The R Project SVN R

Rev

Rev 88273 | Show entire file | Ignore whitespace | Details | Blame | Last modification | View Log | RSS feed

Rev 88273 Rev 90200
Line 29... Line 29...
29 
#include <tre/tre.h>
 29 
#include <tre/tre.h>
30 
 
 30 
 
31 
static SEXP allocMatrixNA(SEXPTYPE, int, int);
 31 
static SEXP allocMatrixNA(SEXPTYPE, int, int);
32 
static void transferVector(SEXP s, SEXP t);
 32 
static void transferVector(SEXP s, SEXP t);
33 
 
 33 
 
- 
 
 34 
/* Build a CHARSXP marked as UTF-8 from the NUL-terminated string 's'.
- 
 
 35 
   DCF files are required to be UTF-8, so 's' is interpreted as UTF-8
- 
 
 36 
   regardless of its actual encoding; any invalid byte sequences are
- 
 
 37 
   repaired by escaping them as "<xx>", exactly as
- 
 
 38 
   iconv(from = "UTF-8", to = "UTF-8", sub = "byte") does (which the R
- 
 
 39 
   code paths in read.dcf()/write.dcf() also use).  The common case of
- 
 
 40 
   already-valid input is handled without conversion. */
- 
 
 41 
static SEXP mkCharUTF8sub(const char *s)
- 
 
 42 
{
- 
 
 43 
    if (utf8Valid(s))
- 
 
 44 
	return mkCharCE(s, CE_UTF8);
- 
 
 45 
 
- 
 
 46 
    /* reEnc3() performs the iconv() repair via Riconv(); subst = 1 selects
- 
 
 47 
       the "<xx>" hexadecimal substitution for invalid bytes.  It returns a
- 
 
 48 
       string allocated with R_alloc() (and freed at the vmaxset() in
- 
 
 49 
       do_readDCF()), or 's' itself if iconv is unavailable.  Either way
- 
 
 50 
       mkCharCE() copies the bytes into the CHARSXP below. */
- 
 
 51 
    const char *repaired = reEnc3(s, "UTF-8", "UTF-8", 1);
- 
 
 52 
    return mkCharCE(repaired, CE_UTF8);
- 
 
 53 
}
- 
 
 54 
 
34 
static void con_cleanup(void *data)
 55 
static void con_cleanup(void *data)
35 
{
 56 
{
36 
    Rconnection con = data;
 57 
    Rconnection con = data;
37 
    if(con->isopen) con->close(con);
 58 
    if(con->isopen) con->close(con);
38 
}
 59 
}
Line 201... Line 222...
201 
			    }
 222 
			    }
202 
			    n_eblanklines = 0;
 223 
			    n_eblanklines = 0;
203 
			}
 224 
			}
204 
			strcat(buf, line + offset);
 225 
			strcat(buf, line + offset);
205 
		    }
 226 
		    }
206 
		    SET_STRING_ELT(retval, lastm + nwhat * k, mkChar(buf));
 227 
		    SET_STRING_ELT(retval, lastm + nwhat * k, mkCharUTF8sub(buf));
207 
		}
 228 
		}
208 
	    } else {
 229 
	    } else {
209 
		if(tre_regexecb(&regline, line, 1, regmatch, 0) == 0) {
 230 
		if(tre_regexecb(&regline, line, 1, regmatch, 0) == 0) {
210 
		    for(m = 0; m < nwhat; m++){
 231 
		    for(m = 0; m < nwhat; m++){
211 
			whatlen = (int) strlen(CHAR(STRING_ELT(what, m)));
 232 
			whatlen = (int) strlen(CHAR(STRING_ELT(what, m)));
Line 228... Line 249...
228 
				if((tre_regexecb(&trailblank, line, 1,
 249 
				if((tre_regexecb(&trailblank, line, 1,
229 
						 regmatch, 0) == 0))
 250 
						 regmatch, 0) == 0))
230 
				    line[regmatch[0].rm_so] = '\0';
 251 
				    line[regmatch[0].rm_so] = '\0';
231 
			    }
 252 
			    }
232 
			    SET_STRING_ELT(retval, m + nwhat * k,
 253 
			    SET_STRING_ELT(retval, m + nwhat * k,
233 
					   mkChar(line + offset));
 254 
					   mkCharUTF8sub(line + offset));
234 
			    break;
 255 
			    break;
235 
			} else {
 256 
			} else {
236 
			    /* This is a field, but not one prespecified */
 257 
			    /* This is a field, but not one prespecified */
237 
			    lastm = -1;
 258 
			    lastm = -1;
238 
			    field_skip = true;
 259 
			    field_skip = true;
Line 272... Line 293...
272 
			    } else buf = tmp;
 293 
			    } else buf = tmp;
273 
			    buflen = need;
 294 
			    buflen = need;
274 
			}
 295 
			}
275 
			strncpy(buf, line, Rf_strchr(line, ':') - line);
 296 
			strncpy(buf, line, Rf_strchr(line, ':') - line);
276 
			buf[Rf_strchr(line, ':') - line] = '\0';
 297 
			buf[Rf_strchr(line, ':') - line] = '\0';
277 
			SET_STRING_ELT(what, nwhat, mkChar(buf));
 298 
			SET_STRING_ELT(what, nwhat, mkCharUTF8sub(buf));
278 
			nwhat++;
 299 
			nwhat++;
279 
			/* lastm uses C indexing, hence nwhat - 1 */
 300 
			/* lastm uses C indexing, hence nwhat - 1 */
280 
			lastm = nwhat - 1;
 301 
			lastm = nwhat - 1;
281 
			field_name = CHAR(STRING_ELT(what, lastm));
 302 
			field_name = CHAR(STRING_ELT(what, lastm));
282 
			if(has_fold_excludes) {
 303 
			if(has_fold_excludes) {
Line 290... Line 311...
290 
			    if((tre_regexecb(&trailblank, line, 1,
 311 
			    if((tre_regexecb(&trailblank, line, 1,
291 
					     regmatch, 0) == 0))
 312 
					     regmatch, 0) == 0))
292 
				line[regmatch[0].rm_so] = '\0';
 313 
				line[regmatch[0].rm_so] = '\0';
293 
			}
 314 
			}
294 
			SET_STRING_ELT(retval, lastm + nwhat * k,
 315 
			SET_STRING_ELT(retval, lastm + nwhat * k,
295 
				       mkChar(line + offset));
 316 
				       mkCharUTF8sub(line + offset));
296 
		    }
 317 
		    }
297 
		} else {
 318 
		} else {
298 
		    /* Must be a regular line with no tag ... */
 319 
		    /* Must be a regular line with no tag ... */
299 
		    line[20] = '\0';
 320 
		    line[20] = '\0';
300 
		    error(_("Line starting '%s ...' is malformed!"), line);
 321 
		    error(_("Line starting '%s ...' is malformed!"), line);