With LAST_UPDATED: "Sun Nov 14 05:49:38 UTC 2004" I get: FAIL: g++.dg/lookup/forscope1.C (test for excess errors) With the message in g++.log being: /gcc/testsuite/g++.dg/lookup/forscope1.C: In function 'void Go()': /gcc/testsuite/g++.dg/lookup/forscope1.C:25: error: NOTE_INSN_BASIC_BLOCK is missing for block 0 /gcc/testsuite/g++.dg/lookup/forscope1.C:25: error: NOTE_INSN_BASIC_BLOCK 8 in middle of basic block 0 /gcc/testsuite/g++.dg/lookup/forscope1.C:25: error: insn outside basic block (insn 30 7 8 0 (set (reg:DI 269) (reg:DI 259 rJ)) -1 (nil) (nil)) /gcc/testsuite/g++.dg/lookup/forscope1.C:25: internal compiler error: in rtl_verify_flow_info, at cfgrtl.c:2255 Last known to work on: "Fri Jul 9 02:34:14 GMT 2004". Known to work on unspecified 3.4 and 3.4 checkouts. Known to fail on: "Sat Sep 11 14:54:38 GMT 2004". Errors with similar appearance in the past have been flaws in the basic-block machinery.
g++.old-deja/g++.niklas/t132.C and g++.old-deja/g++.other/singleton.C of the same test-run seem to be the same bug judging by the g++.log contents.
Confirmed. This is exception handling related.
MMIX bugs are low-priority.
MMIX is not a primary or secondary platform; removing target milestone.
The same problem occurs on alpha-dec-osf5.1b as of 20050627: FAIL: g++.old-deja/g++.other/singleton.C (test for excess errors) Excess errors: /vol/gnu/src/gcc/gcc-dist/gcc/testsuite/g++.old-deja/g++.other/singleton.C:38: error: NOTE_INSN_BASIC_BLOCK is missing for block 0 /vol/gnu/src/gcc/gcc-dist/gcc/testsuite/g++.old-deja/g++.other/singleton.C:38: error: NOTE_INSN_BASIC_BLOCK 7 in middle of basic block 0 /vol/gnu/src/gcc/gcc-dist/gcc/testsuite/g++.old-deja/g++.other/singleton.C:38: error: insn outside basic block (insn 54 6 7 0 (set (mem:DI (reg/f:DI 65 virtual-stack-vars) [0 S8 A64]) (reg:DI 29 $29)) -1 (nil) (nil)) /vol/gnu/src/gcc/gcc-dist/gcc/testsuite/g++.old-deja/g++.other/singleton.C:38: internal compiler error: in rtl_verify_flow_info, at cfgrtl.c:2234
Removing target milestone per Comment #4.
Also seen for new test g++.dg/tree-ssa/pr23164.C /home/hp/combined/combined/gcc/testsuite/g++.dg/tree-ssa/pr23164.C: In function 'void g()':^M /home/hp/combined/combined/gcc/testsuite/g++.dg/tree-ssa/pr23164.C:16: error: NOTE_INSN_BASIC_BLOCK is missing for block 0^M /home/hp/combined/combined/gcc/testsuite/g++.dg/tree-ssa/pr23164.C:16: error: NOTE_INSN_BASIC_BLOCK 12 in middle of basic block 0\ ^M /home/hp/combined/combined/gcc/testsuite/g++.dg/tree-ssa/pr23164.C:16: error: insn outside basic block^M (insn 405 11 12 0 (set (reg:DI 284)^M (reg:DI 259 rJ)) 3 {movdi} (nil)^M (nil))^M /home/hp/combined/combined/gcc/testsuite/g++.dg/tree-ssa/pr23164.C:16: internal compiler error: in rtl_verify_flow_info, at cfgrtl.c:2255^M
Will not be fixed in 4.1.1; adjust target milestone to 4.1.2.
Control flow graph handling has changed significantly in GCC 4.3. Could someone try this with a recent snapshot from the trunk? If the bug still exists, please assign it to me and provide me with some instructions on how to reproduce it with a cross-compiler.
These tests do not fail and no gcc or g++ test fail with the same error with r130591, so best to close this one. FWIW, instructions are at simtest-howto.html (I've given personal instructions at least once).