The R Project SVN R

Rev

Rev 86821 | Show entire file | Ignore whitespace | Details | Blame | Last modification | View Log | RSS feed

Rev 86821 Rev 87891
Line 1... Line 1...
1 
/*
 1 
/*
2 
 *  R : A Computer Language for Statistical Data Analysis
 2 
 *  R : A Computer Language for Statistical Data Analysis
3 
 *  Copyright (C) 2001--2021  The R Core Team
 3 
 *  Copyright (C) 2001--2025  The R Core Team
4 
 *  Copyright (C) 1995, 1996  Robert Gentleman and Ross Ihaka
 4 
 *  Copyright (C) 1995, 1996  Robert Gentleman and Ross Ihaka
5 
 *
 5 
 *
6 
 *  This program is free software; you can redistribute it and/or modify
 6 
 *  This program is free software; you can redistribute it and/or modify
7 
 *  it under the terms of the GNU General Public License as published by
 7 
 *  it under the terms of the GNU General Public License as published by
8 
 *  the Free Software Foundation; either version 2 of the License, or
 8 
 *  the Free Software Foundation; either version 2 of the License, or
Line 174... Line 174...
174 
    UNPROTECT(1);
 174 
    UNPROTECT(1);
175 
}
 175 
}
176 
 
 176 
 
177 
typedef struct parse_info {
 177 
typedef struct parse_info {
178 
    Rconnection con;
 178 
    Rconnection con;
179 
    Rboolean old_latin1;
 179 
    bool old_latin1;
180 
    Rboolean old_utf8;
 180 
    bool old_utf8;
181 
}  parse_cleanup_info;
 181 
}  parse_cleanup_info;
182 
 
 182 
 
183 
static void parse_cleanup(void *data)
 183 
static void parse_cleanup(void *data)
184 
{
 184 
{
185 
    parse_cleanup_info *pci = (parse_cleanup_info *)data;
 185 
    parse_cleanup_info *pci = (parse_cleanup_info *)data;
Line 205... Line 205...
205 
    R_ParseError = 0;
 205 
    R_ParseError = 0;
206 
    R_ParseErrorMsg[0] = '\0';
 206 
    R_ParseErrorMsg[0] = '\0';
207 
 
 207 
 
208 
    int ifile = asInteger(CAR(args));                   args = CDR(args);
 208 
    int ifile = asInteger(CAR(args));                   args = CDR(args);
209 
    Rconnection con = getConnection(ifile);
 209 
    Rconnection con = getConnection(ifile);
210 
    Rboolean wasopen = con->isopen;
 210 
    bool wasopen = con->isopen;
211 
    int num = asInteger(CAR(args));			args = CDR(args);
 211 
    int num = asInteger(CAR(args));			args = CDR(args);
212 
    if (num == 0)
 212 
    if (num == 0)
213 
	return(allocVector(EXPRSXP, 0));
 213 
	return(allocVector(EXPRSXP, 0));
214 
 
 214 
 
215 
    SEXP text = PROTECT(coerceVector(CAR(args), STRSXP));
 215 
    SEXP text = PROTECT(coerceVector(CAR(args), STRSXP));
Line 232... Line 232...
232 
    begincontext(&cntxt, CTXT_CCODE, R_NilValue, R_BaseEnv, R_BaseEnv,
 232 
    begincontext(&cntxt, CTXT_CCODE, R_NilValue, R_BaseEnv, R_BaseEnv,
233 
		 R_NilValue, R_NilValue);
 233 
		 R_NilValue, R_NilValue);
234 
    cntxt.cend = &parse_cleanup;
 234 
    cntxt.cend = &parse_cleanup;
235 
    cntxt.cenddata = &pci;
 235 
    cntxt.cenddata = &pci;
236 
 
 236 
 
237 
    known_to_be_latin1 = known_to_be_utf8 = FALSE;
 237 
    known_to_be_latin1 = known_to_be_utf8 = false;
238 
    Rboolean allKnown = TRUE;
 238 
    bool allKnown = true;
239 
    /* allow 'encoding' to override declaration on 'text'. */
 239 
    /* allow 'encoding' to override declaration on 'text'. */
240 
    if(streql(encoding, "latin1")) {
 240 
    if(streql(encoding, "latin1")) {
241 
	if (!mbcslocale) {
 241 
	if (!mbcslocale) {
242 
	    known_to_be_latin1 = TRUE;
 242 
	    known_to_be_latin1 = true;
243 
	    allKnown = FALSE;
 243 
	    allKnown = false;
244 
	} else
 244 
	} else
245 
	    warning(_("argument encoding=\"latin1\" is ignored in MBCS locales"));
 245 
	    warning(_("argument encoding=\"latin1\" is ignored in MBCS locales"));
246 
    } else if(streql(encoding, "UTF-8"))  {
 246 
    } else if(streql(encoding, "UTF-8"))  {
247 
	if (!mbcslocale || utf8locale) {
 247 
	if (!mbcslocale || utf8locale) {
248 
	    known_to_be_utf8 = TRUE;
 248 
	    known_to_be_utf8 = true;
249 
	    allKnown = FALSE;
 249 
	    allKnown = false;
250 
	} else
 250 
	} else
251 
	    /* the input may be invalid or not parseable when interpreted as
 251 
	    /* the input may be invalid or not parseable when interpreted as
252 
	       in different multi-byte encoding; related to PR#16819 */
 252 
	       in different multi-byte encoding; related to PR#16819 */
253 
	    warning(_("argument encoding=\"UTF-8\" is ignored in MBCS locales"));
 253 
	    warning(_("argument encoding=\"UTF-8\" is ignored in MBCS locales"));
254 
    } else if(!streql(encoding, "unknown") && !streql(encoding, "native.enc"))
 254 
    } else if(!streql(encoding, "unknown") && !streql(encoding, "native.enc"))
Line 273... Line 273...
273 
	*/
 273 
	*/
274 
	if(allKnown)
 274 
	if(allKnown)
275 
	  for(int i = 0; i < length(text); i++)
 275 
	  for(int i = 0; i < length(text); i++)
276 
	    if(!ENC_KNOWN(STRING_ELT(text, i)) &&
 276 
	    if(!ENC_KNOWN(STRING_ELT(text, i)) &&
277 
	       ! IS_ASCII(STRING_ELT(text, i))) {
 277 
	       ! IS_ASCII(STRING_ELT(text, i))) {
278 
		allKnown = FALSE;
 278 
		allKnown = false;
279 
		break;
 279 
		break;
280 
	    }
 280 
	    }
281 
	if(allKnown) {
 281 
	if(allKnown) {
282 
	    known_to_be_latin1 = pci.old_latin1;
 282 
	    known_to_be_latin1 = pci.old_latin1;
283 
	    known_to_be_utf8 = pci.old_utf8;
 283 
	    known_to_be_utf8 = pci.old_utf8;