1 #include "vterm_internal.h"
2 
3 #define UNICODE_INVALID 0xFFFD
4 
5 #if defined(DEBUG) && DEBUG > 1
6 # define DEBUG_PRINT_UTF8
7 #endif
8 
9 struct UTF8DecoderData {
10   // number of bytes remaining in this codepoint
11   int bytes_remaining;
12 
13   // number of bytes total in this codepoint once it's finished
14   // (for detecting overlongs)
15   int bytes_total;
16 
17   int this_cp;
18 };
19 
init_utf8(VTermEncoding * enc,void * data_)20 static void init_utf8(VTermEncoding *enc, void *data_)
21 {
22   struct UTF8DecoderData *data = data_;
23 
24   data->bytes_remaining = 0;
25   data->bytes_total     = 0;
26 }
27 
decode_utf8(VTermEncoding * enc,void * data_,uint32_t cp[],int * cpi,int cplen,const char bytes[],size_t * pos,size_t bytelen)28 static void decode_utf8(VTermEncoding *enc, void *data_,
29                         uint32_t cp[], int *cpi, int cplen,
30                         const char bytes[], size_t *pos, size_t bytelen)
31 {
32   struct UTF8DecoderData *data = data_;
33 
34 #ifdef DEBUG_PRINT_UTF8
35   printf("BEGIN UTF-8\n");
36 #endif
37 
38   for(; *pos < bytelen && *cpi < cplen; (*pos)++) {
39     unsigned char c = bytes[*pos];
40 
41 #ifdef DEBUG_PRINT_UTF8
42     printf(" pos=%zd c=%02x rem=%d\n", *pos, c, data->bytes_remaining);
43 #endif
44 
45     if(c < 0x20)
46       return;
47 
48     else if(c >= 0x20 && c < 0x80) {
49       if(data->bytes_remaining)
50         cp[(*cpi)++] = UNICODE_INVALID;
51 
52       cp[(*cpi)++] = c;
53 #ifdef DEBUG_PRINT_UTF8
54       printf(" UTF-8 char: U+%04x\n", c);
55 #endif
56       data->bytes_remaining = 0;
57     }
58 
59     else if(c >= 0x80 && c < 0xc0) {
60       if(!data->bytes_remaining) {
61         cp[(*cpi)++] = UNICODE_INVALID;
62         continue;
63       }
64 
65       data->this_cp <<= 6;
66       data->this_cp |= c & 0x3f;
67       data->bytes_remaining--;
68 
69       if(!data->bytes_remaining) {
70 #ifdef DEBUG_PRINT_UTF8
71         printf(" UTF-8 raw char U+%04x bytelen=%d ", data->this_cp, data->bytes_total);
72 #endif
73         // Check for overlong sequences
74         switch(data->bytes_total) {
75         case 2:
76           if(data->this_cp <  0x0080) data->this_cp = UNICODE_INVALID; break;
77         case 3:
78           if(data->this_cp <  0x0800) data->this_cp = UNICODE_INVALID; break;
79         case 4:
80           if(data->this_cp < 0x10000) data->this_cp = UNICODE_INVALID; break;
81         case 5:
82           if(data->this_cp < 0x200000) data->this_cp = UNICODE_INVALID; break;
83         case 6:
84           if(data->this_cp < 0x4000000) data->this_cp = UNICODE_INVALID; break;
85         }
86         // Now look for plain invalid ones
87         if((data->this_cp >= 0xD800 && data->this_cp <= 0xDFFF) ||
88            data->this_cp == 0xFFFE ||
89            data->this_cp == 0xFFFF)
90           data->this_cp = UNICODE_INVALID;
91 #ifdef DEBUG_PRINT_UTF8
92         printf(" char: U+%04x\n", data->this_cp);
93 #endif
94         cp[(*cpi)++] = data->this_cp;
95       }
96     }
97 
98     else if(c >= 0xc0 && c < 0xe0) {
99       if(data->bytes_remaining)
100         cp[(*cpi)++] = UNICODE_INVALID;
101 
102       data->this_cp = c & 0x1f;
103       data->bytes_total = 2;
104       data->bytes_remaining = 1;
105     }
106 
107     else if(c >= 0xe0 && c < 0xf0) {
108       if(data->bytes_remaining)
109         cp[(*cpi)++] = UNICODE_INVALID;
110 
111       data->this_cp = c & 0x0f;
