[Bug c++/33962] ICE at call to overloaded template function with variable-length function argument list

2007-11-16 Thread rob dot quill at gmail dot com
--- Comment #1 from rob dot quill at gmail dot com 2007-11-16 15:03 --- Hey, I can confirm this and am trying to investigate the problem. It seems that more_specialized_fn() loops over the arguments to the first overloaded function, where in this case there are two, but the tree chain

[Bug tree-optimization/33092] [4.3 Regrsssion] Using -O1 -fno-tree-salias results in ICE

2007-11-16 Thread rob dot quill at gmail dot com
--- Comment #6 from rob dot quill at gmail dot com 2007-11-16 20:00 --- I'm working on adding a new pass for this. Rob -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33092

[Bug tree-optimization/33092] [4.3 Regrsssion] Using -O1 -fno-tree-salias results in ICE

2007-11-17 Thread rob dot quill at gmail dot com
--- Comment #7 from rob dot quill at gmail dot com 2007-11-17 19:27 --- I believe I have this done, just checking there are no regressions. Rob -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=33092

[Bug tree-optimization/33092] [4.3 Regrsssion] Using -O1 -fno-tree-salias results in ICE

2007-11-19 Thread rob dot quill at gmail dot com
--- Comment #8 from rob dot quill at gmail dot com 2007-11-19 13:46 --- Created an attachment (id=14581) --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=14581&action=view) Patch adds a dummy pass, as discussed in comments Patch attached, no regressions on x86_64. --