it.underlying_iterator().is_begin(); }
bool is_end(const Utf8Iterator& it) { return it.underlying_iterator().is_end(); }
template<typename Iterator, typename T>
2013-01-03 14:30:14 +01:00
void skip_while(Iterator& it, T condition)
{
2012-10-08 19:12:09 +02:00
while (not is_end(it) and condition(*it))
++it;
}
2012-10-08 19:12:09 +02:00
template<typename Iterator, typename T>
2013-01-03 14:30:14 +01:00
void skip_while_reverse(Iterator& it, T condition)
{
2012-10-08 19:12:09 +02:00
while (not is_begin(it) and condition(*it))
--it;
2012-10-08 19:12:09 +02:00
}
Range utf8_range(const Utf8Iterator& first, const Utf8Iterator& last)
2012-10-08 19:12:09 +02:00
{
return {first.underlying_iterator(), last.underlying_iterator()};
}
2012-03-07 20:20:32 +01:00
}
typedef boost::regex_iterator<BufferIterator> RegexIterator;
2012-03-07 20:20:32 +01:00
template<bool punctuation_is_word>
2013-06-01 14:22:57 +02:00
Selection select_to_next_word(const Buffer& buffer, const Selection& selection)
{
2012-10-08 19:12:09 +02:00
Utf8Iterator begin = selection.last();
2012-03-07 20:20:32 +01:00
if (categorize<punctuation_is_word>(*begin) !=
categorize<punctuation_is_word>(*(begin+1)))
++begin;
skip_while(begin, is_eol);
2012-10-08 19:12:09 +02:00
Utf8Iterator end = begin+1;
2012-03-07 20:20:32 +01:00
if (not punctuation_is_word and is_punctuation(*begin))
skip_while(end, is_punctuation);
2012-03-07 20:20:32 +01:00
else if (is_word<punctuation_is_word>(*begin))
skip_while(end, is_word<punctuation_is_word>);
2013-01-03 14:30:14 +01:00
skip_while(end, is_blank);
return utf8_range(begin, end-1);
}
2013-06-01 14:22:57 +02:00
template Selection select_to_next_word<false>(const Buffer&, const Selection&);
template Selection select_to_next_word<true>(const Buffer&, const Selection&);
2012-03-07 20:20:32 +01:00
template<bool punctuation_is_word>
2013-06-01 14:22:57 +02:00
Selection select_to_next_word_end(const Buffer& buffer, const Selection& selection)
{
2012-10-08 19:12:09 +02:00
Utf8Iterator begin = selection.last();
2012-03-07 20:20:32 +01:00
if (categorize<punctuation_is_word>(*begin) !=
categorize<punctuation_is_word>(*(begin+1)))
++begin;
skip_while(begin, is_eol);
2012-10-08 19:12:09 +02:00
Utf8Iterator end = begin;
skip_while(end, is_blank);
2012-03-07 20:20:32 +01:00
if (not punctuation_is_word and is_punctuation(*end))
2013-01-03 14:30:14 +01:00
skip_while(end, is_punctuation);
2012-03-07 20:20:32 +01:00
else if (is_word<punctuation_is_word>(*end))
2013-01-03 14:30:14 +01:00
skip_while(end, is_word<punctuation_is_word>);
return utf8_range(begin, end-1);
}
2013-06-01 14:22:57 +02:00
template Selection select_to_next_word_end<false>(const Buffer&, const Selection&);
template Selection select_to_next_word_end<true>(const Buffer&, const Selection&);
2012-03-07 20:20:32 +01:00
template<bool punctuation_is_word>
2013-06-01 14:22:57 +02:00
Selection select_to_previous_word(const Buffer& buffer, const Selection& selection)
{
2012-10-08 19:12:09 +02:00
Utf8Iterator begin = selection.last();
2012-03-07 20:20:32 +01:00
if (categorize<punctuation_is_word>(*begin) !=
categorize<punctuation_is_word>(*(begin-1)))
--begin;
skip_while_reverse(begin, is_eol);
2012-10-08 19:12:09 +02:00
Utf8Iterator end = begin;
skip_while_reverse(end, is_blank);
bool with_end = false;
2012-03-07 20:20:32 +01:00
if (not punctuation_is_word and is_punctuation(*end))
2013-01-03 14:30:14 +01:00
{
skip_while_reverse(end, is_punctuation);
with_end = is_punctuation(*end);
}
2012-03-07 20:20:32 +01:00
else if (is_word<punctuation_is_word>(*end))
2013-01-03 14:30:14 +01:00