112       data->bytes_total = 3;
113       data->bytes_remaining = 2;
114     }
115 
116     else if(c >= 0xf0 && c < 0xf8) {
117       if(data->bytes_remaining)
118         cp[(*cpi)++] = UNICODE_INVALID;
119 
120       data->this_cp = c & 0x07;
121       data->bytes_total = 4;
122       data->bytes_remaining = 3;
123     }
124 
125     else if(c >= 0xf8 && c < 0xfc) {
126       if(data->bytes_remaining)
127         cp[(*cpi)++] = UNICODE_INVALID;
128 
129       data->this_cp = c & 0x03;
130       data->bytes_total = 5;
131       data->bytes_remaining = 4;
132     }
133 
134     else if(c >= 0xfc && c < 0xfe) {
135       if(data->bytes_remaining)
136         cp[(*cpi)++] = UNICODE_INVALID;
137 
138       data->this_cp = c & 0x01;
139       data->bytes_total = 6;
140       data->bytes_remaining = 5;
141     }
142 
143     else {
144       cp[(*cpi)++] = UNICODE_INVALID;
145     }
146   }
147 }
148 
149 static VTermEncoding encoding_utf8 = {
150   .init   = &init_utf8,
151   .decode = &decode_utf8,
152 };
153 
decode_usascii(VTermEncoding * enc,void * data,uint32_t cp[],int * cpi,int cplen,const char bytes[],size_t * pos,size_t bytelen)154 static void decode_usascii(VTermEncoding *enc, void *data,
155                            uint32_t cp[], int *cpi, int cplen,
156                            const char bytes[], size_t *pos, size_t bytelen)
157 {
158   int is_gr = bytes[*pos] & 0x80;
159 
160   for(; *pos < bytelen && *cpi < cplen; (*pos)++) {
161     unsigned char c = bytes[*pos] ^ is_gr;
162 
163     if(c < 0x20 || c >= 0x80)
164       return;
165 
166     cp[(*cpi)++] = c;
167   }
168 }
169 
170 static VTermEncoding encoding_usascii = {
171   .decode = &decode_usascii,
172 };
173 
174 struct StaticTableEncoding {
175   const VTermEncoding enc;
176   const uint32_t chars[128];
177 };
178 
decode_table(VTermEncoding * enc,void * data,uint32_t cp[],int * cpi,int cplen,const char bytes[],size_t * pos,size_t bytelen)179 static void decode_table(VTermEncoding *enc, void *data,
180                          uint32_t cp[], int *cpi, int cplen,
181                          const char bytes[], size_t *pos, size_t bytelen)
182 {
183   struct StaticTableEncoding *table = (struct StaticTableEncoding *)enc;
184   int is_gr = bytes[*pos] & 0x80;
185 
186   for(; *pos < bytelen && *cpi < cplen; (*pos)++) {
187     unsigned char c = bytes[*pos] ^ is_gr;
188 
189     if(c < 0x20 || c >= 0x80)
190       return;
191 
192     if(table->chars[c])
193       cp[(*cpi)++] = table->chars[c];
194     else
195       cp[(*cpi)++] = c;
196   }
197 }
198 
199 #include "encoding/DECdrawing.inc"
200 #include "encoding/uk.inc"
201 
202 static struct {
203   VTermEncodingType type;
204   char designation;
205   VTermEncoding *enc;
206 }
207 encodings[] = {
208   { ENC_UTF8,      'u', &encoding_utf8 },
209   { ENC_SINGLE_94, '0', (VTermEncoding*)&encoding_DECdrawing },
210   { ENC_SINGLE_94, 'A', (VTermEncoding*)&encoding_uk },
211   { ENC_SINGLE_94, 'B', &encoding_usascii },
212   { 0 },
213 };
214 
215 /* This ought to be INTERNAL but isn't because it's used by unit testing */
vterm_lookup_encoding(VTermEncodingType type,char designation)216 VTermEncoding *vterm_lookup_encoding(VTermEncodingType type, char designation)
217 {
218   for(int i = 0; encodings[i].designation; i++)
219     if(encodings[i].type == type && encodings[i].designation == designation)
220       return encodings[i].enc;
221   return NULL;
222 }
223