ben
/
mwparserfromhell
mirror of https://github.com/earwig/mwparserfromhell
1
1
Fork 0
Code Releases 24 Activity
Browse Source

More reworking of CTokenizer Unicode support (incomplete)

tags/v0.4.1
Ben Kurtovic 9 years ago
parent
8b72c783f0
commit
2072a10b67
5 changed files with 160 additions and 87 deletions
Split View
Diff Options
Show Stats Download Patch File Download Diff File
  1. +17
    -3
      mwparserfromhell/parser/ctokenizer/common.h
  2. +65
    -55
      mwparserfromhell/parser/ctokenizer/tok_parse.c
  3. +18
    -6
      mwparserfromhell/parser/ctokenizer/tok_support.c
  4. +3
    -7
      mwparserfromhell/parser/ctokenizer/tok_support.h
  5. +57
    -16
      mwparserfromhell/parser/ctokenizer/tokenizer.c

+ 17
- 3
mwparserfromhell/parser/ctokenizer/common.h View File

@@ -51,8 +51,12 @@ SOFTWARE.

#ifdef PEP_393
#define Unicode Py_UCS4
#define PyUnicode_FROM_SINGLE(chr) \
PyUnicode_FromKindAndData(PyUnicode_4BYTE_KIND, &(chr), 1)
#else
#define Unicode Py_UNICODE
#define PyUnicode_FROM_SINGLE(chr) \
PyUnicode_FromUnicode(&(chr), 1)
#endif

/* Error handling macros */
@@ -77,7 +81,7 @@ extern PyObject* definitions;

typedef struct {
Py_ssize_t size;
Py_UNICODE* data;
Unicode* data;
} Textbuffer;

struct Stack {
@@ -89,11 +93,21 @@ struct Stack {
typedef struct Stack Stack;

typedef struct {
PyObject* object; /* base PyUnicodeObject object */
Py_ssize_t length; /* length of object, in code points */
#ifdef PEP_393
int kind; /* object's kind value */
void* data; /* object's raw unicode buffer */
#else
Py_UNICODE* buf; /* object's internal buffer */
#endif
} TokenizerInput;

typedef struct {
PyObject_HEAD
PyObject* text; /* text to tokenize */
TokenizerInput text; /* text to tokenize */
Stack* topstack; /* topmost stack */
Py_ssize_t head; /* current position in text */
Py_ssize_t length; /* length of text */
int global; /* global context */
int depth; /* stack recursion depth */
int cycles; /* total number of stack recursions */


+ 65
- 55
mwparserfromhell/parser/ctokenizer/tok_parse.c View File

@@ -190,7 +190,7 @@ static int Tokenizer_parse_template_or_argument(Tokenizer* self)
PyObject *tokenlist;

self->head += 2;
while (Tokenizer_READ(self, 0) == '{' && braces < MAX_BRACES) {
while (Tokenizer_read(self, 0) == '{' && braces < MAX_BRACES) {
self->head++;
braces++;
}
@@ -426,7 +426,7 @@ static int Tokenizer_parse_bracketed_uri_scheme(Tokenizer* self)

