diff options
author | Victor Stinner <victor.stinner@haypocalc.com> | 2011-10-03 11:53:37 (GMT) |
---|---|---|
committer | Victor Stinner <victor.stinner@haypocalc.com> | 2011-10-03 11:53:37 (GMT) |
commit | a3b334da6dd0477e5bf144934d184bc0b3e3779b (patch) | |
tree | 9f52ceabf86e5c992b45cfa8582cd98e1877031e /Include | |
parent | 1b4f9ceca79206770215c9cbb0b2de532dba4aec (diff) | |
download | cpython-a3b334da6dd0477e5bf144934d184bc0b3e3779b.zip cpython-a3b334da6dd0477e5bf144934d184bc0b3e3779b.tar.gz cpython-a3b334da6dd0477e5bf144934d184bc0b3e3779b.tar.bz2 |
PyUnicode_Ready() now sets ascii=1 if maxchar < 128
ascii=1 is no more reserved to PyASCIIObject. Use
PyUnicode_IS_COMPACT_ASCII(obj) to check if obj is a PyASCIIObject (as before).
Diffstat (limited to 'Include')
-rw-r--r-- | Include/unicodeobject.h | 41 |
1 files changed, 25 insertions, 16 deletions
diff --git a/Include/unicodeobject.h b/Include/unicodeobject.h index ba73e56..d7c9fa7 100644 --- a/Include/unicodeobject.h +++ b/Include/unicodeobject.h @@ -224,7 +224,7 @@ typedef struct { PyUnicode_4BYTE_KIND * compact = 1 * ready = 1 - * (ascii = 0) + * ascii = 0 - string created by the legacy API (not ready): @@ -236,7 +236,7 @@ typedef struct { * data.any is NULL * utf8 is NULL * interned = SSTATE_NOT_INTERNED - * (ascii = 0) + * ascii = 0 - string created by the legacy API, ready: @@ -246,7 +246,6 @@ typedef struct { * compact = 0 * ready = 1 * data.any is not NULL - * (ascii = 0) String created by the legacy API becomes ready when calling PyUnicode_READY(). @@ -278,8 +277,9 @@ typedef struct { one block for the PyUnicodeObject struct and another for its data buffer. */ unsigned int compact:1; - /* Compact objects which are ASCII-only also have the state.compact - flag set, and use the PyASCIIObject struct. */ + /* kind is PyUnicode_1BYTE_KIND but data contains only ASCII + characters. If ascii is 1 and compact is 1, use the PyASCIIObject + structure. */ unsigned int ascii:1; /* The ready flag indicates whether the object layout is initialized completely. This means that this is either a compact object, or @@ -304,7 +304,7 @@ typedef struct { /* Strings allocated through PyUnicode_FromUnicode(NULL, len) use the PyUnicodeObject structure. The actual string data is initially in the wstr - block, and copied into the data block using PyUnicode_Ready. */ + block, and copied into the data block using _PyUnicode_Ready. */ typedef struct { PyCompactUnicodeObject _base; union { @@ -327,7 +327,7 @@ PyAPI_DATA(PyTypeObject) PyUnicodeIter_Type; #ifndef Py_LIMITED_API #define PyUnicode_WSTR_LENGTH(op) \ - (((PyASCIIObject*)op)->state.ascii ? \ + (PyUnicode_IS_COMPACT_ASCII(op) ? \ ((PyASCIIObject*)op)->length : \ ((PyCompactUnicodeObject*)op)->wstr_length) @@ -369,10 +369,24 @@ PyAPI_DATA(PyTypeObject) PyUnicodeIter_Type; #define SSTATE_INTERNED_MORTAL 1 #define SSTATE_INTERNED_IMMORTAL 2 -#define PyUnicode_IS_COMPACT_ASCII(op) (((PyASCIIObject*)op)->state.ascii) +/* Return true if the string contains only ASCII characters, or 0 if not. The + string may be compact (PyUnicode_IS_COMPACT_ASCII) or not. No type checks + or Ready calls are performed. */ +#define PyUnicode_IS_ASCII(op) \ + (((PyASCIIObject*)op)->state.ascii) + +/* Return true if the string is compact or 0 if not. + No type checks or Ready calls are performed. */ +#define PyUnicode_IS_COMPACT(op) \ + (((PyASCIIObject*)(op))->state.compact) + +/* Return true if the string is a compact ASCII string (use PyASCIIObject + structure), or 0 if not. No type checks or Ready calls are performed. */ +#define PyUnicode_IS_COMPACT_ASCII(op) \ + (PyUnicode_IS_ASCII(op) && PyUnicode_IS_COMPACT(op)) /* String contains only wstr byte characters. This is only possible - when the string was created with a legacy API and PyUnicode_Ready() + when the string was created with a legacy API and _PyUnicode_Ready() has not been called yet. */ #define PyUnicode_WCHAR_KIND 0 @@ -399,11 +413,6 @@ PyAPI_DATA(PyTypeObject) PyUnicodeIter_Type; #define PyUnicode_2BYTE_DATA(op) ((Py_UCS2*)PyUnicode_DATA(op)) #define PyUnicode_4BYTE_DATA(op) ((Py_UCS4*)PyUnicode_DATA(op)) -/* Return true if the string is compact or 0 if not. - No type checks or Ready calls are performed. */ -#define PyUnicode_IS_COMPACT(op) \ - (((PyASCIIObject*)(op))->state.compact) - /* Return one of the PyUnicode_*_KIND values defined above. */ #define PyUnicode_KIND(op) \ (assert(PyUnicode_Check(op)), \ @@ -500,9 +509,9 @@ PyAPI_DATA(PyTypeObject) PyUnicodeIter_Type; #define PyUnicode_IS_READY(op) (((PyASCIIObject*)op)->state.ready) -/* PyUnicode_READY() does less work than PyUnicode_Ready() in the best +/* PyUnicode_READY() does less work than _PyUnicode_Ready() in the best case. If the canonical representation is not yet set, it will still call - PyUnicode_Ready(). + _PyUnicode_Ready(). Returns 0 on success and -1 on errors. */ #define PyUnicode_READY(op) \ (assert(PyUnicode_Check(op)), \ |