Death in EMRCOprf due to N_id vs. N_num
| Bugzilla Link | 1112 |
| Created on | Jan 24, 2014 20:29 |
| Version | svn |
| OS | Linux |
| Architecture | PC |
| Attachments | bug1107.sac |
Extended Description
Created an attachment (id=1000) source code to reproduce fault I do not know how long this has been broken... sac2c bug1107.sac **** Reducing reference counting instructions ... **** Optimizing reference counting instructions ... TRAVERSE ERROR: node of type 31:N_id found where 32:N_num was expected! Things go wrong in EMRCOprf, in the F_inc_rc block, here: NUM_VAL( PRF_ARG1( alloc)) += NUM_VAL( PRF_ARG2( arg_node)); I think that PRF_ARG1( alloc) should be PRF_ARG3( alloc). arg_node is the F_inc_rc; alloc is an F_suballoc. I would like someone to confirm this before I check in my one-character fix for this. I would like someone to tell me why this thing doesn't break more often...
To upload designs, you'll need to enable LFS and have an admin enable hashed storage. More information