if (Tokenizer_push(self, LC_EXT_LINK_URI))
return -1;
if (Tokenizer_READ(self, 0) == '/' && Tokenizer_READ(self, 1) == '/') {
if (Tokenizer_read(self, 0) == '/' && Tokenizer_read(self, 1) == '/') {
if (Tokenizer_emit_text(self, "//"))
return -1;
self->head += 2;
@@ -435,7 +435,7 @@ static int Tokenizer_parse_bracketed_uri_scheme(Tokenizer* self)
buffer = Textbuffer_new();
if (!buffer)
return -1;
while ((this = Tokenizer_READ(self, 0))) {
while ((this = Tokenizer_read(self, 0))) {
i = 0;
while (1) {
if (!valid[i])
@@ -462,8 +462,8 @@ static int Tokenizer_parse_bracketed_uri_scheme(Tokenizer* self)
return -1;
}
self->head++;
slashes = (Tokenizer_READ(self, 0) == '/' &&
Tokenizer_READ(self, 1) == '/');
slashes = (Tokenizer_read(self, 0) == '/' &&
Tokenizer_read(self, 1) == '/');
if (slashes) {
if (Tokenizer_emit_text(self, "//")) {
Textbuffer_dealloc(buffer);
@@ -528,8 +528,8 @@ static int Tokenizer_parse_free_uri_scheme(Tokenizer* self)
Textbuffer_dealloc(scheme_buffer);
return -1;
}
slashes = (Tokenizer_READ(self, 0) == '/' &&
Tokenizer_READ(self, 1) == '/');
slashes = (Tokenizer_read(self, 0) == '/' &&
Tokenizer_read(self, 1) == '/');
if (!IS_SCHEME(scheme, slashes, 1)) {
Py_DECREF(scheme);
Textbuffer_dealloc(scheme_buffer);
@@ -589,7 +589,7 @@ static int
Tokenizer_is_free_link(Tokenizer* self, Py_UNICODE this, Py_UNICODE next)
{
// Built from Tokenizer_parse()'s end sentinels:
Py_UNICODE after = Tokenizer_READ(self, 2);
Py_UNICODE after = Tokenizer_read(self, 2);
uint64_t ctx = self->topstack->context;

return (!this || this == '\n' || this == '[' || this == ']' ||
@@ -615,22 +615,22 @@ Tokenizer_really_parse_external_link(Tokenizer* self, int brackets,
return NULL;
if (BAD_ROUTE)
return NULL;
this = Tokenizer_READ(self, 0);
this = Tokenizer_read(self, 0);
if (!this || this == '\n' || this == ' ' || this == ']')
return Tokenizer_fail_route(self);
if (!brackets && this == '[')
return Tokenizer_fail_route(self);
while (1) {
this = Tokenizer_READ(self, 0);
next = Tokenizer_READ(self, 1);
this = Tokenizer_read(self, 0);
next = Tokenizer_read(self, 1);
if (this == '&') {
PUSH_TAIL_BUFFER(*extra, NULL)
if (Tokenizer_parse_entity(self))
return NULL;
}
else if (this == '<' && next == '!'
&& Tokenizer_READ(self, 2) == '-'
&& Tokenizer_READ(self, 3) == '-') {
&& Tokenizer_read(self, 2) == '-'
&& Tokenizer_read(self, 3) == '-') {
PUSH_TAIL_BUFFER(*extra, NULL)
if (Tokenizer_parse_comment(self))
return NULL;
@@ -716,7 +716,7 @@ static int Tokenizer_parse_external_link(Tokenizer* self, int brackets)
#define NOT_A_LINK \
if (!brackets && self->topstack->context & LC_DLTERM) \
return Tokenizer_handle_dl_term(self); \
return Tokenizer_emit_char(self, Tokenizer_READ(self, 0))
return Tokenizer_emit_char(self, Tokenizer_read(self, 0))

Py_ssize_t reset = self->head;
PyObject *link, *kwargs;
@@ -787,7 +787,7 @@ static int Tokenizer_parse_heading(Tokenizer* self)

self->global |= GL_HEADING;
self->head += 1;
while (Tokenizer_READ(self, 0) == '=') {
while (Tokenizer_read(self, 0) == '=') {
best++;
self->head++;
}
@@ -862,7 +862,7 @@ static HeadingData* Tokenizer_handle_heading_end(Tokenizer* self)

