On 2007-05-04 18:53, Georg Brandl wrote:
M.-A. Lemburg schrieb:
Hi Walter,
if the bytes type does turn out to be a mutable type as suggested in PEP 358, then please make sure that no code (C code in particular), relies on the constantness of these byte objects.
This is especially important when it comes to codecs, since the error callback logic would allow the callback to manipulate the byte object contents and length without the codec taking note of this change.
I expect there to be other places in the interpreter which would break as well.
Otherwise, you end up opening the door for segfaults and easy DOS attacks on Python3.
If the user does not need to change these bytes objects and this is needed in more places, adding an "immutable" flag for "internal" bytes objects only settable from C, or even an immutable byte base class might be an idea.
+1 I also suggest making all bytes literals immutable to avoid running into any issues like the above. -- Marc-Andre Lemburg eGenix.com Professional Python Services directly from the Source (#1, May 04 2007)
Python/Zope Consulting and Support ... http://www.egenix.com/ mxODBC.Zope.Database.Adapter ... http://zope.egenix.com/ mxODBC, mxDateTime, mxTextTools ... http://python.egenix.com/
:::: Try mxODBC.Zope.DA for Windows,Linux,Solaris,MacOSX for free ! :::: eGenix.com Software, Skills and Services GmbH Pastor-Loeh-Str.48 D-40764 Langenfeld, Germany. CEO Dipl.-Math. Marc-Andre Lemburg Registered at Amtsgericht Duesseldorf: HRB 46611