1 | static List<String> nlTok3(String s) {
|
2 | List<String> tok = new ArrayList<String>(); |
3 | int l = s.length(); |
4 | |
5 | int i = 0; |
6 | while (i < l) {
|
7 | int j = i; |
8 | char c; String cc; |
9 | |
10 | // scan for whitespace |
11 | while (j < l) {
|
12 | c = s.charAt(j); |
13 | cc = s.substring(j, Math.min(j+2, l)); |
14 | if (c == ' ' || c == '\t' || c == '\r' || c == '\n') |
15 | ++j; |
16 | else if (cc.equals("/*")) {
|
17 | do ++j; while (j < l && !s.substring(j, Math.min(j+2, l)).equals("*/"));
|
18 | j = Math.min(j+2, l); |
19 | } else if (cc.equals("//")) {
|
20 | do ++j; while (j < l && "\r\n".indexOf(s.charAt(j)) < 0); |
21 | } else |
22 | break; |
23 | } |
24 | |
25 | tok.add(s.substring(i, j)); |
26 | i = j; |
27 | if (i >= l) break; |
28 | c = s.charAt(i); |
29 | cc = s.substring(i, Math.min(i+2, l)); |
30 | |
31 | // scan for non-whitespace |
32 | if (c == '\u201C' || c == '\u201D') c = '"'; // normalize quotes |
33 | if (c == '\'' || c == '"') {
|
34 | char opener = c; |
35 | ++j; |
36 | while (j < l) {
|
37 | char _c = s.charAt(j); |
38 | if (_c == '\u201C' || _c == '\u201D') _c = '"'; // normalize quotes |
39 | if (_c == opener) {
|
40 | ++j; |
41 | break; |
42 | } else if (s.charAt(j) == '\\' && j+1 < l) |
43 | j += 2; |
44 | else |
45 | ++j; |
46 | } |
47 | if (j-1 >= i+1) {
|
48 | tok.add(opener + s.substring(i+1, j-1) + opener); |
49 | i = j; |
50 | continue; |
51 | } |
52 | } else if (Character.isJavaIdentifierStart(c)) |
53 | do ++j; while (j < l && (Character.isJavaIdentifierPart(s.charAt(j)) /*|| s.charAt(j) == '\''*/ )); |
54 | else if (Character.isDigit(c)) |
55 | do ++j; while (j < l && Character.isDigit(s.charAt(j))); |
56 | else if (cc.equals("[[")) {
|
57 | do ++j; while (j+1 < l && !s.substring(j, j+2).equals("]]"));
|
58 | j = Math.min(j+2, l); |
59 | } else if (cc.equals("[=") && i+2 < l && s.charAt(i+2) == '[') {
|
60 | do ++j; while (j+2 < l && !s.substring(j, j+3).equals("]=]"));
|
61 | j = Math.min(j+3, l); |
62 | } else if (s.substring(j, Math.min(j+3, l)).equals("..."))
|
63 | j += 3; |
64 | else if (c == '$' || c == '#') |
65 | do ++j; while (j < l && Character.isDigit(s.charAt(j))); |
66 | else |
67 | ++j; |
68 | |
69 | tok.add(s.substring(i, j)); |
70 | i = j; |
71 | } |
72 | |
73 | if ((tok.size() % 2) == 0) tok.add("");
|
74 | return tok; |
75 | } |
Began life as a copy of #1000769
download show line numbers debug dex old transpilations
Travelled to 14 computer(s): aoiabmzegqzx, bhatertpkbcr, cbybwowwnfue, cfunsshuasjs, ddnzoavkxhuk, gwrvuhgaqvyk, ishqpsrjomds, lpdgvwnxivlt, mqqgnosmbjvj, pyentgdyhuwx, pzhvpgtvlbxg, tslmcundralx, tvejysmllsmz, vouqrxazstgt
No comments. add comment
| Snippet ID: | #1004163 |
| Snippet name: | nlTok3 - with quoted strings, without ' in identifiers |
| Eternal ID of this version: | #1004163/1 |
| Text MD5: | ad55cef92456fa3948357bea6380b799 |
| Author: | stefan |
| Category: | |
| Type: | JavaX fragment (include) |
| Public (visible to everyone): | Yes |
| Archived (hidden from active list): | No |
| Created/modified: | 2016-08-08 17:41:31 |
| Source code size: | 2300 bytes / 75 lines |
| Pitched / IR pitched: | No / No |
| Views / Downloads: | 512 / 516 |
| Referenced in: | [show references] |