self->head += 1;
best = 1;
while (Tokenizer_READ(self, 0) == '=') {
while (Tokenizer_read(self, 0) == '=') {
best++;
self->head++;
}
@@ -916,7 +916,7 @@ static HeadingData* Tokenizer_handle_heading_end(Tokenizer* self)
*/
static int Tokenizer_really_parse_entity(Tokenizer* self)
{
PyObject *kwargs, *textobj;
PyObject *kwargs, *charobj, *textobj;
Py_UNICODE this;
int numeric, hexadecimal, i, j, zeroes, test;
char *valid, *text, *buffer, *def;
@@ -930,7 +930,7 @@ static int Tokenizer_really_parse_entity(Tokenizer* self)
if (Tokenizer_emit(self, HTMLEntityStart))
return -1;
self->head++;
this = Tokenizer_READ(self, 0);
this = Tokenizer_read(self, 0);
if (!this) {
Tokenizer_fail_route(self);
return 0;
@@ -940,7 +940,7 @@ static int Tokenizer_really_parse_entity(Tokenizer* self)
if (Tokenizer_emit(self, HTMLEntityNumeric))
return -1;
self->head++;
this = Tokenizer_READ(self, 0);
this = Tokenizer_read(self, 0);
if (!this) {
Tokenizer_fail_route(self);
return 0;
@@ -950,7 +950,12 @@ static int Tokenizer_really_parse_entity(Tokenizer* self)
kwargs = PyDict_New();
if (!kwargs)
return -1;
PyDict_SetItemString(kwargs, "char", Tokenizer_read(self, 0));
if (!(charobj = PyUnicode_FROM_SINGLE(this))) {
Py_DECREF(kwargs);
return -1;
}
PyDict_SetItemString(kwargs, "char", charobj);
Py_DECREF(charobj);
if (Tokenizer_emit_kwargs(self, HTMLEntityHex, kwargs))
return -1;
self->head++;
@@ -974,7 +979,7 @@ static int Tokenizer_really_parse_entity(Tokenizer* self)
i = 0;
zeroes = 0;
while (1) {
this = Tokenizer_READ(self, 0);
this = Tokenizer_read(self, 0);
if (this == ';') {
if (i == 0)
FAIL_ROUTE_AND_EXIT()
@@ -1093,15 +1098,15 @@ static int Tokenizer_parse_comment(Tokenizer* self)
if (Tokenizer_push(self, 0))
return -1;
while (1) {
this = Tokenizer_READ(self, 0);
this = Tokenizer_read(self, 0);
if (!this) {
comment = Tokenizer_pop(self);
Py_XDECREF(comment);
self->head = reset;
return Tokenizer_emit_text(self, "<!--");
}
if (this == '-' && Tokenizer_READ(self, 1) == this &&
Tokenizer_READ(self, 2) == '>') {
if (this == '-' && Tokenizer_read(self, 1) == this &&
Tokenizer_read(self, 2) == '>') {
if (Tokenizer_emit_first(self, CommentStart))
return -1;
if (Tokenizer_emit(self, CommentEnd))
@@ -1221,7 +1226,7 @@ Tokenizer_handle_tag_space(Tokenizer* self, TagData* data, Py_UNICODE text)
*/
static int Tokenizer_handle_tag_text(Tokenizer* self, Py_UNICODE text)
{
Py_UNICODE next = Tokenizer_READ(self, 1);
Py_UNICODE next = Tokenizer_read(self, 1);

if (!is_marker(text) || !Tokenizer_CAN_RECURSE(self))
return Tokenizer_emit_char(self, text);
@@ -1291,8 +1296,8 @@ Tokenizer_handle_tag_data(Tokenizer* self, TagData* data, Py_UNICODE chunk)
}
}
else { // data->context & TAG_ATTR_VALUE assured
escaped = (Tokenizer_READ_BACKWARDS(self, 1) == '\\' &&
Tokenizer_READ_BACKWARDS(self, 2) != '\\');
escaped = (Tokenizer_read_backwards(self, 1) == '\\' &&
Tokenizer_read_backwards(self, 2) != '\\');
if (data->context & TAG_NOTE_QUOTE) {
data->context ^= TAG_NOTE_QUOTE;
if ((chunk == '"' || chunk == '\'') && !escaped) {
@@ -1419,8 +1424,8 @@ static PyObject* Tokenizer_handle_blacklisted_tag(Tokenizer* self)
int cmp;

