diff options
Diffstat (limited to 'Objects/stringlib')
| -rw-r--r-- | Objects/stringlib/codecs.h | 220 | ||||
| -rw-r--r-- | Objects/stringlib/ctype.h | 5 | ||||
| -rw-r--r-- | Objects/stringlib/fastsearch.h | 150 | ||||
| -rw-r--r-- | Objects/stringlib/find.h | 82 | ||||
| -rw-r--r-- | Objects/stringlib/find_max_char.h | 5 | ||||
| -rw-r--r-- | Objects/stringlib/join.h | 8 | ||||
| -rw-r--r-- | Objects/stringlib/localeutil.h | 4 | ||||
| -rw-r--r-- | Objects/stringlib/transmogrify.h | 615 | ||||
| -rw-r--r-- | Objects/stringlib/unicode_format.h | 2 |
9 files changed, 741 insertions, 350 deletions
diff --git a/Objects/stringlib/codecs.h b/Objects/stringlib/codecs.h index 0fc6b582d2..43f2f3266f 100644 --- a/Objects/stringlib/codecs.h +++ b/Objects/stringlib/codecs.h @@ -1,6 +1,8 @@ /* stringlib: codec implementations */ -#if STRINGLIB_IS_UNICODE +#if !STRINGLIB_IS_UNICODE +# error "codecs.h is specific to Unicode" +#endif /* Mask to quickly check whether a C 'long' contains a non-ASCII, UTF8-encoded char. */ @@ -260,53 +262,35 @@ STRINGLIB(utf8_encoder)(PyObject *unicode, Py_ssize_t size, const char *errors) { -#define MAX_SHORT_UNICHARS 300 /* largest size we'll do on the stack */ - - Py_ssize_t i; /* index into s of next input byte */ - PyObject *result; /* result string object */ + Py_ssize_t i; /* index into data of next input character */ char *p; /* next free byte in output buffer */ - Py_ssize_t nallocated; /* number of result bytes allocated */ - Py_ssize_t nneeded; /* number of result bytes needed */ #if STRINGLIB_SIZEOF_CHAR > 1 - PyObject *errorHandler = NULL; + PyObject *error_handler_obj = NULL; PyObject *exc = NULL; PyObject *rep = NULL; + _Py_error_handler error_handler = _Py_ERROR_UNKNOWN; #endif #if STRINGLIB_SIZEOF_CHAR == 1 const Py_ssize_t max_char_size = 2; - char stackbuf[MAX_SHORT_UNICHARS * 2]; #elif STRINGLIB_SIZEOF_CHAR == 2 const Py_ssize_t max_char_size = 3; - char stackbuf[MAX_SHORT_UNICHARS * 3]; #else /* STRINGLIB_SIZEOF_CHAR == 4 */ const Py_ssize_t max_char_size = 4; - char stackbuf[MAX_SHORT_UNICHARS * 4]; #endif + _PyBytesWriter writer; assert(size >= 0); + _PyBytesWriter_Init(&writer); - if (size <= MAX_SHORT_UNICHARS) { - /* Write into the stack buffer; nallocated can't overflow. - * At the end, we'll allocate exactly as much heap space as it - * turns out we need. - */ - nallocated = Py_SAFE_DOWNCAST(sizeof(stackbuf), size_t, int); - result = NULL; /* will allocate after we're done */ - p = stackbuf; - } - else { - if (size > PY_SSIZE_T_MAX / max_char_size) { - /* integer overflow */ - return PyErr_NoMemory(); - } - /* Overallocate on the heap, and give the excess back at the end. */ - nallocated = size * max_char_size; - result = PyBytes_FromStringAndSize(NULL, nallocated); - if (result == NULL) - return NULL; - p = PyBytes_AS_STRING(result); + if (size > PY_SSIZE_T_MAX / max_char_size) { + /* integer overflow */ + return PyErr_NoMemory(); } + p = _PyBytesWriter_Alloc(&writer, size * max_char_size); + if (p == NULL) + return NULL; + for (i = 0; i < size;) { Py_UCS4 ch = data[i++]; @@ -326,72 +310,117 @@ STRINGLIB(utf8_encoder)(PyObject *unicode, } #if STRINGLIB_SIZEOF_CHAR > 1 else if (Py_UNICODE_IS_SURROGATE(ch)) { - Py_ssize_t newpos; - Py_ssize_t repsize, k, startpos; + Py_ssize_t startpos, endpos, newpos; + Py_ssize_t k; + if (error_handler == _Py_ERROR_UNKNOWN) { + error_handler = get_error_handler(errors); + } + startpos = i-1; - rep = unicode_encode_call_errorhandler( - errors, &errorHandler, "utf-8", "surrogates not allowed", - unicode, &exc, startpos, startpos+1, &newpos); - if (!rep) - goto error; - - if (PyBytes_Check(rep)) - repsize = PyBytes_GET_SIZE(rep); - else - repsize = PyUnicode_GET_LENGTH(rep); + endpos = startpos+1; + + while ((endpos < size) && Py_UNICODE_IS_SURROGATE(data[endpos])) + endpos++; + + /* Only overallocate the buffer if it's not the last write */ + writer.overallocate = (endpos < size); + + switch (error_handler) + { + case _Py_ERROR_REPLACE: + memset(p, '?', endpos - startpos); + p += (endpos - startpos); + /* fall through the ignore handler */ + case _Py_ERROR_IGNORE: + i += (endpos - startpos - 1); + break; - if (repsize > max_char_size) { - Py_ssize_t offset; + case _Py_ERROR_SURROGATEPASS: + for (k=startpos; k<endpos; k++) { + ch = data[k]; + *p++ = (char)(0xe0 | (ch >> 12)); + *p++ = (char)(0x80 | ((ch >> 6) & 0x3f)); + *p++ = (char)(0x80 | (ch & 0x3f)); + } + i += (endpos - startpos - 1); + break; - if (result == NULL) - offset = p - stackbuf; - else - offset = p - PyBytes_AS_STRING(result); + case _Py_ERROR_BACKSLASHREPLACE: + /* subtract preallocated bytes */ + writer.min_size -= max_char_size * (endpos - startpos); + p = backslashreplace(&writer, p, + unicode, startpos, endpos); + if (p == NULL) + goto error; + i += (endpos - startpos - 1); + break; - if (nallocated > PY_SSIZE_T_MAX - repsize + max_char_size) { - /* integer overflow */ - PyErr_NoMemory(); + case _Py_ERROR_XMLCHARREFREPLACE: + /* subtract preallocated bytes */ + writer.min_size -= max_char_size * (endpos - startpos); + p = xmlcharrefreplace(&writer, p, + unicode, startpos, endpos); + if (p == NULL) goto error; + i += (endpos - startpos - 1); + break; + + case _Py_ERROR_SURROGATEESCAPE: + for (k=startpos; k<endpos; k++) { + ch = data[k]; + if (!(0xDC80 <= ch && ch <= 0xDCFF)) + break; + *p++ = (char)(ch & 0xff); } - nallocated += repsize - max_char_size; - if (result != NULL) { - if (_PyBytes_Resize(&result, nallocated) < 0) - goto error; - } else { - result = PyBytes_FromStringAndSize(NULL, nallocated); - if (result == NULL) - goto error; - Py_MEMCPY(PyBytes_AS_STRING(result), stackbuf, offset); + if (k >= endpos) { + i += (endpos - startpos - 1); + break; } - p = PyBytes_AS_STRING(result) + offset; - } + startpos = k; + assert(startpos < endpos); + /* fall through the default handler */ + default: + rep = unicode_encode_call_errorhandler( + errors, &error_handler_obj, "utf-8", "surrogates not allowed", + unicode, &exc, startpos, endpos, &newpos); + if (!rep) + goto error; - if (PyBytes_Check(rep)) { - char *prep = PyBytes_AS_STRING(rep); - for(k = repsize; k > 0; k--) - *p++ = *prep++; - } else /* rep is unicode */ { - enum PyUnicode_Kind repkind; - void *repdata; + /* subtract preallocated bytes */ + writer.min_size -= max_char_size * (newpos - startpos); - if (PyUnicode_READY(rep) < 0) - goto error; - repkind = PyUnicode_KIND(rep); - repdata = PyUnicode_DATA(rep); - - for(k=0; k<repsize; k++) { - Py_UCS4 c = PyUnicode_READ(repkind, repdata, k); - if (0x80 <= c) { - raise_encode_exception(&exc, "utf-8", - unicode, - i-1, i, + if (PyBytes_Check(rep)) { + p = _PyBytesWriter_WriteBytes(&writer, p, + PyBytes_AS_STRING(rep), + PyBytes_GET_SIZE(rep)); + } + else { + /* rep is unicode */ + if (PyUnicode_READY(rep) < 0) + goto error; + + if (!PyUnicode_IS_ASCII(rep)) { + raise_encode_exception(&exc, "utf-8", unicode, + startpos, endpos, "surrogates not allowed"); goto error; } - *p++ = (char)c; + + p = _PyBytesWriter_WriteBytes(&writer, p, + PyUnicode_DATA(rep), + PyUnicode_GET_LENGTH(rep)); } + + if (p == NULL) + goto error; + Py_CLEAR(rep); + + i = newpos; } - Py_CLEAR(rep); + + /* If overallocation was disabled, ensure that it was the last + write. Otherwise, we missed an optimization */ + assert(writer.overallocate || i == size); } else #if STRINGLIB_SIZEOF_CHAR > 2 @@ -416,35 +445,20 @@ STRINGLIB(utf8_encoder)(PyObject *unicode, #endif /* STRINGLIB_SIZEOF_CHAR > 1 */ } - if (result == NULL) { - /* This was stack allocated. */ - nneeded = p - stackbuf; - assert(nneeded <= nallocated); - result = PyBytes_FromStringAndSize(stackbuf, nneeded); - } - else { - /* Cut back to size actually needed. */ - nneeded = p - PyBytes_AS_STRING(result); - assert(nneeded <= nallocated); - _PyBytes_Resize(&result, nneeded); - } - #if STRINGLIB_SIZEOF_CHAR > 1 - Py_XDECREF(errorHandler); + Py_XDECREF(error_handler_obj); Py_XDECREF(exc); #endif - return result; + return _PyBytesWriter_Finish(&writer, p); #if STRINGLIB_SIZEOF_CHAR > 1 error: Py_XDECREF(rep); - Py_XDECREF(errorHandler); + Py_XDECREF(error_handler_obj); Py_XDECREF(exc); - Py_XDECREF(result); + _PyBytesWriter_Dealloc(&writer); return NULL; #endif - -#undef MAX_SHORT_UNICHARS } /* The pattern for constructing UCS2-repeated masks. */ @@ -806,5 +820,3 @@ STRINGLIB(utf32_encode)(const STRINGLIB_CHAR *in, #undef SWAB4 #endif - -#endif /* STRINGLIB_IS_UNICODE */ diff --git a/Objects/stringlib/ctype.h b/Objects/stringlib/ctype.h index 739cf3d9eb..f0546256ed 100644 --- a/Objects/stringlib/ctype.h +++ b/Objects/stringlib/ctype.h @@ -1,5 +1,6 @@ -/* NOTE: this API is -ONLY- for use with single byte character strings. */ -/* Do not use it with Unicode. */ +#if STRINGLIB_IS_UNICODE +# error "ctype.h only compatible with byte-wise strings" +#endif #include "bytes_methods.h" diff --git a/Objects/stringlib/fastsearch.h b/Objects/stringlib/fastsearch.h index cda68e77c8..98165ad114 100644 --- a/Objects/stringlib/fastsearch.h +++ b/Objects/stringlib/fastsearch.h @@ -32,52 +32,98 @@ #define STRINGLIB_BLOOM(mask, ch) \ ((mask & (1UL << ((ch) & (STRINGLIB_BLOOM_WIDTH -1))))) - Py_LOCAL_INLINE(Py_ssize_t) -STRINGLIB(fastsearch_memchr_1char)(const STRINGLIB_CHAR* s, Py_ssize_t n, - STRINGLIB_CHAR ch, unsigned char needle, - int mode) +STRINGLIB(find_char)(const STRINGLIB_CHAR* s, Py_ssize_t n, STRINGLIB_CHAR ch) { - if (mode == FAST_SEARCH) { - const STRINGLIB_CHAR *ptr = s; - const STRINGLIB_CHAR *e = s + n; - while (ptr < e) { - void *candidate = memchr((const void *) ptr, needle, (e - ptr) * sizeof(STRINGLIB_CHAR)); - if (candidate == NULL) - return -1; - ptr = (const STRINGLIB_CHAR *) _Py_ALIGN_DOWN(candidate, sizeof(STRINGLIB_CHAR)); - if (sizeof(STRINGLIB_CHAR) == 1 || *ptr == ch) - return (ptr - s); - /* False positive */ - ptr++; - } + const STRINGLIB_CHAR *p, *e; + + p = s; + e = s + n; + if (n > 10) { +#if STRINGLIB_SIZEOF_CHAR == 1 + p = memchr(s, ch, n); + if (p != NULL) + return (p - s); return -1; +#else + /* use memchr if we can choose a needle without two many likely + false positives */ + unsigned char needle = ch & 0xff; + /* If looking for a multiple of 256, we'd have too + many false positives looking for the '\0' byte in UCS2 + and UCS4 representations. */ + if (needle != 0) { + while (p < e) { + void *candidate = memchr(p, needle, + (e - p) * sizeof(STRINGLIB_CHAR)); + if (candidate == NULL) + return -1; + p = (const STRINGLIB_CHAR *) + _Py_ALIGN_DOWN(candidate, sizeof(STRINGLIB_CHAR)); + if (*p == ch) + return (p - s); + /* False positive */ + p++; + } + return -1; + } +#endif } + while (p < e) { + if (*p == ch) + return (p - s); + p++; + } + return -1; +} + +Py_LOCAL_INLINE(Py_ssize_t) +STRINGLIB(rfind_char)(const STRINGLIB_CHAR* s, Py_ssize_t n, STRINGLIB_CHAR ch) +{ + const STRINGLIB_CHAR *p; #ifdef HAVE_MEMRCHR /* memrchr() is a GNU extension, available since glibc 2.1.91. it doesn't seem as optimized as memchr(), but is still quite - faster than our hand-written loop in FASTSEARCH below */ - else if (mode == FAST_RSEARCH) { - while (n > 0) { - const STRINGLIB_CHAR *found; - void *candidate = memrchr((const void *) s, needle, n * sizeof(STRINGLIB_CHAR)); - if (candidate == NULL) - return -1; - found = (const STRINGLIB_CHAR *) _Py_ALIGN_DOWN(candidate, sizeof(STRINGLIB_CHAR)); - n = found - s; - if (sizeof(STRINGLIB_CHAR) == 1 || *found == ch) - return n; - /* False positive */ - } + faster than our hand-written loop below */ + + if (n > 10) { +#if STRINGLIB_SIZEOF_CHAR == 1 + p = memrchr(s, ch, n); + if (p != NULL) + return (p - s); return -1; - } +#else + /* use memrchr if we can choose a needle without two many likely + false positives */ + unsigned char needle = ch & 0xff; + /* If looking for a multiple of 256, we'd have too + many false positives looking for the '\0' byte in UCS2 + and UCS4 representations. */ + if (needle != 0) { + while (n > 0) { + void *candidate = memrchr(s, needle, + n * sizeof(STRINGLIB_CHAR)); + if (candidate == NULL) + return -1; + p = (const STRINGLIB_CHAR *) + _Py_ALIGN_DOWN(candidate, sizeof(STRINGLIB_CHAR)); + n = p - s; + if (*p == ch) + return n; + /* False positive */ + } + return -1; + } #endif - else { - assert(0); /* Should never get here */ - return 0; } - -#undef DO_MEMCHR +#endif /* HAVE_MEMRCHR */ + p = s + n; + while (p > s) { + p--; + if (*p == ch) + return (p - s); + } + return -1; } Py_LOCAL_INLINE(Py_ssize_t) @@ -99,25 +145,11 @@ FASTSEARCH(const STRINGLIB_CHAR* s, Py_ssize_t n, if (m <= 0) return -1; /* use special case for 1-character strings */ - if (n > 10 && (mode == FAST_SEARCH -#ifdef HAVE_MEMRCHR - || mode == FAST_RSEARCH -#endif - )) { - /* use memchr if we can choose a needle without two many likely - false positives */ - unsigned char needle; - needle = p[0] & 0xff; -#if STRINGLIB_SIZEOF_CHAR > 1 - /* If looking for a multiple of 256, we'd have too - many false positives looking for the '\0' byte in UCS2 - and UCS4 representations. */ - if (needle != 0) -#endif - return STRINGLIB(fastsearch_memchr_1char) - (s, n, p[0], needle, mode); - } - if (mode == FAST_COUNT) { + if (mode == FAST_SEARCH) + return STRINGLIB(find_char)(s, n, p[0]); + else if (mode == FAST_RSEARCH) + return STRINGLIB(rfind_char)(s, n, p[0]); + else { /* FAST_COUNT */ for (i = 0; i < n; i++) if (s[i] == p[0]) { count++; @@ -125,14 +157,6 @@ FASTSEARCH(const STRINGLIB_CHAR* s, Py_ssize_t n, return maxcount; } return count; - } else if (mode == FAST_SEARCH) { - for (i = 0; i < n; i++) - if (s[i] == p[0]) - return i; - } else { /* FAST_RSEARCH */ - for (i = n - 1; i > -1; i--) - if (s[i] == p[0]) - return i; } return -1; } diff --git a/Objects/stringlib/find.h b/Objects/stringlib/find.h index 