| 1 |
|
|---|
| 2 | /* Parser-tokenizer link implementation */
|
|---|
| 3 |
|
|---|
| 4 | #include "pgenheaders.h"
|
|---|
| 5 | #include "tokenizer.h"
|
|---|
| 6 | #include "node.h"
|
|---|
| 7 | #include "grammar.h"
|
|---|
| 8 | #include "parser.h"
|
|---|
| 9 | #include "parsetok.h"
|
|---|
| 10 | #include "errcode.h"
|
|---|
| 11 | #include "graminit.h"
|
|---|
| 12 |
|
|---|
| 13 | int Py_TabcheckFlag;
|
|---|
| 14 |
|
|---|
| 15 |
|
|---|
| 16 | /* Forward */
|
|---|
| 17 | static node *parsetok(struct tok_state *, grammar *, int, perrdetail *, int);
|
|---|
| 18 | static void initerr(perrdetail *err_ret, const char* filename);
|
|---|
| 19 |
|
|---|
| 20 | /* Parse input coming from a string. Return error code, print some errors. */
|
|---|
| 21 | node *
|
|---|
| 22 | PyParser_ParseString(const char *s, grammar *g, int start, perrdetail *err_ret)
|
|---|
| 23 | {
|
|---|
| 24 | return PyParser_ParseStringFlagsFilename(s, NULL, g, start, err_ret, 0);
|
|---|
| 25 | }
|
|---|
| 26 |
|
|---|
| 27 | node *
|
|---|
| 28 | PyParser_ParseStringFlags(const char *s, grammar *g, int start,
|
|---|
| 29 | perrdetail *err_ret, int flags)
|
|---|
| 30 | {
|
|---|
| 31 | return PyParser_ParseStringFlagsFilename(s, NULL,
|
|---|
| 32 | g, start, err_ret, flags);
|
|---|
| 33 | }
|
|---|
| 34 |
|
|---|
| 35 | node *
|
|---|
| 36 | PyParser_ParseStringFlagsFilename(const char *s, const char *filename,
|
|---|
| 37 | grammar *g, int start,
|
|---|
| 38 | perrdetail *err_ret, int flags)
|
|---|
| 39 | {
|
|---|
| 40 | struct tok_state *tok;
|
|---|
| 41 |
|
|---|
| 42 | initerr(err_ret, filename);
|
|---|
| 43 |
|
|---|
| 44 | if ((tok = PyTokenizer_FromString(s)) == NULL) {
|
|---|
| 45 | err_ret->error = PyErr_Occurred() ? E_DECODE : E_NOMEM;
|
|---|
| 46 | return NULL;
|
|---|
| 47 | }
|
|---|
| 48 |
|
|---|
| 49 | tok->filename = filename ? filename : "<string>";
|
|---|
| 50 | if (Py_TabcheckFlag || Py_VerboseFlag) {
|
|---|
| 51 | tok->altwarning = (tok->filename != NULL);
|
|---|
| 52 | if (Py_TabcheckFlag >= 2)
|
|---|
| 53 | tok->alterror++;
|
|---|
| 54 | }
|
|---|
| 55 |
|
|---|
| 56 | return parsetok(tok, g, start, err_ret, flags);
|
|---|
| 57 | }
|
|---|
| 58 |
|
|---|
| 59 | /* Parse input coming from a file. Return error code, print some errors. */
|
|---|
| 60 |
|
|---|
| 61 | node *
|
|---|
| 62 | PyParser_ParseFile(FILE *fp, const char *filename, grammar *g, int start,
|
|---|
| 63 | char *ps1, char *ps2, perrdetail *err_ret)
|
|---|
| 64 | {
|
|---|
| 65 | return PyParser_ParseFileFlags(fp, filename, g, start, ps1, ps2,
|
|---|
| 66 | err_ret, 0);
|
|---|
| 67 | }
|
|---|
| 68 |
|
|---|
| 69 | node *
|
|---|
| 70 | PyParser_ParseFileFlags(FILE *fp, const char *filename, grammar *g, int start,
|
|---|
| 71 | char *ps1, char *ps2, perrdetail *err_ret, int flags)
|
|---|
| 72 | {
|
|---|
| 73 | struct tok_state *tok;
|
|---|
| 74 |
|
|---|
| 75 | initerr(err_ret, filename);
|
|---|
| 76 |
|
|---|
| 77 | if ((tok = PyTokenizer_FromFile(fp, ps1, ps2)) == NULL) {
|
|---|
| 78 | err_ret->error = E_NOMEM;
|
|---|
| 79 | return NULL;
|
|---|
| 80 | }
|
|---|
| 81 | tok->filename = filename;
|
|---|
| 82 | if (Py_TabcheckFlag || Py_VerboseFlag) {
|
|---|
| 83 | tok->altwarning = (filename != NULL);
|
|---|
| 84 | if (Py_TabcheckFlag >= 2)
|
|---|
| 85 | tok->alterror++;
|
|---|
| 86 | }
|
|---|
| 87 |
|
|---|
| 88 |
|
|---|
| 89 | return parsetok(tok, g, start, err_ret, flags);
|
|---|
| 90 | }
|
|---|
| 91 |
|
|---|
| 92 | /* Parse input coming from the given tokenizer structure.