while (1) {
this = Tokenizer_READ(self, 0);
next = Tokenizer_READ(self, 1);
this = Tokenizer_read(self, 0);
next = Tokenizer_read(self, 1);
if (!this)
return Tokenizer_fail_route(self);
else if (this == '<' && next == '/') {
@@ -1429,7 +1434,7 @@ static PyObject* Tokenizer_handle_blacklisted_tag(Tokenizer* self)
buffer = Textbuffer_new();
if (!buffer)
return NULL;
while ((this = Tokenizer_READ(self, 0)), 1) {
while ((this = Tokenizer_read(self, 0)), 1) {
if (this == '>') {
buf_tmp = Textbuffer_render(buffer);
if (!buf_tmp)
@@ -1576,8 +1581,8 @@ static PyObject* Tokenizer_really_parse_tag(Tokenizer* self)
return NULL;
}
while (1) {
this = Tokenizer_READ(self, 0);
next = Tokenizer_READ(self, 1);
this = Tokenizer_read(self, 0);
next = Tokenizer_read(self, 1);
can_exit = (!(data->context & (TAG_QUOTED | TAG_NAME)) ||
data->context & TAG_NOTE_SPACE);
if (!this) {
@@ -1652,7 +1657,7 @@ static int Tokenizer_handle_invalid_tag_start(Tokenizer* self)
if (!buf)
return -1;
while (1) {
this = Tokenizer_READ(self, pos);
this = Tokenizer_read(self, pos);
if (Py_UNICODE_ISSPACE(this) || is_marker(this)) {
name = Textbuffer_render(buf);
if (!name) {
@@ -1879,7 +1884,7 @@ static PyObject* Tokenizer_parse_style(Tokenizer* self)
uint64_t context = self->topstack->context, ticks = 2, i;

self->head += 2;
while (Tokenizer_READ(self, 0) == '\'') {
while (Tokenizer_read(self, 0) == '\'') {
self->head++;
ticks++;
}
@@ -1941,15 +1946,20 @@ static PyObject* Tokenizer_parse_style(Tokenizer* self)
*/
static int Tokenizer_handle_list_marker(Tokenizer* self)
{
PyObject *markup = Tokenizer_read(self, 0), *kwargs;
Py_UNICODE code = *PyUnicode_AS_UNICODE(markup);
PyObject *kwargs, *markup;
Py_UNICODE code = Tokenizer_read(self, 0);

if (code == ';')
self->topstack->context |= LC_DLTERM;
kwargs = PyDict_New();
if (!kwargs)
return -1;
if (!(markup = PyUnicode_FROM_SINGLE(code))) {
Py_DECREF(kwargs);
return -1;
}
PyDict_SetItemString(kwargs, "wiki_markup", markup);
Py_DECREF(markup);
if (Tokenizer_emit_kwargs(self, TagOpenOpen, kwargs))
return -1;
if (Tokenizer_emit_text(self, GET_HTML_TAG(code)))
@@ -1964,7 +1974,7 @@ static int Tokenizer_handle_list_marker(Tokenizer* self)
*/
static int Tokenizer_handle_list(Tokenizer* self)
{
Py_UNICODE marker = Tokenizer_READ(self, 1);
Py_UNICODE marker = Tokenizer_read(self, 1);

if (Tokenizer_handle_list_marker(self))
return -1;
@@ -1973,7 +1983,7 @@ static int Tokenizer_handle_list(Tokenizer* self)
self->head++;
if (Tokenizer_handle_list_marker(self))
return -1;
marker = Tokenizer_READ(self, 1);
marker = Tokenizer_read(self, 1);
}
return 0;
}
@@ -1994,7 +2004,7 @@ static int Tokenizer_handle_hr(Tokenizer* self)
if (Textbuffer_write(&buffer, '-'))
return -1;
}
while (Tokenizer_READ(self, 1) == '-') {
while (Tokenizer_read(self, 1) == '-') {
if (Textbuffer_write(&buffer, '-'))
return -1;
self->head++;
@@ -2023,7 +2033,7 @@ static int Tokenizer_handle_hr(Tokenizer* self)
static int Tokenizer_handle_dl_term(Tokenizer* self)
{
self->topstack->context ^= LC_DLTERM;
if (Tokenizer_READ(self, 0) == ':')
if (Tokenizer_read(self, 0) == ':')
return Tokenizer_handle_list_marker(self);
return Tokenizer_emit_char(self, '\n');
}
@@ -2130,7 +2140,7 @@ static PyObject* Tokenizer_handle_table_style(Tokenizer* self, char end_token)
data->context = TAG_ATTR_READY;

