Collision-resistant hashing

[Mark Mitchell]

>>>>> "Jonathan" == Jonathan Schilling <jls at sco.com> writes:

    Jonathan> Another thought on this is that there are currently
    Jonathan> times when users *have* to use the names linkers see for
    Jonathan> symbols, such as when passing in -B options to the
    Jonathan> linker (such as -Bsymbolic to protect a name in a
    Jonathan> dynamic library from being overridden).  If the linker
    Jonathan> symbol table name is based on a hash, it may be more
    Jonathan> difficult for users to figure out the name to pass in.
    Jonathan> The general problem is that there often aren't "mangle"
    Jonathan> utilities that linkers can use, and thus you can't pass
    Jonathan> in a real C++ name.

Yup.  Then again, typing a 2K mangled identifier to use with -B isn't
a lot of fun either!

These usability points are valid.  I'm not convinced either way, but
I'm glad we're having this discussion; that was my point in suggesting
this option.

But, nm/c++filt could have an option whereby it prints out the mapping
from demangled names to hashed names so that the user could figure out
what hashed name to use.
  
--
Mark Mitchell                   mark at codesourcery.com
CodeSourcery, LLC               http://www.codesourcery.com