IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
CR1829: Fix memory leak when unpacking 'B' buffers in Python.
(PyList_Append creates a new reference to the added object, and so we
have to release the existing one.)
- combine py_foo_proto.h and py_foo.h
- move #include of Python.h into one place so we can get rid of that
annoying compiler warning about HAVE_FSTAT being redefined
calls to init_unistr2() in the code and every one of them got the 3rd
argument incorrect, so I thought it best just to remove the argument.
The incorrect usage was caused by callers using strlen() to determine
the length of the string. The 3rd argument to init_unistr2() was
supposed to be the character length, not the byte length of the
string, so for non-english this could come out wrong.
I also removed the bogus 'always allocate at least 256 bytes'
hack. There may be some code that relies on this, but if there is then
the code is broken and needs fixing.
* (pytdbunpack_item): PyList_Append creates an additional reference to
the appended object. Therefore, release the initial reference after
it's added to the list.
on encoding. (Tough luck if your Samba codepage is not 850.) It's
much cleaner and possibly more efficient now too: just build one big
list of strings, then zip it up at the end.
Still need to do this for unpacking.
rather than as just a string. Makes the code more messy, but needed
for compatibility with existing PSA Python code which seems to be too
knotty to separate out.
