about summary refs log tree commit diff stats
path: root/014literal_string.cc
blob: eb5422ff481f7bf338e804587767c45ec0825852 (plain) (blame)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
//: For convenience, some instructions will take literal arrays of characters
//: (text or strings).
//:
//: Instead of quotes, we'll use [] to delimit strings. That'll reduce the
//: need for escaping since we can support nested brackets. And we can also
//: imagine that 'recipe' might one day itself be defined in Mu, doing its own
//: parsing.

:(scenarios load)
:(scenario string_literal)
def main [
  1:address:array:character <- copy [abc def]
]
+parse:   ingredient: {"abc def": "literal-string"}

:(scenario string_literal_with_colons)
def main [
  1:address:array:character <- copy [abc:def/ghi]
]
+parse:   ingredient: {"abc:def/ghi": "literal-string"}

:(before "End Mu Types Initialization")
put(Type_ordinal, "literal-string", 0);

:(before "End next_word Special-cases")
if (in.peek() == '[') {
  string result = slurp_quoted(in);
  skip_whitespace_and_comments_but_not_newline(in);
  return result;
}

:(code)
string slurp_quoted(istream& in) {
  ostringstream out;
  assert(has_data(in));  assert(in.peek() == '[');  out << static_cast<char>(in.get());  // slurp the '['
  if (is_code_string(in, out))
    slurp_quoted_comment_aware(in, out);
  else
    slurp_quoted_comment_oblivious(in, out);
  return out.str();
}

// A string is a code string (ignores comments when scanning for matching
// brackets) if it contains a newline at the start before any non-whitespace.
bool is_code_string(istream& in, ostream& out) {
  while (has_data(in)) {
    char c = in.get();
    if (!isspace(c)) {
      in.putback(c);
      return false;
    }
    out << c;
    if (c == '\n') {
      return true;
    }
  }
  return false;
}

// Read a regular string. Regular strings can only contain other regular
// strings.
void slurp_quoted_comment_oblivious(istream& in, ostream& out) {
  int brace_depth = 1;
  while (has_data(in)) {
    char c = in.get();
    if (c == '\\') {
      slurp_one_past_backslashes(in, out);
      continue;
    }
    out << c;
    if (c == '[') ++brace_depth;
    if (c == ']') --brace_depth;
    if (brace_depth == 0) break;
  }
  if (!has_data(in) && brace_depth > 0) {
    raise << "unbalanced '['\n" << end();
    out.clear();
  }
}

// Read a code string. Code strings can contain either code or regular strings.
void slurp_quoted_comment_aware(istream& in, ostream& out) {
  char c;
  while (in >> c) {
    if (c == '\\') {
      slurp_one_past_backslashes(in, out);
      continue;
    }
    if (c == '#') {
      out << c;
      while (has_data(in) && in.peek() != '\n') out << static_cast<char>(in.get());
      continue;
    }
    if (c == '[') {
      in.putback(c);
      // recurse
      out << slurp_quoted(in);
      continue;
    }
    out << c;
    if (c == ']') return;
  }
  raise << "unbalanced '['\n" << end();
  out.clear();
}

:(after "Parsing reagent(string s)")
if (starts_with(s, "[")) {
  if (*s.rbegin() != ']') return;  // unbalanced bracket; handled elsewhere
  name = s;
  // delete [] delimiters
  name.erase(0, 1);
  strip_last(name);
  type = new type_tree("literal-string", 0);
  return;
}

//: Unlike other reagents, escape newlines in literal strings to make them
//: more friendly to trace().

