ruby-changes:32455
From: usa <ko1@a...>
Date: Thu, 9 Jan 2014 17:42:49 +0900 (JST)
Subject: [ruby-changes:32455] usa:r44534 (trunk): * hash.c (rb_any_hash): should treat the return value of rb_objid_hash()
usa 2014-01-09 17:42:42 +0900 (Thu, 09 Jan 2014) New Revision: 44534 http://svn.ruby-lang.org/cgi-bin/viewvc.cgi?view=rev&revision=44534 Log: * hash.c (rb_any_hash): should treat the return value of rb_objid_hash() as `long', because ruby assumes the object id of an object is `long'. this fixes test failures on mswin64 introduced at r44525. Modified files: trunk/ChangeLog trunk/hash.c Index: ChangeLog =================================================================== --- ChangeLog (revision 44533) +++ ChangeLog (revision 44534) @@ -1,3 +1,9 @@ https://github.com/ruby/ruby/blob/trunk/ChangeLog#L1 +Thu Jan 9 17:40:28 2014 NAKAMURA Usaku <usa@r...> + + * hash.c (rb_any_hash): should treat the return value of rb_objid_hash() + as `long', because ruby assumes the object id of an object is `long'. + this fixes test failures on mswin64 introduced at r44525. + Thu Jan 9 09:55:20 2014 Aaron Patterson <aaron@t...> * ext/psych/lib/psych/visitors/yaml_tree.rb: dumping strings with Index: hash.c =================================================================== --- hash.c (revision 44533) +++ hash.c (revision 44534) @@ -132,7 +132,8 @@ rb_any_hash(VALUE a) https://github.com/ruby/ruby/blob/trunk/hash.c#L132 if (SPECIAL_CONST_P(a)) { if (a == Qundef) return 0; - hnum = rb_objid_hash((st_index_t)a); + /* assume hnum is long, so need to drop upper dword on LLP64. */ + hnum = (long)rb_objid_hash((st_index_t)a); } else if (BUILTIN_TYPE(a) == T_STRING) { hnum = rb_str_hash(a); -- ML: ruby-changes@q... Info: http://www.atdot.net/~ko1/quickml/