[RFA] PR c++/39637

[Dodji Seketeli]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hello,

In this PR, we crash (on assert) in dependent_type_p(type), called by
cxx_sizeof_or_alignof_type because the former function expects the
processing_template_decl flag to be set when TYPE is a TEMPLATE_TYPE_PARM.

Using the more robust function uses_template_params instead of dependent_type_p
seems to address the issue.

The patch has been bootstrapped and regtested on amd64 for c++ and java.

OK for trunk ?

ps: I am not sure if it should be applied to 4.3.3 and 4.4 as it's not tagged as
a regression (yet).
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Remi - http://enigmail.mozdev.org

iEYEARECAAYFAknaGtQACgkQPejI7lrem2FkIgCeIFKJU/jvNtD8fuOPowvQ1WIW
NH8AnjZ9IDuBWWyAvlrJUFQwqzaQsCJV
=0u2o
-----END PGP SIGNATURE-----
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: PR39637-patch.txt
URL: <http://gcc.gnu.org/pipermail/gcc-patches/attachments/20090406/29868052/attachment.txt>