I think my last note is a red herring: the revision which actually causes the failure
is svn r 5147, which doesn't involve the same files as svn r5140.
The 'tview' string in my version of the traceback is definitely a corrupt version of
the affected file: it doesn't match any revision of the file in subversion at all, and
in fact contains a Python syntax error.
The 'tview_len' value corresponds exactly to the length of the file as checked in
to r5147.
I think my last note is a red herring: the revision which actually causes the failure
is svn r 5147, which doesn't involve the same files as svn r5140.
The 'tview' string in my version of the traceback is definitely a corrupt version of
the affected file: it doesn't match any revision of the file in subversion at all, and
in fact contains a Python syntax error.
The 'tview_len' value corresponds exactly to the length of the file as checked in
to r5147.