📄 pcremodule.c
字号:
/* Pcre objects */
#include "Python.h"
#ifndef Py_eval_input
/* For Python 1.4, graminit.h has to be explicitly included */
#include "graminit.h"
#define Py_eval_input eval_input
#endif
#ifndef FOR_PYTHON
#define FOR_PYTHON
#endif
#include "pcre.h"
#include "pcre-int.h"
static PyObject *ErrorObject;
typedef struct {
PyObject_HEAD
pcre *regex;
pcre_extra *regex_extra;
int num_groups;
} PcreObject;
staticforward PyTypeObject Pcre_Type;
#define PcreObject_Check(v) ((v)->ob_type == &Pcre_Type)
#define NORMAL 0
#define CHARCLASS 1
#define REPLACEMENT 2
#define CHAR 0
#define MEMORY_REFERENCE 1
#define SYNTAX 2
#define NOT_SYNTAX 3
#define SET 4
#define WORD_BOUNDARY 5
#define NOT_WORD_BOUNDARY 6
#define BEGINNING_OF_BUFFER 7
#define END_OF_BUFFER 8
#define STRING 9
static PcreObject *
newPcreObject(PyObject *args)
{
PcreObject *self;
self = PyObject_New(PcreObject, &Pcre_Type);
if (self == NULL)
return NULL;
self->regex = NULL;
self->regex_extra = NULL;
return self;
}
/* Pcre methods */
static void
PyPcre_dealloc(PcreObject *self)
{
if (self->regex) (pcre_free)(self->regex);
if (self->regex_extra) (pcre_free)(self->regex_extra);
PyObject_Del(self);
}
static PyObject *
PyPcre_exec(PcreObject *self, PyObject *args)
{
char *string;
int stringlen, pos = 0, options=0, endpos = -1, i, count;
int offsets[100*2];
PyObject *list;
if (!PyArg_ParseTuple(args, "t#|iii:match", &string, &stringlen,
&pos, &endpos, &options))
return NULL;
if (endpos == -1) {endpos = stringlen;}
count = pcre_exec(self->regex, self->regex_extra,
string, endpos, pos, options,
offsets, sizeof(offsets)/sizeof(int) );
/* If an error occurred during the match, and an exception was raised,
just return NULL and leave the exception alone. The most likely
problem to cause this would be running out of memory for
the failure stack. */
if (PyErr_Occurred())
{
return NULL;
}
if (count==PCRE_ERROR_NOMATCH) {Py_INCREF(Py_None); return Py_None;}
if (count<0)
{
PyObject *errval = Py_BuildValue("si", "Regex execution error", count);
PyErr_SetObject(ErrorObject, errval);
Py_XDECREF(errval);
return NULL;
}
list=PyList_New(self->num_groups+1);
if (list==NULL) return NULL;
for(i=0; i<=self->num_groups; i++)
{
PyObject *v;
int start=offsets[i*2], end=offsets[i*2+1];
/* If the group wasn't affected by the match, return -1, -1 */
if (start<0 || count<=i)
{start=end=-1;}
v=Py_BuildValue("ii", start, end);
if (v==NULL) {Py_DECREF(list); return NULL;}
PyList_SetItem(list, i, v);
}
return list;
}
static PyMethodDef Pcre_methods[] = {
{"match", (PyCFunction)PyPcre_exec, 1},
{NULL, NULL} /* sentinel */
};
static PyObject *
PyPcre_getattr(PcreObject *self, char *name)
{
return Py_FindMethod(Pcre_methods, (PyObject *)self, name);
}
staticforward PyTypeObject Pcre_Type = {
PyObject_HEAD_INIT(NULL)
0, /*ob_size*/
"pcre.Pcre", /*tp_name*/
sizeof(PcreObject), /*tp_basicsize*/
0, /*tp_itemsize*/
/* methods */
(destructor)PyPcre_dealloc, /*tp_dealloc*/
0, /*tp_print*/
(getattrfunc)PyPcre_getattr, /*tp_getattr*/
0, /*tp_setattr*/
0, /*tp_compare*/
0, /*tp_repr*/
0, /*tp_as_number*/
0, /*tp_as_sequence*/
0, /*tp_as_mapping*/
0, /*tp_hash*/
};
/* --------------------------------------------------------------------- */
static PyObject *
PyPcre_compile(PyObject *self, PyObject *args)
{
PcreObject *rv;
PyObject *dictionary;
char *pattern;
const char *error;
int options, erroroffset;
if (!PyArg_ParseTuple(args, "siO!:pcre_compile", &pattern, &options,
&PyDict_Type, &dictionary))