{
skip_while_reverse(end, is_word<punctuation_is_word>);
with_end = is_word<punctuation_is_word>(*end);
}
return utf8_range(begin, with_end ? end : end+1);
}
2013-06-01 14:22:57 +02:00
template Selection select_to_previous_word<false>(const Buffer&, const Selection&);
template Selection select_to_previous_word<true>(const Buffer&, const Selection&);
2013-06-01 14:22:57 +02:00
Selection select_line(const Buffer& buffer, const Selection& selection)
{
2012-10-08 19:12:09 +02:00
Utf8Iterator first = selection.last();
if (*first == '\n' and not is_end(first + 1))
++first;
2012-10-08 19:12:09 +02:00
while (not is_begin(first) and *(first - 1) != '\n')
--first;
2012-10-08 19:12:09 +02:00
Utf8Iterator last = first;
while (not is_end(last + 1) and *last != '\n')
++last;
return utf8_range(first, last);
}
2013-06-01 14:22:57 +02:00
Selection select_matching(const Buffer& buffer, const Selection& selection)
{
2012-10-11 01:17:29 +02:00
std::vector<Codepoint> matching_pairs = { '(', ')', '{', '}', '[', ']', '<', '>' };
2012-10-08 19:12:09 +02:00
Utf8Iterator it = selection.last();
2012-10-11 01:17:29 +02:00
std::vector<Codepoint>::iterator match = matching_pairs.end();
while (not is_eol(*it))
{
match = std::find(matching_pairs.begin(), matching_pairs.end(), *it);
if (match != matching_pairs.end())
break;
++it;
}
if (match == matching_pairs.end())
return selection;
2012-10-08 19:12:09 +02:00
Utf8Iterator begin = it;
if (((match - matching_pairs.begin()) % 2) == 0)
{
int level = 0;
2012-10-11 01:17:29 +02:00
const Codepoint opening = *match;
const Codepoint closing = *(match+1);
2012-10-08 19:12:09 +02:00
while (not is_end(it))
{
if (*it == opening)
++level;
else if (*it == closing and --level == 0)
return utf8_range(begin, it);
++it;
}
}
else
{
int level = 0;
2012-10-11 01:17:29 +02:00
const Codepoint opening = *(match-1);
const Codepoint closing = *match;
while (true)
{
if (*it == closing)
++level;
else if (*it == opening and --level == 0)
return utf8_range(begin, it);
if (is_begin(it))
break;
--it;
}
}
return selection;
}
// c++14 will add std::optional, so we use boost::optional until then
using boost::optional;
static optional<Range> find_surrounding(const BufferIterator& pos,
const CodepointPair& matching,
ObjectFlags flags)
{
const bool to_begin = flags & ObjectFlags::ToBegin;
const bool to_end = flags & ObjectFlags::ToEnd;
const bool nestable = matching.first != matching.second;
Utf8Iterator first = pos;
if (to_begin)
{
int level = 0;
while (not is_begin(first))
{
if (nestable and first != pos and *first == matching.second)
++level;
else if (*first == matching.first)
{
if (level == 0)
break;
else
--level;
}
--first;
}
if (level != 0 or *first != matching.first)
return optional<Range>{};
}
Utf8Iterator last = pos;
if (to_end)
{
int level = 0;
last = first + 1;
while (not is_end(last))
{
if (nestable and *last == matching.first)
++level;
else if (*last == matching.second)
{
if (level == 0)
break;
else
--level;
}
++last;
}
if (level != 0 or *last != matching.second)
return optional<Range>{};
}
if (flags & ObjectFlags::Inner)
{
if (to_begin)
++first;
if (to_end and first != last)
--last;
}
return to_end ? utf8_range(first, last) : utf8_range(last, first);
}
2013-06-01 14:22:57 +02:00
Selection select_surrounding(const Buffer& buffer, const Selection& selection,
const CodepointPair& matching,
ObjectFlags flags)
{
auto res = find_surrounding(selection.last(), matching, flags);
if (not res)
return selection;
if (flags == (ObjectFlags::ToBegin | ObjectFlags::ToEnd) and
matching.first != matching.second and not res->last().is_end() and
