http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52388
Bug #: 52388 Summary: binutils-2.21 and newer break ACATS on arm-linux-gnueabi due to widechr compile failure Classification: Unclassified Product: gcc Version: 4.7.0 Status: UNCONFIRMED Severity: normal Priority: P3 Component: ada AssignedTo: unassig...@gcc.gnu.org ReportedBy: mi...@it.uu.se On armv5tel-linux-gnueabi, if I let PATH refer to binutils-2.21 or newer, then the ACATS test suite disappears from the final summary report. Inspecting gcc/testsuite/ada/acats/acats.log shows ... splitting /mnt/scratch/objdir46/gcc/testsuite/ada/acats/support/impdefh.a into: impdef-annex_h.ads splitting /mnt/scratch/objdir46/gcc/testsuite/ada/acats/support/widechr.a into: widechr.adb **** Failed to compile widechr and nothing else. This happens with every gcc branch (4.4 -> 4.7), and binutils-2.21, 2.22, and cvs head as of early December 2011. Reverting to binutils-2.20.1 eliminates the above error and gives me clean ACATS results. The "gnat" test results look the same regardless of the binutils version. I have traced the regression to one specific change in binutils (http://sourceware.org/ml/binutils/2010-05/msg00074.html), commited as part of a group of changes (http://sourceware.org/ml/binutils-cvs/2010-05/msg00035.html), but I don't see anything obviously wrong with it. Before I report this to the sourceware bugzilla, is there any way to get a better error diagnostic from the widechr compilation attempt?