/[LeafOK_CVS]/lbbs/utils/lib/str_process.inc.php
ViewVC logotype

Annotation of /lbbs/utils/lib/str_process.inc.php

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1.3 - (hide annotations)
Wed Jul 2 04:17:33 2025 UTC (8 months, 2 weeks ago) by sysadm
Branch: MAIN
Changes since 1.2: +19 -9 lines
Support UTF8 instead of GBK

1 sysadm 1.1 <?php
2 sysadm 1.2 function str_length(string $str) : int
3 sysadm 1.1 {
4     $len = strlen($str);
5     $ret = 0;
6    
7     for ($i = 0; $i < $len; $i++)
8     {
9     $c = $str[$i];
10    
11 sysadm 1.3 // Process UTF-8 Chinese characters
12 sysadm 1.1 $v1 = ord($c);
13 sysadm 1.3 if ($v1 & 0b10000000) //head of multi-byte character
14 sysadm 1.1 {
15 sysadm 1.3 $v2 = ($v1 & 0b01110000) << 1;
16     while ($v2 & 0b10000000)
17     {
18     $i++;
19     $c .= $str[$i];
20     $v2 = ($v2 & 0b01111111) << 1;
21     }
22 sysadm 1.1
23     $ret += 2;
24     }
25     else
26     {
27     $ret++;
28     }
29     }
30    
31     return $ret;
32     }
33    
34     function split_line(string $str, string $prefix = "", int $width = 76, int $lines_limit = PHP_INT_MAX, string $end_of_line = "\n") : string
35     {
36     if ($width <= 0)
37     {
38     return $str;
39     }
40    
41     $result = "";
42     $len = strlen($str);
43     $prefix_len = str_length($prefix);
44    
45     $lines_count = 0;
46    
47     $line = $prefix;
48     $line_len = $prefix_len;
49     for ($i = 0; $i < $len && $lines_count < $lines_limit; $i++)
50     {
51     $c = $str[$i];
52    
53     // Skip special characters
54     if ($c == "\r" || $c == "\7")
55     {
56     continue;
57     }
58    
59     if ($c == "\n")
60     {
61     if ($lines_count + 1 >= $lines_limit)
62     {
63     break;
64     }
65    
66     $result .= ($line . $end_of_line);
67     $lines_count++;
68     $line = $prefix;
69     $line_len = $prefix_len;
70     continue;
71     }
72    
73 sysadm 1.3 // Process UTF-8 Chinese characters
74 sysadm 1.1 $v1 = ord($c);
75 sysadm 1.3 if ($v1 & 0b10000000) //head of multi-byte character
76 sysadm 1.1 {
77 sysadm 1.3 $v2 = ($v1 & 0b01110000) << 1;
78     while ($v2 & 0b10000000)
79     {
80     $i++;
81     $c .= $str[$i];
82     $v2 = ($v2 & 0b01111111) << 1;
83     }
84 sysadm 1.1
85 sysadm 1.3 // Each UTF-8 CJK character should use two character length for display
86 sysadm 1.1 if ($line_len + 2 > $width)
87     {
88     if ($lines_count + 1 >= $lines_limit)
89     {
90     break;
91     }
92    
93     $result .= ($line . $end_of_line);
94     $lines_count++;
95     $line = $prefix;
96     $line_len = $prefix_len;
97     }
98     $line_len += 2;
99     }
100     else
101     {
102     $line_len++;
103     }
104    
105     if ($line_len > $width)
106     {
107     if ($lines_count + 1 >= $lines_limit)
108     {
109     break;
110     }
111    
112     $result .= ($line . $end_of_line);
113     $lines_count++;
114     $line = $prefix;
115     $line_len = $prefix_len + 1;
116     }
117    
118     $line .= $c;
119     }
120    
121     if ($lines_count < $lines_limit)
122     {
123     $result .= $line;
124     }
125    
126     return $result;
127     }

webmaster@leafok.com
ViewVC Help
Powered by ViewVC 1.3.0-beta1