14815f6e62..509b929739 100644 --- a/Objects/stringlib/find.h +++ b/Objects/stringlib/find.h @@ -117,85 +117,3 @@ STRINGLIB(parse_args_finds)(const char * function_name, PyObject *args, } #undef FORMAT_BUFFER_SIZE - -#if STRINGLIB_IS_UNICODE - -/* -Wraps stringlib_parse_args_finds() and additionally ensures that the -first argument is a unicode object. - -Note that we receive a pointer to the pointer of the substring object, -so when we create that object in this function we don't DECREF it, -because it continues living in the caller functions (those functions, -after finishing using the substring, must DECREF it). -*/ - -Py_LOCAL_INLINE(int) -STRINGLIB(parse_args_finds_unicode)(const char * function_name, PyObject *args, - PyObject **substring, - Py_ssize_t *start, Py_ssize_t *end) -{ - PyObject *tmp_substring; - - if(STRINGLIB(parse_args_finds)(function_name, args, &tmp_substring, - start, end)) { - tmp_substring = PyUnicode_FromObject(tmp_substring); - if (!tmp_substring) - return 0; - *substring = tmp_substring; - return 1; - } - return 0; -} - -#else /* !STRINGLIB_IS_UNICODE */ - -/* -Wraps stringlib_parse_args_finds() and additionally checks whether the -first argument is an integer in range(0, 256). - -If this is the case, writes the integer value to the byte parameter -and sets subobj to NULL. Otherwise, sets the first argument to subobj -and doesn't touch byte. The other parameters are similar to those of -stringlib_parse_args_finds(). -*/ - -Py_LOCAL_INLINE(int) -STRINGLIB(parse_args_finds_byte)(const char *function_name, PyObject *args, - PyObject **subobj, char *byte, - Py_ssize_t *start, Py_ssize_t *end) -{ - PyObject *tmp_subobj; - Py_ssize_t ival; - PyObject *err; - - if(!STRINGLIB(parse_args_finds)(function_name, args, &tmp_subobj, - start, end)) - return 0; - - if (!PyNumber_Check(tmp_subobj)) { - *subobj = tmp_subobj; - return 1; - } - - ival = PyNumber_AsSsize_t(tmp_subobj, PyExc_OverflowError); - if (ival == -1) { - err = PyErr_Occurred(); - if (err && !PyErr_GivenExceptionMatches(err, PyExc_OverflowError)) { - PyErr_Clear(); - *subobj = tmp_subobj; - return 1; - } - } - - if (ival < 0 || ival > 255) { - PyErr_SetString(PyExc_ValueError, "byte must be in range(0, 256)"); - return 0; - } - - *subobj = NULL; - *byte = (char)ival; - return 1; -} - -#endif /* STRINGLIB_IS_UNICODE */ diff --git a/Objects/stringlib/find_max_char.h b/Objects/stringlib/find_max_char.h index eb3fe886e2..8ccbc30944 100644 --- a/Objects/stringlib/find_max_char.h +++ b/Objects/stringlib/find_max_char.h @@ -1,6 +1,8 @@ /* Finding the optimal width of unicode characters in a buffer */ -#if STRINGLIB_IS_UNICODE +#if !STRINGLIB_IS_UNICODE +# error "find_max_char.h is specific to Unicode" +#endif /* Mask to quickly check whether a C 'long' contains a non-ASCII, UTF8-encoded char. */ @@ -129,5 +131,4 @@ STRINGLIB(find_max_char)(const STRINGLIB_CHAR *begin, const STRINGLIB_CHAR *end) #undef MAX_CHAR_UCS4 #endif /* STRINGLIB_SIZEOF_CHAR == 1 */ -#endif /* STRINGLIB_IS_UNICODE */ diff --git a/Objects/stringlib/join.h b/Objects/stringlib/join.h index cbf81be170..6f314e1524 100644 --- a/Objects/stringlib/join.h +++ b/Objects/stringlib/join.h @@ -1,6 +1,6 @@ /* stringlib: bytes joining implementation */ -#if STRINGLIB_SIZEOF_CHAR != 1 +#if STRINGLIB_IS_UNICODE #error join.h only compatible with byte-wise strings #endif @@ -107,7 +107,7 @@ STRINGLIB(bytes_join)(PyObject *sep, PyObject *iterable) for (i = 0; i < nbufs; i++) { Py_ssize_t n = buffers[i].len; char *q = buffers[i].buf; - Py_MEMCPY(p, q, n); + memcpy(p, q, n); p += n; } goto done; @@ -116,12 +116,12 @@ STRINGLIB(bytes_join)(PyObject *sep, PyObject *iterable) Py_ssize_t n; char *q; if (i) { - Py_MEMCPY(p, sepstr, seplen); + memcpy(p, sepstr, seplen); p += seplen; } n = buffers[i].len; q = buffers[i].buf; - Py_MEMCPY(p, q, n); + memcpy(p, q, n); p += n; } goto done; diff --git a/Objects/stringlib/localeutil.h b/Objects/stringlib/localeutil.h index 6e2f07342c..df501ed05c 100644 --- a/Objects/stringlib/localeutil.h +++ b/Objects/stringlib/localeutil.h @@ -2,8 +2,8 @@ #include <locale.h> -#ifndef STRINGLIB_IS_UNICODE -# error "localeutil is specific to Unicode" +#if !STRINGLIB_IS_UNICODE +# error "localeutil.h is specific to Unicode" #endif typedef struct { diff --git a/Objects/stringlib/transmogrify.h b/Objects/stringlib/transmogrify.h index b559b5356b..a314572a72 100644 --- a/Objects/stringlib/transmogrify.h +++ b/Objects/stringlib/transmogrify.h @@ -1,14 +1,21 @@ -/* NOTE: this API is -ONLY- for use with single byte character strings. */ -/* Do not use it with Unicode. */ +#if STRINGLIB_IS_UNICODE +# error "transmogrify.h only compatible with byte-wise strings" +#endif /* the more complicated methods. parts of these should be pulled out into the shared code in bytes_methods.c to cut down on duplicate code bloat. */ -PyDoc_STRVAR(expandtabs__doc__, -"B.expandtabs(tabsize=8) -> copy of B\n\ -\n\ -Return a copy of B where all tab characters are expanded using spaces.