return NULL;
rv = newPcreObject(args);
if ( rv == NULL )
return NULL;
rv->regex = pcre_compile((char*)pattern, options,
&error, &erroroffset, dictionary);
if (rv->regex==NULL)
{
Py_DECREF(rv);
if (!PyErr_Occurred())
{
PyObject *errval = Py_BuildValue("si", error, erroroffset);
PyErr_SetObject(ErrorObject, errval);
Py_XDECREF(errval);
}
return NULL;
}
rv->regex_extra=pcre_study(rv->regex, 0, &error);
if (rv->regex_extra==NULL && error!=NULL)
{
PyObject *errval = Py_BuildValue("si", error, 0);
Py_DECREF(rv);
PyErr_SetObject(ErrorObject, errval);
Py_XDECREF(errval);
return NULL;
}
rv->num_groups = pcre_info(rv->regex, NULL, NULL);
if (rv->num_groups<0)
{
PyObject *errval = Py_BuildValue("si", error, rv->num_groups);
PyErr_SetObject(ErrorObject, errval);
Py_XDECREF(errval);
Py_DECREF(rv);
return NULL;
}
return (PyObject *)rv;
}
static PyObject *
PyPcre_expand_escape(unsigned char *pattern, int pattern_len,
int *indexptr, int *typeptr)
{
unsigned char c;
int index = *indexptr;
if (pattern_len<=index)
{
PyErr_SetString(ErrorObject, "escape ends too soon");
return NULL;
}
c=pattern[index]; index++;
*typeptr=CHAR;
switch (c)
{
case('t'):
*indexptr=index;
return Py_BuildValue("c", (char)9);
case('n'):
*indexptr = index;
return Py_BuildValue("c", (char)10);
case('v'):
*indexptr = index;
return Py_BuildValue("c", (char)11);
case('r'):
*indexptr = index;
return Py_BuildValue("c", (char)13);
case('f'):
*indexptr = index;
return Py_BuildValue("c", (char)12);
case('a'):
*indexptr = index;
return Py_BuildValue("c", (char)7);
case('b'):
*indexptr=index;
return Py_BuildValue("c", (char)8);
case('\\'):
*indexptr=index;
return Py_BuildValue("c", '\\');
case('x'):
{
int x, ch, end;
x = 0; end = index;
while ( (end<pattern_len && pcre_ctypes[ pattern[end] ] & ctype_xdigit) != 0)
{
ch = pattern[end];
x = x * 16 + pcre_lcc[ch] -
(((pcre_ctypes[ch] & ctype_digit) != 0)? '0' : 'W');
x &= 255;
end++;
}
if (end==index)
{
PyErr_SetString(ErrorObject, "\\x must be followed by hex digits");
return NULL;
}
*indexptr = end;
return Py_BuildValue("c", (char)x);
}
case('E'): case('G'): case('L'): case('Q'):
case('U'): case('l'): case('u'):
{
char message[50];
PyOS_snprintf(message, sizeof(message),
"\\%c is not allowed", c);
PyErr_SetString(ErrorObject, message);
return NULL;
}
case('g'):
{
int end, i;
int group_num = 0, is_number=0;
if (pattern_len<=index)
{
PyErr_SetString(ErrorObject, "unfinished symbolic reference");
return NULL;
}
if (pattern[index]!='<')
{
PyErr_SetString(ErrorObject, "missing < in symbolic reference");
return NULL;
}
index++;
end=index;
while (end<pattern_len && pattern[end]!='>')
end++;
if (end==pattern_len)
{
PyErr_SetString(ErrorObject, "unfinished symbolic reference");
return NULL;
}
if (index==end) /* Zero-length name */
{
/* XXX should include the text of the reference */
PyErr_SetString(ErrorObject, "zero-length symbolic reference");
return NULL;
}
if ((pcre_ctypes[pattern[index]] & ctype_digit)) /* First char. a digit */
{
is_number = 1;
group_num = pattern[index] - '0';
}
for(i=index+1; i<end; i++)
{
if (is_number &&
!(pcre_ctypes[pattern[i]] & ctype_digit) )
{
/* XXX should include the text of the reference */
PyErr_SetString(ErrorObject, "illegal non-digit character in \\g<...> starting with digit");
return NULL;
}
else {group_num = group_num * 10 + pattern[i] - '0';}
if (!(pcre_ctypes[pattern[i]] & ctype_word) )
⌨️ 快捷键说明
复制代码
Ctrl + C
搜索代码
Ctrl + F
全屏模式
F11
切换主题
Ctrl + Shift + D
显示快捷键
?
增大字号
Ctrl + =
减小字号
Ctrl + -