while (1) {
this = Tokenizer_READ(self, 0);
this = Tokenizer_read(self, 0);
can_exit = (!(data->context & TAG_QUOTED) || data->context & TAG_NOTE_SPACE);
if (this == end_token && can_exit) {
if (data->context & (TAG_ATTR_NAME | TAG_ATTR_VALUE)) {
@@ -2432,7 +2442,7 @@ Tokenizer_verify_safe(Tokenizer* self, uint64_t context, Py_UNICODE data)
} else if (data == '\n' || data == '[' || data == '}' || data == '>') {
return -1;
} else if (data == '<') {
if (Tokenizer_READ(self, 1) == '!')
if (Tokenizer_read(self, 1) == '!')
self->topstack->context |= LC_FAIL_NEXT;
else
return -1;
@@ -2448,7 +2458,7 @@ Tokenizer_verify_safe(Tokenizer* self, uint64_t context, Py_UNICODE data)
self->topstack->context |= LC_HAS_TEMPLATE | LC_FAIL_NEXT;
return 0;
}
if (data == '}' || (data == '<' && Tokenizer_READ(self, 1) == '!')) {
if (data == '}' || (data == '<' && Tokenizer_read(self, 1) == '!')) {
self->topstack->context |= LC_FAIL_NEXT;
return 0;
}
@@ -2475,8 +2485,8 @@ Tokenizer_verify_safe(Tokenizer* self, uint64_t context, Py_UNICODE data)
}
}
else if (context & LC_FAIL_ON_LBRACE) {
if (data == '{' || (Tokenizer_READ_BACKWARDS(self, 1) == '{' &&
Tokenizer_READ_BACKWARDS(self, 2) == '{')) {
if (data == '{' || (Tokenizer_read_backwards(self, 1) == '{' &&
Tokenizer_read_backwards(self, 2) == '{')) {
if (context & LC_TEMPLATE)
self->topstack->context |= LC_FAIL_ON_EQUALS;
else
@@ -2509,7 +2519,7 @@ static int Tokenizer_has_leading_whitespace(Tokenizer* self)
int offset = 1;
Py_UNICODE current_character;
while (1) {
current_character = Tokenizer_READ_BACKWARDS(self, offset);
current_character = Tokenizer_read_backwards(self, offset);
if (!current_character || current_character == '\n')
return 1;
else if (!Py_UNICODE_ISSPACE(current_character))
@@ -2533,7 +2543,7 @@ PyObject* Tokenizer_parse(Tokenizer* self, uint64_t context, int push)
return NULL;
}
while (1) {
this = Tokenizer_READ(self, 0);
this = Tokenizer_read(self, 0);
this_context = self->topstack->context;
if (this_context & AGG_UNSAFE) {
if (Tokenizer_verify_safe(self, this_context, this) < 0) {
@@ -2552,8 +2562,8 @@ PyObject* Tokenizer_parse(Tokenizer* self, uint64_t context, int push)
}
if (!this)
return Tokenizer_handle_end(self, this_context);
next = Tokenizer_READ(self, 1);
last = Tokenizer_READ_BACKWARDS(self, 1);
next = Tokenizer_read(self, 1);
last = Tokenizer_read_backwards(self, 1);
if (this == next && next == '{') {
if (Tokenizer_CAN_RECURSE(self)) {
if (Tokenizer_parse_template_or_argument(self))
@@ -2577,7 +2587,7 @@ PyObject* Tokenizer_parse(Tokenizer* self, uint64_t context, int push)
return NULL;
}
else if (this == next && next == '}' && this_context & LC_ARGUMENT) {
if (Tokenizer_READ(self, 2) == '}') {
if (Tokenizer_read(self, 2) == '}') {
return Tokenizer_handle_argument_end(self);
}
if (Tokenizer_emit_char(self, this))
@@ -2624,15 +2634,15 @@ PyObject* Tokenizer_parse(Tokenizer* self, uint64_t context, int push)
return NULL;
}
else if (this == '<' && next == '!') {
next_next = Tokenizer_READ(self, 2);
if (next_next == Tokenizer_READ(self, 3) && next_next == '-') {
next_next = Tokenizer_read(self, 2);
if (next_next == Tokenizer_read(self, 3) && next_next == '-') {
if (Tokenizer_parse_comment(self))
return NULL;
}
else if (Tokenizer_emit_char(self, this))
return NULL;
}
else if (this == '<' && next == '/' && Tokenizer_READ(self, 2)) {
else if (this == '<' && next == '/' && Tokenizer_read(self, 2)) {
if (this_context & LC_TAG_BODY ?
Tokenizer_handle_tag_open_close(self) :
Tokenizer_handle_invalid_tag_start(self))
@@ -2658,8 +2668,8 @@ PyObject* Tokenizer_parse(Tokenizer* self, uint64_t context, int push)
return NULL;
}
else if ((!last || last == '\n') && (this == '-' && this == next &&
this == Tokenizer_READ(self, 2) &&
this == Tokenizer_READ(self, 3))) {
this == Tokenizer_read(self, 2) &&
this == Tokenizer_read(self, 3))) {
if (Tokenizer_handle_hr(self))
return NULL;
}


