Re: analysis of libiberty/cp-demangle.c for libsupc++/__cxa_demangle

[Ian Lance Taylor <ian at wasabisystems dot com>]

Benjamin Kosnik <bkoz@redhat.com> writes:

> FAIL: demangle/regression/cw-11.cc execution test
> string: _X11TransParseAddress
> libiberty: _X11TransParseAddress, -2: invalid mangled name
> libstdc++: _X11TransParseAddress
> 
> Same: this is twisted, but not invalid.

Hmmm.  I'm not really sure what the issue is here.  That is not a
valid mangled name.  So my version of the function returns failure
with a status of -2.  What is it supposed to return?

The ABI says ``The return value is a pointer to a null-terminated
array of characters, the demangled name.''  And it says that for an
error in demangling, -2 means an ``invalid mangled name.''  And the
mangled-name argument is required to be ``either an external name,
i.e. with a "_Z" prefix, or an internal NTBS mangling, e.g. of a type
for type_info.''

So by my reading, when given something like _X11TransParseAddress
which is not a valid mangled name, it is correct for __cxa_demangle()
to return NULL with status set to -2.

What other reading could there be?

Note that it necessarily matters, but I note that my interpretation is
the same as the historical cplus_demangle() interface in libiberty;
when given a name which can not be interpreted as a mangled name, it
returns NULL.

Ian