|
|---|
| 93 | Return error code. */
|
|---|
| 94 |
|
|---|
| 95 | static char with_msg[] =
|
|---|
| 96 | "%s:%d: Warning: 'with' will become a reserved keyword in Python 2.6\n";
|
|---|
| 97 |
|
|---|
| 98 | static char as_msg[] =
|
|---|
| 99 | "%s:%d: Warning: 'as' will become a reserved keyword in Python 2.6\n";
|
|---|
| 100 |
|
|---|
| 101 | static void
|
|---|
| 102 | warn(const char *msg, const char *filename, int lineno)
|
|---|
| 103 | {
|
|---|
| 104 | if (filename == NULL)
|
|---|
| 105 | filename = "<string>";
|
|---|
| 106 | PySys_WriteStderr(msg, filename, lineno);
|
|---|
| 107 | }
|
|---|
| 108 |
|
|---|
| 109 | static node *
|
|---|
| 110 | parsetok(struct tok_state *tok, grammar *g, int start, perrdetail *err_ret,
|
|---|
| 111 | int flags)
|
|---|
| 112 | {
|
|---|
| 113 | parser_state *ps;
|
|---|
| 114 | node *n;
|
|---|
| 115 | int started = 0, handling_import = 0, handling_with = 0;
|
|---|
| 116 |
|
|---|
| 117 | if ((ps = PyParser_New(g, start)) == NULL) {
|
|---|
| 118 | fprintf(stderr, "no mem for new parser\n");
|
|---|
| 119 | err_ret->error = E_NOMEM;
|
|---|
| 120 | PyTokenizer_Free(tok);
|
|---|
| 121 | return NULL;
|
|---|
| 122 | }
|
|---|
| 123 | #ifdef PY_PARSER_REQUIRES_FUTURE_KEYWORD
|
|---|
| 124 | if (flags & PyPARSE_WITH_IS_KEYWORD)
|
|---|
| 125 | ps->p_flags |= CO_FUTURE_WITH_STATEMENT;
|
|---|
| 126 | #endif
|
|---|
| 127 |
|
|---|
| 128 | for (;;) {
|
|---|
| 129 | char *a, *b;
|
|---|
| 130 | int type;
|
|---|
| 131 | size_t len;
|
|---|
| 132 | char *str;
|
|---|
| 133 | int col_offset;
|
|---|
| 134 |
|
|---|
| 135 | type = PyTokenizer_Get(tok, &a, &b);
|
|---|
| 136 | if (type == ERRORTOKEN) {
|
|---|
| 137 | err_ret->error = tok->done;
|
|---|
| 138 | break;
|
|---|
| 139 | }
|
|---|
| 140 | if (type == ENDMARKER && started) {
|
|---|
| 141 | type = NEWLINE; /* Add an extra newline */
|
|---|
| 142 | handling_with = handling_import = 0;
|
|---|
| 143 | started = 0;
|
|---|
| 144 | /* Add the right number of dedent tokens,
|
|---|
| 145 | except if a certain flag is given --
|
|---|
| 146 | codeop.py uses this. */
|
|---|
| 147 | if (tok->indent &&
|
|---|
| 148 | !(flags & PyPARSE_DONT_IMPLY_DEDENT))
|
|---|
| 149 | {
|
|---|
| 150 | tok->pendin = -tok->indent;
|
|---|
| 151 | tok->indent = 0;
|
|---|
| 152 | }
|
|---|
| 153 | }
|
|---|
| 154 | else
|
|---|
| 155 | started = 1;
|
|---|
| 156 | len = b - a; /* XXX this may compute NULL - NULL */
|
|---|
| 157 | str = (char *) PyObject_MALLOC(len + 1);
|
|---|
| 158 | if (str == NULL) {
|
|---|
| 159 | fprintf(stderr, "no mem for next token\n");
|
|---|
| 160 | err_ret->error = E_NOMEM;
|
|---|
| 161 | break;
|
|---|
| 162 | }
|
|---|
| 163 | if (len > 0)
|
|---|
| 164 | strncpy(str, a, len);
|
|---|
| 165 | str[len] = '\0';
|
|---|
| 166 |
|
|---|
| 167 | #ifdef PY_PARSER_REQUIRES_FUTURE_KEYWORD
|
|---|
| 168 | /* This is only necessary to support the "as" warning, but
|
|---|
| 169 | we don't want to warn about "as" in import statements. */
|
|---|
| 170 | if (type == NAME &&
|
|---|
| 171 | len == 6 && str[0] == 'i' && strcmp(str, "import") == 0)
|
|---|
| 172 | handling_import = 1;
|
|---|
| 173 |
|
|---|
| 174 | /* Warn about with as NAME */