+ 18
- 6
mwparserfromhell/parser/ctokenizer/tok_support.c View File

@@ -198,7 +198,7 @@ int Tokenizer_emit_token_kwargs(Tokenizer* self, PyObject* token,
/*
Write a Unicode codepoint to the current textbuffer.
*/
int Tokenizer_emit_char(Tokenizer* self, Py_UNICODE code)
int Tokenizer_emit_char(Tokenizer* self, Unicode code)
{
return Textbuffer_write(&(self->topstack->textbuffer), code);
}
@@ -337,26 +337,38 @@ int Tokenizer_emit_text_then_stack(Tokenizer* self, const char* text)
}

/*
Internal function to read the codepoint at the given index from the input.
*/
static Unicode read_codepoint(TokenizerInput* text, Py_ssize_t index)
{
#ifdef PEP_393
return PyUnicode_READ(text->kind, text->data, index);
#else
return text->buf[index];
#endif
}

/*
Read the value at a relative point in the wikicode, forwards.
*/
PyObject* Tokenizer_read(Tokenizer* self, Py_ssize_t delta)
Unicode Tokenizer_read(Tokenizer* self, Py_ssize_t delta)
{
Py_ssize_t index = self->head + delta;

if (index >= self->length)
if (index >= self->text.length)
return EMPTY;
return PyList_GET_ITEM(self->text, index);
return read_codepoint(&self->text, index);
}

/*
Read the value at a relative point in the wikicode, backwards.
*/
PyObject* Tokenizer_read_backwards(Tokenizer* self, Py_ssize_t delta)
Unicode Tokenizer_read_backwards(Tokenizer* self, Py_ssize_t delta)
{
Py_ssize_t index;

if (delta > self->head)
return EMPTY;
index = self->head - delta;
return PyList_GET_ITEM(self->text, index);
return read_codepoint(&self->text, index);
}

+ 3
- 7
mwparserfromhell/parser/ctokenizer/tok_support.h View File

@@ -35,24 +35,20 @@ void* Tokenizer_fail_route(Tokenizer*);

int Tokenizer_emit_token(Tokenizer*, PyObject*, int);
int Tokenizer_emit_token_kwargs(Tokenizer*, PyObject*, PyObject*, int);
int Tokenizer_emit_char(Tokenizer*, Py_UNICODE);
int Tokenizer_emit_char(Tokenizer*, Unicode);
int Tokenizer_emit_text(Tokenizer*, const char*);
int Tokenizer_emit_textbuffer(Tokenizer*, Textbuffer*, int);
int Tokenizer_emit_all(Tokenizer*, PyObject*);
int Tokenizer_emit_text_then_stack(Tokenizer*, const char*);

PyObject* Tokenizer_read(Tokenizer*, Py_ssize_t);
PyObject* Tokenizer_read_backwards(Tokenizer*, Py_ssize_t);
Unicode Tokenizer_read(Tokenizer*, Py_ssize_t);
Unicode Tokenizer_read_backwards(Tokenizer*, Py_ssize_t);

/* Macros */

#define MAX_DEPTH 40
#define MAX_CYCLES 100000

#define Tokenizer_READ(self, delta) \
(*PyUnicode_AS_UNICODE(Tokenizer_read(self, delta)))
#define Tokenizer_READ_BACKWARDS(self, delta) \
(*PyUnicode_AS_UNICODE(Tokenizer_read_backwards(self, delta)))
#define Tokenizer_CAN_RECURSE(self) \
(self->depth < MAX_DEPTH && self->cycles < MAX_CYCLES)