(*res == selection or Range{res->last(), res->first()} == selection))
{
res = find_surrounding(res->last() + 1, matching, flags);
return res ? Selection{*res} : selection;
}
return *res;
}
2013-06-01 14:22:57 +02:00
Selection select_to(const Buffer& buffer, const Selection& selection,
Codepoint c, int count, bool inclusive)
{
2012-10-08 19:12:09 +02:00
Utf8Iterator begin = selection.last();
Utf8Iterator end = begin;
do
{
++end;
skip_while(end, [c](Codepoint cur) { return cur != c; });
if (is_end(end))
return selection;
}
while (--count > 0);
return utf8_range(begin, inclusive ? end : end-1);
}
2013-06-01 14:22:57 +02:00
Selection select_to_reverse(const Buffer& buffer, const Selection& selection,
Codepoint c, int count, bool inclusive)
{
2012-10-08 19:12:09 +02:00
Utf8Iterator begin = selection.last();
Utf8Iterator end = begin;
do
{
--end;
skip_while_reverse(end, [c](Codepoint cur) { return cur != c; });
if (is_begin(end))
return selection;
}
while (--count > 0);
return utf8_range(begin, inclusive ? end : end+1);
}
2013-06-01 14:22:57 +02:00
Selection select_to_eol(const Buffer& buffer, const Selection& selection)
{
2012-10-08 19:12:09 +02:00
Utf8Iterator begin = selection.last();
Utf8Iterator end = begin + 1;
skip_while(end, [](Codepoint cur) { return not is_eol(cur); });
return utf8_range(begin, end-1);
}
2013-06-01 14:22:57 +02:00
Selection select_to_eol_reverse(const Buffer& buffer, const Selection& selection)
{
2012-10-08 19:12:09 +02:00
Utf8Iterator begin = selection.last();
Utf8Iterator end = begin - 1;
skip_while_reverse(end, [](Codepoint cur) { return not is_eol(cur); });
return utf8_range(begin, is_begin(end) ? end : end+1);
}
template<bool punctuation_is_word>
2013-06-01 14:22:57 +02:00
Selection select_whole_word(const Buffer& buffer, const Selection& selection, ObjectFlags flags)
{
2012-10-08 19:12:09 +02:00
Utf8Iterator first = selection.last();
Utf8Iterator last = first;
if (is_word<punctuation_is_word>(*first))
{
if (flags & ObjectFlags::ToBegin)
{
skip_while_reverse(first, is_word<punctuation_is_word>);
if (not is_word<punctuation_is_word>(*first))
++first;
}
if (flags & ObjectFlags::ToEnd)
{
skip_while(last, is_word<punctuation_is_word>);
if (not (flags & ObjectFlags::Inner))
skip_while(last, is_blank);
--last;
}
}
else if (not (flags & ObjectFlags::Inner))
{
if (flags & ObjectFlags::ToBegin)
{
skip_while_reverse(first, is_blank);
if (not is_word<punctuation_is_word>(*first))
return selection;
skip_while_reverse(first, is_word<punctuation_is_word>);
if (not is_word<punctuation_is_word>(*first))
++first;
}
if (flags & ObjectFlags::ToEnd)
{
skip_while(last, is_blank);
--last;
}
}
return (flags & ObjectFlags::ToEnd) ? utf8_range(first, last)
: utf8_range(last, first);
}
2013-06-01 14:22:57 +02:00
template Selection select_whole_word<false>(const Buffer&, const Selection&, ObjectFlags);
template Selection select_whole_word<true>(const Buffer&, const Selection&, ObjectFlags);
2013-06-01 14:22:57 +02:00
Selection select_whole_sentence(const Buffer& buffer, const Selection& selection, ObjectFlags flags)
2013-04-30 14:21:48 +02:00
{
BufferIterator first = selection.last();
BufferIterator last = first;
2013-04-30 14:21:48 +02:00
if (flags & ObjectFlags::ToBegin)
2013-04-30 14:21:48 +02:00
{
bool saw_non_blank = false;
while (not is_begin(first))
2013-04-30 14:21:48 +02:00
{
char cur = *first;
char prev = *(first-1);
if (not is_blank(cur))
saw_non_blank = true;
if (is_eol(prev) and is_eol(cur))
{
++first;
break;
}
else if (prev == '.' or prev == ';' or prev == '!' or prev == '?')
