Hi! As the testcase shows, while the grammar has attribute-list: attribute[opt] attribute-list , attribute[opt] attribute ... attribute-list , attribute ... we are actually parsing it as if the last 2 lines were attribute[opt] ... attribute-list , attribute[opt] ... and ICEing when there is no attribute. Fixed thusly, bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk/6.3?
2016-09-19 Jakub Jelinek <ja...@redhat.com> PR c++/77637 * parser.c (cp_parser_std_attribute_list): Reject ... without preceding attribute. * g++.dg/cpp0x/gen-attrs-62.C: New test. --- gcc/cp/parser.c.jj 2016-09-16 22:19:39.000000000 +0200 +++ gcc/cp/parser.c 2016-09-19 10:33:51.481904739 +0200 @@ -24220,8 +24220,12 @@ cp_parser_std_attribute_list (cp_parser if (token->type == CPP_ELLIPSIS) { cp_lexer_consume_token (parser->lexer); - TREE_VALUE (attribute) - = make_pack_expansion (TREE_VALUE (attribute)); + if (attribute == NULL_TREE) + error_at (token->location, + "expected attribute before %<...%>"); + else + TREE_VALUE (attribute) + = make_pack_expansion (TREE_VALUE (attribute)); token = cp_lexer_peek_token (parser->lexer); } if (token->type != CPP_COMMA) --- gcc/testsuite/g++.dg/cpp0x/gen-attrs-62.C.jj 2016-09-19 10:37:59.414772726 +0200 +++ gcc/testsuite/g++.dg/cpp0x/gen-attrs-62.C 2016-09-19 10:37:28.000000000 +0200 @@ -0,0 +1,5 @@ +// PR c++/77637 +// { dg-do compile { target c++11 } } + +int [[...]] a; // { dg-error "expected attribute before '...'" } +int [[,,...]] b; // { dg-error "expected attribute before '...'" } Jakub