lex.c - 9base - revived minimalist port of Plan 9 userland to Unix
(HTM) git clone git://git.suckless.org/9base
(DIR) Log
(DIR) Files
(DIR) Refs
(DIR) README
(DIR) LICENSE
---
lex.c (6203B)
---
1 #include "rc.h"
2 #include "exec.h"
3 #include "io.h"
4 #include "getflags.h"
5 #include "fns.h"
6 int getnext(void);
7
8 int
9 wordchr(int c)
10 {
11 return !strchr("\n \t#;&|^$=`'{}()<>", c) && c!=EOF;
12 }
13
14 int
15 idchr(int c)
16 {
17 /*
18 * Formerly:
19 * return 'a'<=c && c<='z' || 'A'<=c && c<='Z' || '0'<=c && c<='9'
20 * || c=='_' || c=='*';
21 */
22 return c>' ' && !strchr("!\"#$%&'()+,-./:;<=>?@[\\]^`{|}~", c);
23 }
24 int future = EOF;
25 int doprompt = 1;
26 int inquote;
27 int incomm;
28 /*
29 * Look ahead in the input stream
30 */
31
32 int
33 nextc(void)
34 {
35 if(future==EOF)
36 future = getnext();
37 return future;
38 }
39 /*
40 * Consume the lookahead character.
41 */
42
43 int
44 advance(void)
45 {
46 int c = nextc();
47 lastc = future;
48 future = EOF;
49 return c;
50 }
51 /*
52 * read a character from the input stream
53 */
54
55 int
56 getnext(void)
57 {
58 int c;
59 static int peekc = EOF;
60 if(peekc!=EOF){
61 c = peekc;
62 peekc = EOF;
63 return c;
64 }
65 if(runq->eof)
66 return EOF;
67 if(doprompt)
68 pprompt();
69 c = rchr(runq->cmdfd);
70 if(!inquote && c=='\\'){
71 c = rchr(runq->cmdfd);
72 if(c=='\n' && !incomm){ /* don't continue a comment */
73 doprompt = 1;
74 c=' ';
75 }
76 else{
77 peekc = c;
78 c='\\';
79 }
80 }
81 doprompt = doprompt || c=='\n' || c==EOF;
82 if(c==EOF)
83 runq->eof++;
84 else if(flag['V'] || ndot>=2 && flag['v']) pchr(err, c);
85 return c;
86 }
87
88 void
89 pprompt(void)
90 {
91 var *prompt;
92 if(runq->iflag){
93 pstr(err, promptstr);
94 flush(err);
95 prompt = vlook("prompt");
96 if(prompt->val && prompt->val->next)
97 promptstr = prompt->val->next->word;
98 else
99 promptstr="\t";
100 }
101 runq->lineno++;
102 doprompt = 0;
103 }
104
105 void
106 skipwhite(void)
107 {
108 int c;
109 for(;;){
110 c = nextc();
111 /* Why did this used to be if(!inquote && c=='#') ?? */
112 if(c=='#'){
113 incomm = 1;
114 for(;;){
115 c = nextc();
116 if(c=='\n' || c==EOF) {
117 incomm = 0;
118 break;
119 }
120 advance();
121 }
122 }
123 if(c==' ' || c=='\t')
124 advance();
125 else return;
126 }
127 }
128
129 void
130 skipnl(void)
131 {
132 int c;
133 for(;;){
134 skipwhite();
135 c = nextc();
136 if(c!='\n')
137 return;
138 advance();
139 }
140 }
141
142 int
143 nextis(int c)
144 {
145 if(nextc()==c){
146 advance();
147 return 1;
148 }
149 return 0;
150 }
151
152 char*
153 addtok(char *p, int val)
154 {
155 if(p==0)
156 return 0;
157 if(p==&tok[NTOK-1]){
158 *p = 0;
159 yyerror("token buffer too short");
160 return 0;
161 }
162 *p++=val;
163 return p;
164 }
165
166 char*
167 addutf(char *p, int c)
168 {
169 p = addtok(p, c);
170 if(twobyte(c)) /* 2-byte escape */
171 return addtok(p, advance());
172 if(threebyte(c)){ /* 3-byte escape */
173 p = addtok(p, advance());
174 return addtok(p, advance());
175 }
176 return p;
177 }
178 int lastdol; /* was the last token read '$' or '$#' or '"'? */
179 int lastword; /* was the last token read a word or compound word terminator? */
180
181 int
182 yylex(void)
183 {
184 int c, d = nextc();
185 char *w = tok;
186 struct tree *t;
187 yylval.tree = 0;
188 /*
189 * Embarassing sneakiness: if the last token read was a quoted or unquoted
190 * WORD then we alter the meaning of what follows. If the next character
191 * is `(', we return SUB (a subscript paren) and consume the `('. Otherwise,
192 * if the next character is the first character of a simple or compound word,
193 * we insert a `^' before it.
