On Mon, 03 Dec 2007, C. Michael Pilato wrote:
> Eric Gillespie wrote:
> > firstname.lastname@example.org writes:
> >> Author: epg
> >> Date: Mon Dec 3 18:06:06 2007
> >> New Revision: 28226
> >> Log:
> >> Replace bare SubversionException implemented in C with
> >> SubversionException written in Python, exposing the full chain of
> >> svn_error_t objects.
> >> * subversion/bindings/swig/python/svn/core.py
> >> (SubversionException): Add Exception class with apr_err, message,
> >> child, file, and line attributes, plus args attribute that is
> >> compatible with pre-1.5 bindings.
> >> * subversion/bindings/swig/python/libsvn_swig_py/swigutil_py.c
> >> (SubversionException, svn_swig_py_exception_type,
> >> svn_swig_py_register_exception): Remove.
> >> (svn_swig_py_svn_exception): Turn svn_error_t chain into
> >> svn.core.SubversionException chain and raise that.
> >> * subversion/bindings/swig/python/libsvn_swig_py/swigutil_py.h
> >> (svn_swig_py_exception_type, svn_swig_py_register_exception): Remove.
> >> * subversion/bindings/swig/core.i
> >> Drop reference to svn_swig_py_exception_type.
> >> Don't register SubversionException here.
> >> * subversion/bindings/swig/python/tests/client.py
> >> * subversion/bindings/swig/python/tests/mergeinfo.py
> >> * subversion/bindings/swig/python/tests/repository.py
> >> * subversion/bindings/swig/python/tests/wc.py
> >> Import SubversionException from svn.core instead of libsvn.core.
> > Do we need backward compatibility for this? I think that, even
> > though the Python binding has no documentation, libsvn is not
> > part of the "documented" interface, but only an implementation
> > detail. No one should be importing anything from libsvn.
> I don't think we need it, but it does come awfully cheap. :-)
Seems cheap enough to stick it in there. Then again, I can't think of
anything to stop consumers from doing the same thing...
Received on Tue Dec 4 08:00:05 2007
- application/pgp-signature attachment: stored