\n\ -If tabsize is not given, a tab size of 8 characters is assumed."); +static inline PyObject * +return_self(PyObject *self) +{ +#if !STRINGLIB_MUTABLE + if (STRINGLIB_CHECK_EXACT(self)) { + Py_INCREF(self); + return self; + } +#endif + return STRINGLIB_NEW(STRINGLIB_STR(self), STRINGLIB_LEN(self)); +} static PyObject* stringlib_expandtabs(PyObject *self, PyObject *args, PyObject *kwds) @@ -83,7 +90,7 @@ stringlib_expandtabs(PyObject *self, PyObject *args, PyObject *kwds) return NULL; } -Py_LOCAL_INLINE(PyObject *) +static inline PyObject * pad(PyObject *self, Py_ssize_t left, Py_ssize_t right, char fill) { PyObject *u; @@ -93,39 +100,25 @@ pad(PyObject *self, Py_ssize_t left, Py_ssize_t right, char fill) if (right < 0) right = 0; - if (left == 0 && right == 0 && STRINGLIB_CHECK_EXACT(self)) { -#if STRINGLIB_MUTABLE - /* We're defined as returning a copy; If the object is mutable - * that means we must make an identical copy. */ - return STRINGLIB_NEW(STRINGLIB_STR(self), STRINGLIB_LEN(self)); -#else - Py_INCREF(self); - return (PyObject *)self; -#endif /* STRINGLIB_MUTABLE */ + if (left == 0 && right == 0) { + return return_self(self); } - u = STRINGLIB_NEW(NULL, - left + STRINGLIB_LEN(self) + right); + u = STRINGLIB_NEW(NULL, left + STRINGLIB_LEN(self) + right); if (u) { if (left) memset(STRINGLIB_STR(u), fill, left); - Py_MEMCPY(STRINGLIB_STR(u) + left, - STRINGLIB_STR(self), - STRINGLIB_LEN(self)); + memcpy(STRINGLIB_STR(u) + left, + STRINGLIB_STR(self), + STRINGLIB_LEN(self)); if (right) memset(STRINGLIB_STR(u) + left + STRINGLIB_LEN(self), - fill, right); + fill, right); } return u; } -PyDoc_STRVAR(ljust__doc__, -"B.ljust(width[, fillchar]) -> copy of B\n" -"\n" -"Return B left justified in a string of length width. Padding is\n" -"done using the specified fill character (default is a space)."); - static PyObject * stringlib_ljust(PyObject *self, PyObject *args) { @@ -135,27 +128,14 @@ stringlib_ljust(PyObject *self, PyObject *args) if (!PyArg_ParseTuple(args, "n|c:ljust", &width, &fillchar)) return NULL; - if (STRINGLIB_LEN(self) >= width && STRINGLIB_CHECK_EXACT(self)) { -#if STRINGLIB_MUTABLE - /* We're defined as returning a copy; If the object is mutable - * that means we must make an identical copy. */ - return STRINGLIB_NEW(STRINGLIB_STR(self), STRINGLIB_LEN(self)); -#else - Py_INCREF(self); - return (PyObject*) self; -#endif + if (STRINGLIB_LEN(self) >= width) { + return return_self(self); } return pad(self, 0, width - STRINGLIB_LEN(self), fillchar); } -PyDoc_STRVAR(rjust__doc__, -"B.rjust(width[, fillchar]) -> copy of B\n" -"\n" -"Return B right justified in a string of length width. Padding is\n" -"done using the specified fill character (default is a space)"); - static PyObject * stringlib_rjust(PyObject *self, PyObject *args) { @@ -165,27 +145,14 @@ stringlib_rjust(PyObject *self, PyObject *args) if (!PyArg_ParseTuple(args, "n|c:rjust", &width, &fillchar)) return NULL; - if (STRINGLIB_LEN(self) >= width && STRINGLIB_CHECK_EXACT(self)) { -#if STRINGLIB_MUTABLE - /* We're defined as returning a copy; If the object is mutable - * that means we must make an identical copy. */ - return STRINGLIB_NEW(STRINGLIB_STR(self), STRINGLIB_LEN(self)); -#else - Py_INCREF(self); - return (PyObject*) self; -#endif + if (STRINGLIB_LEN(self) >= width) { + return return_self(self); } return pad(self, width - STRINGLIB_LEN(self), 0, fillchar); } -PyDoc_STRVAR(center__doc__, -"B.center(width[, fillchar]) -> copy of B\n" -"\n" -"Return B centered in a string of length width. Padding is\n" -"done using the specified fill character (default is a space)."); - static PyObject * stringlib_center(PyObject *self, PyObject *args) { @@ -196,15 +163,8 @@ stringlib_center(PyObject *self, PyObject *args) if (!PyArg_ParseTuple(args, "n|c:center", &width, &fillchar)) return NULL; - if (STRINGLIB_LEN(self) >= width && STRINGLIB_CHECK_EXACT(self)) { -#if STRINGLIB_MUTABLE - /* We're defined as returning a copy; If the object is mutable - * that means we must make an identical copy. */ - return STRINGLIB_NEW(STRINGLIB_STR(self), STRINGLIB_LEN(self)); -#else - Py_INCREF(self); - return (PyObject*) self; -#endif + if (STRINGLIB_LEN(self) >= width) { + return return_self(self); } marg = width - STRINGLIB_LEN(self); @@ -213,12 +173,6 @@ stringlib_center(PyObject *self, PyObject *args) return pad(self, left, marg - left, fillchar); } -PyDoc_STRVAR(zfill__doc__, -"B.zfill(width) -> copy of B\n" -"\n" -"Pad a numeric string B with zeros on the left, to fill a field\n" -"of the specified width. B is never truncated."); - static PyObject * stringlib_zfill(PyObject *self, PyObject *args) { @@ -231,21 +185,7 @@ stringlib_zfill(PyObject *self, PyObject *args) return NULL; if (STRINGLIB_LEN(self) >= width) { - if (STRINGLIB_CHECK_EXACT(self)) { -#if STRINGLIB_MUTABLE - /* We're defined as returning a copy; If the object is mutable - * that means we must make an identical copy. */ - return STRINGLIB_NEW(STRINGLIB_STR(self), STRINGLIB_LEN(self)); -#else - Py_INCREF(self); - return (PyObject*) self; -#endif - } - else - return STRINGLIB_NEW( - STRINGLIB_STR(self), - STRINGLIB_LEN(self) - ); + return return_self(self); } fill = width - STRINGLIB_LEN(self); @@ -262,5 +202,500 @@ stringlib_zfill(PyObject *self, PyObject *args) p[fill] = '0'; } - return (PyObject*) s; + return s; +} + + +/* find and count characters and substrings */ + +#define findchar(target, target_len, c) \ + ((char *)memchr((const void *)(target), c, target_len)) + + +static Py_ssize_t +countchar(const char *target, Py_ssize_t target_len, char c, + Py_ssize_t maxcount) +{ + Py_ssize_t count = 0; + const char *start = target; + const char *end = target + target_len; + + while ((start = findchar(start, end - start, c)) != NULL) { + count++; + if (count >= maxcount) + break; + start += 1; + } + return count; +} + + +/* Algorithms for different cases of string replacement */ + +/* len(self)>=1, from="", len(to)>=1, maxcount>=1 */ +static PyObject * +stringlib_replace_interleave(PyObject *self, + const char *to_s, Py_ssize_t to_len, + Py_ssize_t maxcount) +{ + const char *self_s; + char *result_s; + Py_ssize_t self_len, result_len; + Py_ssize_t count, i; + PyObject *result; + + self_len = STRINGLIB_LEN(self); + + /* 1 at the end plus 1 after every character; + count = min(maxcount, self_len + 1) */ + if (maxcount <= self_len) { + count = maxcount; + } + else { + /* Can't overflow: self_len + 1 <= maxcount <= PY_SSIZE_T_MAX. */ + count = self_len + 1; + } + + /* Check for overflow */ + /* result_len = count * to_len + self_len; */ + assert(count > 0); + if (to_len > (PY_SSIZE_T_MAX - self_len) / count) { + PyErr_SetString(PyExc_OverflowError, + "replace bytes are too long"); + return NULL; + } + result_len = count * to_len + self_len; + result = STRINGLIB_NEW(NULL, result_len); + if (result == NULL) { + return NULL; + } + + self_s = STRINGLIB_STR(self); + result_s = STRINGLIB_STR(result); + + if (to_len > 1) { + /* Lay the first one down (guaranteed this will occur) */ + memcpy(result_s, to_s, to_len); + result_s += to_len; + count -= 1; + + for (i = 0; i < count; i++) { + *result_s++ = *self_s++; + memcpy(result_s, to_s, to_len); + result_s += to_len; + } + } + else { + result_s[0] = to_s[0]; + result_s += to_len; + count -= 1; + for (i = 0; i < count; i++) { + *result_s++ = *self_s++; + result_s[0] = to_s[0]; + result_s += to_len; + } + } + + /* Copy the rest of the original string */ + memcpy(result_s, self_s, self_len - i); + + return result; +} + +/* Special case for deleting a single character */ +/* len(self)>=1, len(from)==1, to="", maxcount>=1 */ +static PyObject * +stringlib_replace_delete_single_character(PyObject *self, + char from_c, Py_ssize_t maxcount) +{ + const char *self_s, *start, *next, *end; + char *result_s; + Py_ssize_t self_len, result_len; + Py_ssize_t count; + PyObject *result; + + self_len = STRINGLIB_LEN(self); + self_s = STRINGLIB_STR(self); + + count = countchar(self_s, self_len, from_c, maxcount); + if (count == 0) { + return return_self(self); + } + + result_len = self_len - count; /* from_len == 1 */ + assert(result_len>=0); + + result = STRINGLIB_NEW(NULL, result_len); + if (result == NULL) { + return NULL; + } + result_s = STRINGLIB_STR(result); + + start = self_s; + end = self_s + self_len; + while (count-- > 0) { + next = findchar(start, end - start, from_c); + if (next == NULL) + break; + memcpy(result_s, start, next - start); + result_s += (next - start); + start = next + 1; + } + memcpy(result_s, start, end - start); + + return result; +} + +/* len(self)>=1, len(from)>=2, to="", maxcount>=1 */ + +static PyObject * +stringlib_replace_delete_substring(PyObject *self, + const char *from_s, Py_ssize_t from_len, + Py_ssize_t maxcount) +{ + const char *self_s, *start, *next, *end; + char *result_s; + Py_ssize_t self_len, result_len; + Py_ssize_t count, offset; + PyObject *result; + + self_len = STRINGLIB_LEN(self); + self_s = STRINGLIB_STR(self); + + count = stringlib_count(self_s, self_len, + from_s, from_len, + maxcount); + + if (count == 0) { + /* no matches */ + return return_self(self); + } + + result_len = self_len - (count * from_len); + assert (result_len>=0); + + result = STRINGLIB_NEW(NULL, result_len); + if (result == NULL) { + return NULL; + } + result_s = STRINGLIB_STR(result); + + start = self_s; + end = self_s + self_len; + while (count-- > 0) { + offset = stringlib_find(start, end - start, + from_s, from_len, + 0); + if (offset == -1) + break; + next = start + offset; + + memcpy(result_s, start, next - start); + + result_s += (next - start); + start = next + from_len; + } + memcpy(result_s, start, end - start); + return result; +} + +/* len(self)>=1, len(from)==len(to)==1, maxcount>=1 */ +static PyObject * +stringlib_replace_single_character_in_place(PyObject *self, + char from_c, char to_c, + Py_ssize_t maxcount) +{ + const char *self_s, *end; + char *result_s, *start, *next; + Py_ssize_t self_len; + PyObject *result; + + /* The result string will be the same size */ + self_s = STRINGLIB_STR(self); + self_len = STRINGLIB_LEN(self); + + next = findchar(self_s, self_len, from_c); + + if (next == NULL) { + /* No matches; return the original bytes */ + return return_self(self); + } + + /* Need to make a new bytes */ + result = STRINGLIB_NEW(NULL, self_len); + if (result == NULL) { + return NULL; + } + result_s = STRINGLIB_STR(result); + memcpy(result_s, self_s, self_len); + + /* change everything in-place, starting with this one */ + start = result_s + (next - self_s); + *start = to_c; + start++; + end = result_s + self_len; + + while (--maxcount > 0) { + next = findchar(start, end - start, from_c); + if (next == NULL) + break; + *next = to_c; + start = next + 1; + } + + return result; } + +/* len(self)>=1, len(from)==len(to)>=2, maxcount>=1 */ +static PyObject * +stringlib_replace_substring_in_place(PyObject *self, + const char *from_s, Py_ssize_t from_len, + const char *to_s, Py_ssize_t to_len, + Py_ssize_t maxcount) +{ + const char *self_s, *end; + char *result_s, *start; + Py_ssize_t self_len, offset; + PyObject *result; + + /* The result bytes will be the same size */ + + self_s = STRINGLIB_STR(self); + self_len = STRINGLIB_LEN(self); + + offset = stringlib_find(self_s, self_len, + from_s, from_len, + 0); + if (offset == -1) { + /* No matches; return the original bytes */ + return return_self(self); + } + + /* Need to make a new bytes */ + result = STRINGLIB_NEW(NULL, self_len); + if (result == NULL) { + return NULL; + } + result_s = STRINGLIB_STR(result); + memcpy(result_s, self_s, self_len); + + /* change everything in-place, starting with this one */ + start = result_s + offset; + memcpy(start, to_s, from_len); + start += from_len; + end = result_s + self_len; + + while ( --maxcount > 0) { + offset = stringlib_find(start, end - start, + from_s, from_len, + 0); + if (offset == -1) + break; + memcpy(start + offset, to_s, from_len); + start += offset + from_len; + } + + return result; +} + +/* len(self)>=1, len(from)==1, len(to)>=2, maxcount>=1 */ +static PyObject * +stringlib_replace_single_character(PyObject *self, + char from_c, + const char *to_s, Py_ssize_t to_len, + Py_ssize_t maxcount) +{ + const char *self_s, *start, *next, *end; + char *result_s; + Py_ssize_t self_len, result_len; + Py_ssize_t count; + PyObject *result; + + self_s = STRINGLIB_STR(self); + self_len = STRINGLIB_LEN(self); + + count = countchar(self_s, self_len, from_c, maxcount); + if (count == 0) { + /* no matches, return unchanged */ + return return_self(self); + } + + /* use the difference between current and new, hence the "-1" */ + /* result_len = self_len + count * (to_len-1) */ + assert(count > 0); + if (to_len - 1 > (PY_SSIZE_T_MAX - self_len) / count) { + PyErr_SetString(PyExc_OverflowError, "replace bytes is too long"); + return NULL; + } + result_len = self_len + count * (to_len - 1); + + result = STRINGLIB_NEW(NULL, result_len); + if (result == NULL) { + return NULL; + } + result_s = STRINGLIB_STR(result); + + start = self_s; + end = self_s + self_len; + while (count-- > 0) { + next = findchar(start, end - start, from_c); + if (next == NULL) + break; + + if (next == start) { + /* replace with the 'to' */ + memcpy(result_s, to_s, to_len); + result_s += to_len; + start += 1; + } else { + /* copy the unchanged old then the 'to' */ + memcpy(result_s, start, next - start); + result_s += (next - start); + memcpy(result_s, to_s, to_len); + result_s += to_len; + start = next + 1; + } + } + /* Copy the remainder of the remaining bytes */ + memcpy(result_s, start, end - start); + + return result; +} + +/* len(self)>=1, len(from)>=2, len(to)>=2, maxcount>=1 */ +static PyObject * +stringlib_replace_substring(PyObject *self, + const char *from_s, Py_ssize_t from_len, + const char *to_s, Py_ssize_t to_len, + Py_ssize_t maxcount) +{ + const char *self_s, *start, *next, *end; + char *result_s; + Py_ssize_t self_len, result_len; + Py_ssize_t count, offset; + PyObject *result; + + self_s = STRINGLIB_STR(self); + self_len = STRINGLIB_LEN(self); + + count = stringlib_count(self_s, self_len, + from_s, from_len, + maxcount); + + if (count == 0) { + /* no matches, return unchanged */ + return return_self(self); + } + + /* Check for overflow */ + /* result_len = self_len + count * (to_len-from_len) */ + assert(count > 0); + if (to_len - from_len > (PY_SSIZE_T_MAX - self_len) / count) { + PyErr_SetString(PyExc_OverflowError, "replace bytes is too long"); + return NULL; + } + result_len = self_len + count * (to_len - from_len); + + result = STRINGLIB_NEW(NULL, result_len); + if (result == NULL) { + return NULL; + } + result_s = STRINGLIB_STR(result); + + start = self_s; + end = self_s + self_len; + while (count-- > 0) { + offset = stringlib_find(start, end - start, + from_s, from_len, + 0); + if (offset == -1) + break; + next = start + offset; + if (next == start) { + /* replace with the 'to' */ + memcpy(result_s, to_s, to_len); + result_s += to_len; + start += from_len; + } else { + /* copy the unchanged old then the 'to' */ + memcpy(result_s, start, next - start); + result_s += (next - start); + memcpy(result_s, to_s, to_len); + result_s += to_len; + start = next + from_len; + } + } + /* Copy the remainder of the remaining bytes */ + memcpy(result_s, start, end - start); + + return result; +} + + +static PyObject * +stringlib_replace(PyObject *self, + const char *from_s, Py_ssize_t from_len, + const char *to_s, Py_ssize_t to_len, + Py_ssize_t maxcount) +{ + if (maxcount < 0) { + maxcount = PY_SSIZE_T_MAX; + } else if (maxcount == 0 || STRINGLIB_LEN(self) == 0) { + /* nothing to do; return the original bytes */ + return return_self(self); + } + + /* Handle zero-length special cases */ + if (from_len == 0) { + if (to_len == 0) { + /* nothing to do; return the original bytes */ + return return_self(self); + } + /* insert the 'to' bytes everywhere. */ + /* >>> b"Python".replace(b"", b".") */ + /* b'.P.y.t.h.o.n.' */ + return stringlib_replace_interleave(self, to_s, to_len, maxcount); + } + + /* Except for b"".replace(b"", b"A") == b"A" there is no way beyond this */ + /* point for an empty self bytes to generate a non-empty bytes */ + /* Special case so the remaining code always gets a non-empty bytes */ + if (STRINGLIB_LEN(self) == 0) { + return return_self(self); + } + + if (to_len == 0) { + /* delete all occurrences of 'from' bytes */ + if (from_len == 1) { + return stringlib_replace_delete_single_character( + self, from_s[0], maxcount); + } else { + return stringlib_replace_delete_substring( + self, from_s, from_len, maxcount); + } + } + + /* Handle special case where both bytes have the same length */ + + if (from_len == to_len) { + if (from_len == 1) { + return stringlib_replace_single_character_in_place( + self, from_s[0], to_s[0], maxcount); + } else { + return stringlib_replace_substring_in_place( + self, from_s, from_len, to_s, to_len, maxcount); + } + } + + /* Otherwise use the more generic algorithms */ + if (from_len == 1) { + return stringlib_replace_single_character( + self, from_s[0], to_s, to_len, maxcount); + } else { + /* len('from')>=2, len('to')>=1 */ + return stringlib_replace_substring( + self, from_s, from_len, to_s, to_len, maxcount); + } +} + +#undef findchar diff --git a/Objects/stringlib/unicode_format.h b/Objects/stringlib/unicode_format.h index be09b5f6fa..14fa28ea54 100644 --- a/Objects/stringlib/unicode_format.h +++ b/Objects/stringlib/unicode_format.h @@ -67,7 +67,7 @@ SubString_new_object(SubString *str) return PyUnicode_Substring(str->str, str->start, str->end); } -/* return a new string. if str->str is NULL, return None */ +/* return a new string. if str->str is NULL, return a new empty string */ Py_LOCAL_INLINE(PyObject *) SubString_new_object_or_empty(SubString *str) { |