{
if (saw_non_blank)
break;
else if (flags & ObjectFlags::ToEnd)
last = first-1;
}
--first;
2013-04-30 14:21:48 +02:00
}
skip_while(first, is_blank);
2013-04-30 14:21:48 +02:00
}
if (flags & ObjectFlags::ToEnd)
2013-04-30 14:21:48 +02:00
{
while (not is_end(last))
{
char cur = *last;
if (cur == '.' or cur == ';' or cur == '!' or cur == '?' or
(is_eol(cur) and (is_end(last+1) or is_eol(*last+1))))
break;
++last;
}
if (not (flags & ObjectFlags::Inner) and not is_end(last))
{
++last;
skip_while(last, is_blank);
--last;
}
2013-04-30 14:21:48 +02:00
}
return (flags & ObjectFlags::ToEnd) ? Selection{first, last}
: Selection{last, first};
2013-04-30 14:21:48 +02:00
}
2013-06-01 14:22:57 +02:00
Selection select_whole_paragraph(const Buffer& buffer, const Selection& selection, ObjectFlags flags)
2013-04-30 14:29:18 +02:00
{
BufferIterator first = selection.last();
BufferIterator last = first;
2013-04-30 14:29:18 +02:00
if (flags & ObjectFlags::ToBegin and not is_begin(first))
2013-04-30 14:29:18 +02:00
{
skip_while_reverse(first, is_eol);
if (flags & ObjectFlags::ToEnd)
last = first;
while (not is_begin(first))
2013-04-30 14:29:18 +02:00
{
char cur = *first;
char prev = *(first-1);
if (is_eol(prev) and is_eol(cur))
{
++first;
break;
}
--first;
2013-04-30 14:29:18 +02:00
}
}
if (flags & ObjectFlags::ToEnd)
2013-04-30 14:29:18 +02:00
{
while (not is_end(last))
2013-04-30 14:29:18 +02:00
{
char cur = *last;
char prev = *(last-1);
if (is_eol(cur) and is_eol(prev))
{
if (not (flags & ObjectFlags::Inner))
skip_while(last, is_eol);
break;
}
++last;
2013-04-30 14:29:18 +02:00
}
--last;
2013-04-30 14:29:18 +02:00
}
return (flags & ObjectFlags::ToEnd) ? Selection{first, last}
: Selection{last, first};
2013-04-30 14:29:18 +02:00
}
2013-06-01 14:22:57 +02:00
Selection select_whole_lines(const Buffer& buffer, const Selection& selection)
{
2013-01-03 14:30:14 +01:00
// no need to be utf8 aware for is_eol as we only use \n as line seperator
BufferIterator first = selection.first();
BufferIterator last = selection.last();
BufferIterator& to_line_start = first <= last ? first : last;
BufferIterator& to_line_end = first <= last ? last : first;
2013-01-03 14:30:14 +01:00
--to_line_start;
skip_while_reverse(to_line_start, [](char cur) { return not is_eol(cur); });
if (is_eol(*to_line_start))
++to_line_start;
2011-11-22 19:58:05 +01:00
2013-01-03 14:30:14 +01:00
skip_while(to_line_end, [](char cur) { return not is_eol(cur); });
if (is_end(to_line_end))
--to_line_end;
2013-01-03 14:30:14 +01:00
return Selection(first, last);
}
2013-06-01 14:22:57 +02:00
Selection trim_partial_lines(const Buffer& buffer, const Selection& selection)
2013-04-16 14:30:11 +02:00
{
// same as select_whole_lines
BufferIterator first = selection.first();
BufferIterator last = selection.last();
BufferIterator& to_line_start = first <= last ? first : last;
BufferIterator& to_line_end = first <= last ? last : first;
while (not is_begin(to_line_start) and *(to_line_start-1) != '\n')
++to_line_start;
while (*(to_line_end+1) != '\n' and to_line_end != to_line_start)
--to_line_end;
return Selection(first, last);
}
2013-06-01 14:22:57 +02:00
Selection select_whole_buffer(const Buffer& buffer, const Selection&)