+ 57
- 16
mwparserfromhell/parser/ctokenizer/tokenizer.c View File

@@ -52,12 +52,20 @@ Tokenizer_new(PyTypeObject* type, PyObject* args, PyObject* kwds)
}

/*
Deallocate the given tokenizer's text field.
*/
static void dealloc_tokenizer_text(TokenizerInput* text)
{
Py_XDECREF(text->object);
}

/*
Deallocate the given tokenizer object.
*/
static void Tokenizer_dealloc(Tokenizer* self)
{
Stack *this = self->topstack, *next;
Py_XDECREF(self->text);
dealloc_tokenizer_text(&self->text);

while (this) {
Py_DECREF(this->stack);
@@ -70,6 +78,22 @@ static void Tokenizer_dealloc(Tokenizer* self)
}

/*
Initialize a new tokenizer instance's text field.
*/
static void init_tokenizer_text(TokenizerInput* text)
{
text->object = Py_None;
Py_INCREF(Py_None);
text->length = 0;
#ifdef PEP_393
text->kind = PyUnicode_WCHAR_KIND;
text->data = NULL;
#else
text->buf = NULL;
#endif
}

/*
Initialize a new tokenizer instance by setting instance attributes.
*/
static int Tokenizer_init(Tokenizer* self, PyObject* args, PyObject* kwds)
@@ -78,46 +102,63 @@ static int Tokenizer_init(Tokenizer* self, PyObject* args, PyObject* kwds)

if (!PyArg_ParseTupleAndKeywords(args, kwds, "", kwlist))
return -1;
self->text = Py_None;
Py_INCREF(Py_None);
init_tokenizer_text(&self->text);
self->topstack = NULL;
self->head = self->length = self->global = self->depth = self->cycles = 0;
self->head = self->global = self->depth = self->cycles = 0;
self->route_context = self->route_state = 0;
self->skip_style_tags = 0;
return 0;
}

/*
Load input text into the tokenizer.
*/
static int load_tokenizer_text(TokenizerInput* text, PyObject *input)
{
dealloc_tokenizer_text(text);
text->object = input;

#ifdef PEP_393
if (PyUnicode_READY(input) < 0)
return -1;
text->length = PyUnicode_GET_LENGTH(input);
text->kind = PyUnicode_KIND(input);
text->data = PyUnicode_DATA(input);
#else
text->length = PyUnicode_GET_SIZE(input);
text->buf = PyUnicode_AS_UNICODE(input);
#endif
}

/*
Build a list of tokens from a string of wikicode and return it.
*/
static PyObject* Tokenizer_tokenize(Tokenizer* self, PyObject* args)
{
PyObject *text, *temp, *tokens;
PyObject *input, *tokens;
uint64_t context = 0;
int skip_style_tags = 0;

if (PyArg_ParseTuple(args, "U|ii", &text, &context, &skip_style_tags)) {
Py_XDECREF(self->text);
self->text = PySequence_Fast(text, "expected a sequence");
if (PyArg_ParseTuple(args, "U|ii", &input, &context, &skip_style_tags)) {
if (load_tokenizer_text(&self->text, input))
return NULL;
}
else {
const char* encoded;
const char *encoded;
Py_ssize_t size;

/* Failed to parse a Unicode object; try a string instead. */
PyErr_Clear();
if (!PyArg_ParseTuple(args, "s#|ii", &encoded, &size, &context,
&skip_style_tags))
return NULL;
temp = PyUnicode_FromStringAndSize(encoded, size);
if (!text)
if (!(input = PyUnicode_FromStringAndSize(encoded, size)))
return NULL;
if (load_tokenizer_text(&self->text, input))
return NULL;
Py_XDECREF(self->text);
text = PySequence_Fast(temp, "expected a sequence");
Py_XDECREF(temp);
self->text = text;
}

self->head = self->global = self->depth = self->cycles = 0;
self->length = PyList_GET_SIZE(self->text);
self->skip_style_tags = skip_style_tags;
tokens = Tokenizer_parse(self, context, 1);



Write Preview
Loading…
Cancel
Save