194 */
195 if(lastword){
196 lastword = 0;
197 if(d=='('){
198 advance();
199 strcpy(tok, "( [SUB]");
200 return SUB;
201 }
202 if(wordchr(d) || d=='\'' || d=='`' || d=='$' || d=='"'){
203 strcpy(tok, "^");
204 return '^';
205 }
206 }
207 inquote = 0;
208 skipwhite();
209 switch(c = advance()){
210 case EOF:
211 lastdol = 0;
212 strcpy(tok, "EOF");
213 return EOF;
214 case '$':
215 lastdol = 1;
216 if(nextis('#')){
217 strcpy(tok, "$#");
218 return COUNT;
219 }
220 if(nextis('"')){
221 strcpy(tok, "$\"");
222 return '"';
223 }
224 strcpy(tok, "$");
225 return '$';
226 case '&':
227 lastdol = 0;
228 if(nextis('&')){
229 skipnl();
230 strcpy(tok, "&&");
231 return ANDAND;
232 }
233 strcpy(tok, "&");
234 return '&';
235 case '|':
236 lastdol = 0;
237 if(nextis(c)){
238 skipnl();
239 strcpy(tok, "||");
240 return OROR;
241 }
242 case '<':
243 case '>':
244 lastdol = 0;
245 /*
246 * funny redirection tokens:
247 * redir: arrow | arrow '[' fd ']'
248 * arrow: '<' | '<<' | '>' | '>>' | '|'
249 * fd: digit | digit '=' | digit '=' digit
250 * digit: '0'|'1'|'2'|'3'|'4'|'5'|'6'|'7'|'8'|'9'
251 * some possibilities are nonsensical and get a message.
252 */
253 *w++=c;
254 t = newtree();
255 switch(c){
256 case '|':
257 t->type = PIPE;
258 t->fd0 = 1;
259 t->fd1 = 0;
260 break;
261 case '>':
262 t->type = REDIR;
263 if(nextis(c)){
264 t->rtype = APPEND;
265 *w++=c;
266 }
267 else t->rtype = WRITE;
268 t->fd0 = 1;
269 break;
270 case '<':
271 t->type = REDIR;
272 if(nextis(c)){
273 t->rtype = HERE;
274 *w++=c;
275 } else if (nextis('>')){
276 t->rtype = RDWR;
277 *w++=c;
278 } else t->rtype = READ;
279 t->fd0 = 0;
280 break;
281 }
282 if(nextis('[')){
283 *w++='[';
284 c = advance();
285 *w++=c;
286 if(c<'0' || '9'<c){
287 RedirErr:
288 *w = 0;
289 yyerror(t->type==PIPE?"pipe syntax"
290 :"redirection syntax");
291 return EOF;
292 }
293 t->fd0 = 0;
294 do{
295 t->fd0 = t->fd0*10+c-'0';
296 *w++=c;
297 c = advance();
298 }while('0'<=c && c<='9');
299 if(c=='='){
300 *w++='=';
301 if(t->type==REDIR)
302 t->type = DUP;
303 c = advance();
304 if('0'<=c && c<='9'){
305 t->rtype = DUPFD;
306 t->fd1 = t->fd0;
307 t->fd0 = 0;
308 do{
309 t->fd0 = t->fd0*10+c-'0';
310 *w++=c;
311 c = advance();
312 }while('0'<=c && c<='9');
313 }
314 else{
315 if(t->type==PIPE)
316 goto RedirErr;
317 t->rtype = CLOSE;
318 }
319 }
320 if(c!=']'
321 || t->type==DUP && (t->rtype==HERE || t->rtype==APPEND))
322 goto RedirErr;
323 *w++=']';
324 }
325 *w='\0';
326 yylval.tree = t;
327 if(t->type==PIPE)
328 skipnl();
329 return t->type;
330 case '\'':
331 lastdol = 0;
332 lastword = 1;
333 inquote = 1;
334 for(;;){
335 c = advance();
336 if(c==EOF)
337 break;
338 if(c=='\''){
339 if(nextc()!='\'')
340 break;
341 advance();
342 }
343 w = addutf(w, c);
344 }
345 if(w!=0)
346 *w='\0';
347 t = token(tok, WORD);
348 t->quoted = 1;
349 yylval.tree = t;
350 return t->type;
351 }
352 if(!wordchr(c)){
353 lastdol = 0;
354 tok[0] = c;
355 tok[1]='\0';
356 return c;
357 }
358 for(;;){
359 /* next line should have (char)c==GLOB, but ken's compiler is broken */
360 if(c=='*' || c=='[' || c=='?' || c==(unsigned char)GLOB)
361 w = addtok(w, GLOB);
362 w = addutf(w, c);
363 c = nextc();
364 if(lastdol?!idchr(c):!wordchr(c)) break;
365 advance();
366 }
367
368 lastword = 1;
369 lastdol = 0;
370 if(w!=0)
371 *w='\0';
372 t = klook(tok);
373 if(t->type!=WORD)
374 lastword = 0;
375 t->quoted = 0;
376 yylval.tree = t;
377 return t->type;
378 }