:(after "string to_string(const reagent& r)")
  if (is_literal_text(r))
    return emit_literal_string(r.name);

:(code)
bool is_literal_text(const reagent& x) {
  return x.type && x.type->name == "literal-string";
}

string emit_literal_string(string name) {
  size_t pos = 0;
  while (pos != string::npos)
    pos = replace(name, "\n", "\\n", pos);
  return "{\""+name+"\": \"literal-string\"}";
}

size_t replace(string& str, const string& from, const string& to, size_t n) {
  size_t result = str.find(from, n);
  if (result != string::npos)
    str.replace(result, from.length(), to);
  return result;
}

void strip_last(string& s) {
  if (!s.empty()) s.erase(SIZE(s)-1);
}

void slurp_one_past_backslashes(istream& in, ostream& out) {
  // When you encounter a backslash, strip it out and pass through any
  // following run of backslashes. If we 'escaped' a single following
  // character, then the character '\' would be:
  //   '\\' escaped once
  //   '\\\\' escaped twice
  //   '\\\\\\\\' escaped thrice (8 backslashes)
  // ..and so on. With our approach it'll be:
  //   '\\' escaped once
  //   '\\\' escaped twice
  //   '\\\\' escaped thrice
  // This only works as long as backslashes aren't also overloaded to create
  // special characters. So Mu doesn't follow C's approach of overloading
  // backslashes both to escape quote characters and also as a notation for
  // unprintable characters like '\n'.
  while (has_data(in)) {
    char c = in.get();
    out << c;
    if (c != '\\') break;
  }
}

:(scenario string_literal_nested)
def main [
  1:address:array:character <- copy [abc [def]]
]
+parse:   ingredient: {"abc [def]": "literal-string"}

:(scenario string_literal_escaped)
def main [
  1:address:array:character <- copy [abc \[def]
]
+parse:   ingredient: {"abc [def": "literal-string"}

:(scenario string_literal_escaped_twice)
def main [
  1:address:array:character <- copy [
abc \\[def]
]
+parse:   ingredient: {"\nabc \[def": "literal-string"}

:(scenario string_literal_and_comment)
def main [
  1:address:array:character <- copy [abc]  # comment
]
+parse: --- defining main
+parse: instruction: copy
+parse:   number of ingredients: 1
+parse:   ingredient: {"abc": "literal-string"}
+parse:   product: {1: ("address" "array" "character")}

:(scenario string_literal_escapes_newlines_in_trace)
def main [
  copy [abc
def]
]
+parse:   ingredient: {"abc\ndef": "literal-string"}

:(scenario string_literal_can_skip_past_comments)
def main [
  copy [
    # ']' inside comment
    bar
  ]
]
+parse:   ingredient: {"\n    # ']' inside comment\n    bar\n  ": "literal-string"}

:(scenario string_literal_empty)
def main [
  copy []
]
+parse:   ingredient: {"": "literal-string"}