{
2012-10-08 19:12:09 +02:00
return Selection(buffer.begin(), utf8::previous(buffer.end()));
}
using MatchResults = boost::match_results<BufferIterator>;
static bool find_last_match(BufferIterator begin, const BufferIterator& end,
MatchResults& res, const Regex& regex)
{
MatchResults matches;
while (boost::regex_search(begin, end, matches, regex))
{
if (begin == matches[0].second)
break;
begin = matches[0].second;
res.swap(matches);
}
return not res.empty();
}
template<bool forward>
2013-06-01 14:22:57 +02:00
bool find_match_in_buffer(const Buffer& buffer, const BufferIterator pos,
MatchResults& matches, const Regex& ex)
{
if (forward)
2013-06-01 14:22:57 +02:00
return (boost::regex_search(pos, buffer.end(), matches, ex) or
boost::regex_search(buffer.begin(), pos, matches, ex));
else
2013-06-01 14:22:57 +02:00
return (find_last_match(buffer.begin(), pos, matches, ex) or
find_last_match(pos, buffer.end(), matches, ex));
}
template<bool forward>
2013-06-01 14:22:57 +02:00
Selection select_next_match(const Buffer& buffer, const Selection& selection, const Regex& regex)
{
// regex matching do not use Utf8Iterator as boost::regex handle utf8
// decoding itself
BufferIterator begin = selection.last();
BufferIterator end = begin;
CaptureList captures;
MatchResults matches;
2013-06-01 14:22:57 +02:00
if (find_match_in_buffer<forward>(buffer, utf8::next(begin), matches, regex))
{
begin = matches[0].first;
end = matches[0].second;
for (auto& match : matches)
captures.push_back(String(match.first, match.second));
}
else
throw runtime_error("'" + regex.str() + "': no matches found");
if (begin == end)
++end;
end = utf8::previous(end);
if (not forward)
std::swap(begin, end);
return Selection{begin, end, std::move(captures)};
}
2013-06-01 14:22:57 +02:00
template Selection select_next_match<true>(const Buffer&, const Selection&, const Regex&);
template Selection select_next_match<false>(const Buffer&, const Selection&, const Regex&);
2013-06-01 14:22:57 +02:00
SelectionList select_all_matches(const Buffer& buffer, const Selection& selection, const Regex& regex)
{
auto sel_end = utf8::next(selection.max());
RegexIterator re_it(selection.min(), sel_end, regex);
RegexIterator re_end;
SelectionList result;
for (; re_it != re_end; ++re_it)
{
BufferIterator begin = (*re_it)[0].first;
BufferIterator end = (*re_it)[0].second;
if (begin == sel_end)
continue;
CaptureList captures;
for (auto& match : *re_it)
captures.push_back(String(match.first, match.second));
result.push_back(Selection(begin, begin == end ? end : utf8::previous(end),
std::move(captures)));
}
return result;
}
2013-06-01 14:22:57 +02:00
SelectionList split_selection(const Buffer& buffer, const Selection& selection,
const Regex& regex)
{
auto sel_end = utf8::next(selection.max());
RegexIterator re_it(selection.min(), sel_end, regex,
boost::regex_constants::match_nosubs);
RegexIterator re_end;
SelectionList result;
BufferIterator begin = selection.min();
for (; re_it != re_end; ++re_it)
{
BufferIterator end = (*re_it)[0].first;
result.push_back(Selection(begin, (begin == end) ? end : utf8::previous(end)));
begin = (*re_it)[0].second;
}
result.push_back(Selection(begin, selection.max()));
return result;
}
}