From: Alex Martelli Date: Sun, 2 Nov 2003 17:11:03 +0000 (+0000) Subject: fixed buggy comment as per SF bug #827856 X-Git-Tag: v2.3.3c1~86 X-Git-Url: http://git.ipfire.org/gitweb.cgi?a=commitdiff_plain;h=64e028ce27e8cf081e0304a005b0e48e05870d0b;p=thirdparty%2FPython%2Fcpython.git fixed buggy comment as per SF bug #827856 --- diff --git a/Include/object.h b/Include/object.h index 252e99c94d5e..3eb98a0c4042 100644 --- a/Include/object.h +++ b/Include/object.h @@ -695,22 +695,22 @@ It takes a while to get used to the proper usage of reference counts. Functions that create an object set the reference count to 1; such new objects must be stored somewhere or destroyed again with Py_DECREF(). -Functions that 'store' objects such as PyTuple_SetItem() and -PyDict_SetItemString() +Some functions that 'store' objects, such as PyTuple_SetItem() and +PyList_SetItem(), don't increment the reference count of the object, since the most frequent use is to store a fresh object. Functions that 'retrieve' -objects such as PyTuple_GetItem() and PyDict_GetItemString() also +objects, such as PyTuple_GetItem() and PyDict_GetItemString(), also don't increment the reference count, since most frequently the object is only looked at quickly. Thus, to retrieve an object and store it again, the caller must call Py_INCREF() explicitly. -NOTE: functions that 'consume' a reference count like -PyList_SetItemString() even consume the reference if the object wasn't -stored, to simplify error handling. +NOTE: functions that 'consume' a reference count, like +PyList_SetItem(), consume the reference even if the object wasn't +successfully stored, to simplify error handling. It seems attractive to make other functions that take an object as -argument consume a reference count; however this may quickly get +argument consume a reference count; however, this may quickly get confusing (even the current practice is already confusing). Consider it carefully, it may save lots of calls to Py_INCREF() and Py_DECREF() at times.