:(scenario multiple_unfinished_recipes)
% Hide_errors = true;
def f1 [
def f2 [
+error: unbalanced '['
/span>Cursor_line+1].data Lines[Cursor_line].fragments = nil table.remove(Lines, Cursor_line+1) end end save_to_disk(Lines, Filename) elseif chord == 'up' then assert(Lines[Cursor_line].mode == 'text') local new_cursor_line = Cursor_line while new_cursor_line > 1 do new_cursor_line = new_cursor_line-1 if Lines[new_cursor_line].mode == 'text' then local old_x = Text.cursor_x(Lines[new_cursor_line].data, Cursor_pos) Cursor_line = new_cursor_line Cursor_pos = Text.nearest_cursor_pos(Lines[Cursor_line].data, old_x, Cursor_pos) break end end if Cursor_line < Screen_top_line then Screen_top_line = Cursor_line end elseif chord == 'down' then assert(Lines[Cursor_line].mode == 'text') local new_cursor_line = Cursor_line while new_cursor_line < #Lines do new_cursor_line = new_cursor_line+1 if Lines[new_cursor_line].mode == 'text' then local old_x = Text.cursor_x(Lines[new_cursor_line].data, Cursor_pos) Cursor_line = new_cursor_line Cursor_pos = Text.nearest_cursor_pos(Lines[Cursor_line].data, old_x, Cursor_pos) break end end if Cursor_line > Screen_bottom_line then Screen_top_line = Cursor_line end end end function Text.move_cursor_down_to_next_text_line_while_scrolling_again_if_necessary() while Cursor_line <= #Lines do if Lines[Cursor_line].mode == 'text' then break end Cursor_line = Cursor_line + 1 end -- hack: insert a text line at bottom of file if necessary if Cursor_line > #Lines then assert(Cursor_line == #Lines+1) table.insert(Lines, {mode='text', data=''}) end if Cursor_line > Screen_bottom_line then Screen_top_line = Cursor_line Text.scroll_up_while_cursor_on_screen() end end function Text.scroll_up_while_cursor_on_screen() local y = Screen_height - math.floor(15*Zoom) -- for Cursor_line while true do if Screen_top_line == 1 then break end y = y - math.floor(15*Zoom) if Lines[Screen_top_line].mode == 'drawing' then y = y - Drawing.pixels(Lines[Screen_top_line].h) end if y < math.floor(15*Zoom) then break end Screen_top_line = Screen_top_line - 1 end end function Text.in_line(line, x,y) if line.y == nil then return false end -- outside current page if x < 16 then return false end if y < line.y then return false end if line.screen_line_starting_pos == nil then return y < line.y + math.floor(15*Zoom) end return y < line.y + #line.screen_line_starting_pos * math.floor(15*Zoom) end function Text.move_cursor(line_index, line, mx, my) Cursor_line = line_index if line.screen_line_starting_pos == nil then --? print('single screen line') Cursor_pos = Text.nearest_cursor_pos(line.data, mx) return end assert(line.fragments) assert(my >= line.y) --? print('move_cursor', mx, my) if my < line.y + math.floor(15*Zoom) then --? print('first screen line') Cursor_pos = Text.nearest_cursor_pos(line.data, mx) return end -- duplicate some logic from Text.draw local y = line.y for _,screen_line_starting_pos in ipairs(line.screen_line_starting_pos) do --? print('screen line:', screen_line_starting_pos) local nexty = y + math.floor(15*Zoom) if my < nexty then local s = string.sub(line.data, screen_line_starting_pos) Cursor_pos = screen_line_starting_pos + Text.nearest_cursor_pos(s, mx) - 1 return end y = nexty end assert(false) end -- manual test: -- line: abc -- def -- gh -- fragments: abc, def, gh -- click inside e -- line_starting_pos = 1 + 3 = 4 -- nearest_cursor_pos('defgh', mx) = 2 -- Cursor_pos = 4 + 2 - 1 = 5 -- manual test: -- click inside h -- line_starting_pos = 1 + 3 + 3 = 7 -- nearest_cursor_pos('gh', mx) = 2 -- Cursor_pos = 7 + 2 - 1 = 8 function Text.nearest_cursor_pos(line, x, hint) if x == 0 then return 1 end local len = utf8.len(line) local max_x = Text.cursor_x(line, len+1) if x > max_x then return len+1 end if hint then local currx = Text.cursor_x(line, hint) if currx > x-2 and currx < x+2 then return hint end end local left, right = 1, len+1 if hint then if currx > x then right = hint else left = hint end end --? print('--') while true do local curr = math.floor((left+right)/2) local currxmin = Text.cursor_x(line, curr) local currxmax = Text.cursor_x(line, curr+1) --? print(x, left, right, curr, currxmin, currxmax) if currxmin <= x and x < currxmax then return curr end if left >= right-1 then return right end if currxmin > x then right = curr else left = curr end end assert(false) end function Text.cursor_x(line_data, cursor_pos) local line_before_cursor = line_data:sub(1, cursor_pos-1) local text_before_cursor = love.graphics.newText(love.graphics.getFont(), line_before_cursor) return 25 + math.floor(text_before_cursor:getWidth()*Zoom) end function Text.cursor_x2(s, cursor_pos) local s_before_cursor = s:sub(1, cursor_pos-1) local text_before_cursor = love.graphics.newText(love.graphics.getFont(), s_before_cursor) return math.floor(text_before_cursor:getWidth()*Zoom) end return Text