Fundef preliminary freed
Bugzilla Link | 169 |
Created on | Dec 05, 2005 23:02 |
Resolution | FIXED |
Resolved on | Dec 07, 2005 14:00 |
Version | 1.00beta |
OS | Linux |
Architecture | PC |
Attachments | tutu2.sac |
Extended Description
I'm not sure if this is my doing or sac2c having fun. 1. I changed ArrayArith.sac RELationals from NUM to BUILT_IN, to get _neq_ and _eq_ defined on char arrays. 2. The attached dtb2.sac fails in several mysterious ways: a. sac2c -noOPT generates code that produces an invalid pointer on free(m), followed by a run-time error on the above code, unable to find a char[*]-> char[*]-> Bool[*] != function. Likely my doing, but I don't see why it waits so long to complain. b. sac2c -O3 offers a vague complaint about negative shape vector elements, probably in the area of APEXReshape. I'm chasing this to inline the offending macro and see if I can suss it out. sac2c rev#14459