|
|---|
| 175 | if (type == NAME &&
|
|---|
| 176 | !(ps->p_flags & CO_FUTURE_WITH_STATEMENT)) {
|
|---|
| 177 | if (len == 4 && str[0] == 'w' && strcmp(str, "with") == 0)
|
|---|
| 178 | warn(with_msg, err_ret->filename, tok->lineno);
|
|---|
| 179 | else if (!(handling_import || handling_with) &&
|
|---|
| 180 | len == 2 && str[0] == 'a' &&
|
|---|
| 181 | strcmp(str, "as") == 0)
|
|---|
| 182 | warn(as_msg, err_ret->filename, tok->lineno);
|
|---|
| 183 | }
|
|---|
| 184 | else if (type == NAME &&
|
|---|
| 185 | (ps->p_flags & CO_FUTURE_WITH_STATEMENT) &&
|
|---|
| 186 | len == 4 && str[0] == 'w' && strcmp(str, "with") == 0)
|
|---|
| 187 | handling_with = 1;
|
|---|
| 188 | #endif
|
|---|
| 189 | if (a >= tok->line_start)
|
|---|
| 190 | col_offset = a - tok->line_start;
|
|---|
| 191 | else
|
|---|
| 192 | col_offset = -1;
|
|---|
| 193 |
|
|---|
| 194 | if ((err_ret->error =
|
|---|
| 195 | PyParser_AddToken(ps, (int)type, str, tok->lineno, col_offset,
|
|---|
| 196 | &(err_ret->expected))) != E_OK) {
|
|---|
| 197 | if (err_ret->error != E_DONE) {
|
|---|
| 198 | PyObject_FREE(str);
|
|---|
| 199 | err_ret->token = type;
|
|---|
| 200 | }
|
|---|
| 201 | break;
|
|---|
| 202 | }
|
|---|
| 203 | }
|
|---|
| 204 |
|
|---|
| 205 | if (err_ret->error == E_DONE) {
|
|---|
| 206 | n = ps->p_tree;
|
|---|
| 207 | ps->p_tree = NULL;
|
|---|
| 208 | }
|
|---|
| 209 | else
|
|---|
| 210 | n = NULL;
|
|---|
| 211 |
|
|---|
| 212 | PyParser_Delete(ps);
|
|---|
| 213 |
|
|---|
| 214 | if (n == NULL) {
|
|---|
| 215 | if (tok->lineno <= 1 && tok->done == E_EOF)
|
|---|
| 216 | err_ret->error = E_EOF;
|
|---|
| 217 | err_ret->lineno = tok->lineno;
|
|---|
| 218 | if (tok->buf != NULL) {
|
|---|
| 219 | size_t len;
|
|---|
| 220 | assert(tok->cur - tok->buf < INT_MAX);
|
|---|
| 221 | err_ret->offset = (int)(tok->cur - tok->buf);
|
|---|
| 222 | len = tok->inp - tok->buf;
|
|---|
| 223 | err_ret->text = (char *) PyObject_MALLOC(len + 1);
|
|---|
| 224 | if (err_ret->text != NULL) {
|
|---|
| 225 | if (len > 0)
|
|---|
| 226 | strncpy(err_ret->text, tok->buf, len);
|
|---|
| 227 | err_ret->text[len] = '\0';
|
|---|
| 228 | }
|
|---|
| 229 | }
|
|---|
| 230 | } else if (tok->encoding != NULL) {
|
|---|
| 231 | node* r = PyNode_New(encoding_decl);
|
|---|
| 232 | if (!r) {
|
|---|
| 233 | err_ret->error = E_NOMEM;
|
|---|
| 234 | n = NULL;
|
|---|
| 235 | goto done;
|
|---|
| 236 | }
|
|---|
| 237 | r->n_str = tok->encoding;
|
|---|
| 238 | r->n_nchildren = 1;
|
|---|
| 239 | r->n_child = n;
|
|---|
| 240 | tok->encoding = NULL;
|
|---|
| 241 | n = r;
|
|---|
| 242 | }
|
|---|
| 243 |
|
|---|
| 244 | done:
|
|---|
| 245 | PyTokenizer_Free(tok);
|
|---|
| 246 |
|
|---|
| 247 | return n;
|
|---|
| 248 | }
|
|---|
| 249 |
|
|---|
| 250 | static void
|
|---|
| 251 | initerr(perrdetail *err_ret, const char *filename)
|
|---|
| 252 | {
|
|---|
| 253 | err_ret->error = E_OK;
|
|---|
| 254 | err_ret->filename = filename;
|
|---|
| 255 | err_ret->lineno = 0;
|
|---|
| 256 | err_ret->offset = 0;
|
|---|
| 257 | err_ret->text = NULL;
|
|---|
| 258 | err_ret->token = -1;
|
|---|
| 259 | err_ret->expected = -1;
|
|---|
| 260 | }
|
|---|