While trying to compile a nightly VTK snapshot: g++ -O2 -march=pentium3 -msse -g -Wno-deprecated \ -I/usr/X11R6/include -DvtkImaging_EXPORTS -fPIC \ -I/home/mmutz/ITK/VTK/Imaging -I/home/mmutz/ITK/vtkbuild \ -I/home/mmutz/ITK/VTK/Parallel -I/home/mmutz/ITK/VTK/Hybrid \ -I/home/mmutz/ITK/VTK/Patented \ -I/home/mmutz/ITK/VTK/Rendering -I/home/mmutz/ITK/VTK/IO \ -I/home/mmutz/ITK/VTK/Imaging -I/home/mmutz/ITK/VTK/Graphics \ -I/home/mmutz/ITK/VTK/Filtering -I/home/mmutz/ITK/VTK/Common \ -I/home/mmutz/ITK/VTK/Common/Testing/Cxx \ -I/home/mmutz/ITK/vtkbuild/Utilities/jpeg \ -I/home/mmutz/ITK/VTK/Utilities/jpeg \ -I/home/mmutz/ITK/vtkbuild/Utilities/tiff \ -I/home/mmutz/ITK/VTK/Utilities/tiff \ -I/home/mmutz/ITK/vtkbuild/Utilities/expat \ -I/home/mmutz/ITK/VTK/Utilities/expat \ -I/home/mmutz/ITK/vtkbuild/Utilities/png \ -I/home/mmutz/ITK/VTK/Utilities/png \ -I/home/mmutz/ITK/vtkbuild/Utilities/zlib -I/home/mmutz/ITK/VTK/Utilities/zlib \ -I/home/mmutz/ITK/VTK/Utilities/freetype/include \ -I/home/mmutz/ITK/vtkbuild/Utilities/freetype \ -I/home/mmutz/ITK/VTK/Utilities/ftgl/src \ -I/home/mmutz/ITK/vtkbuild/Utilities/ftgl \ -c /home/mmutz/ITK/VTK/Imaging/vtkImageLogic.cxx \ -o vtkImageLogic.o /home/mmutz/ITK/VTK/Imaging/vtkImageLogic.cxx: In function `void vtkImageLogicExecute1(vtkImageLogic*, vtkImageData*, vtkImageData*, int*, int, T*) [with T = VTK_TT]': /home/mmutz/ITK/VTK/Imaging/vtkImageLogic.cxx:240: instantiated from here /home/mmutz/ITK/VTK/Imaging/vtkImageLogic.cxx:91: could not split insn (insn:TI 422 478 296 (set (reg:SF 21 exmm0 [100]) (if_then_else:SF (ne (mem:SF (reg/v/f:SI 1 edx [77]) [46 S4 A32]) (reg:SF 21 exmm0 [100])) (reg:SF 23 exmm2) (const_double:SF 0 [0x0] 0 [0x0] 0 [0x0] 0 [0x0] 0 [0x0]))) 643 {*sse_movsfcc_const0_3} (insn_list 481 (insn_list 478 (nil))) (expr_list:REG_DEAD (reg:SF 23 exmm2) (expr_list:REG_EQUIV (mem:SF (reg/v/f:SI 0 eax [80]) [46 S4 A32]) (nil)))) /home/mmutz/ITK/VTK/Imaging/vtkImageLogic.cxx:91: Internal compiler error in final_scan_insn, at final.c:2622 Please submit a full bug report, with preprocessed source if appropriate. ---- $ g++ -v Reading specs from /usr//bin/../lib/gcc-lib/i586-mandrake-linux-gnu/3.2/specs Configured with: ../configure --prefix=/usr --libdir=/usr/lib --with-slibdir=/lib --mandir=/usr/share/man --infodir=/usr/share/info --enable-shared --enable-threads=posix --disable-checking --enable-long-long --enable-__cxa_atexit --enable-languages=c,c++,ada,f77,objc,java --host=i586-mandrake-linux-gnu --with-system-zlib Thread model: posix gcc version 3.2 (Mandrake Linux 9.0 3.2-1mdk) ---- Preprocessed source is in vtkImageLogic.ii. More examples are available on request or via nightly snapshots from VTK: http://www.vtk.org/files/nightly/vtkNightlySrc.tar.gz These two files give this internal compiler error: Imaging/vtkImageEuclideanDistance.cxx Imaging/vtkImageLogic.cxx ---- Option combinations tried: -march=pentium -O2 : OK -march=pentiumpro -O2 : OK -march=pentium3 -O : Error -march=pentium3 -O2 : Error -march=pentium3 : OK -O2 : OK -march=pentium4 -O2 : Error -march=athlon -O2 : OK -march=athlon-xp -O2 : Error -march=pentium3 -O2 -mno-sse : OK -march=pentium3 -O2 -msse : Error Release: gcc version 3.2 (Mandrake Linux 9.0 3.2-1mdk) Environment: Mandrake Linux 9.0: Linux kernel: 2.4.19-16mdk $ g++ -v Reading specs from /usr//bin/../lib/gcc-lib/i586-mandrake-linux-gnu/3.2/specs Configured with: ../configure --prefix=/usr --libdir=/usr/lib --with-slibdir=/lib --mandir=/usr/share/man --infodir=/usr/share/info --enable-shared --enable-threads=posix --disable-checking --enable-long-long --enable-__cxa_atexit --enable-languages=c,c++,ada,f77,objc,java --host=i586-mandrake-linux-gnu --with-system-zlib Thread model: posix gcc version 3.2 (Mandrake Linux 9.0 3.2-1mdk) How-To-Repeat: g++ -O -msse -o foo.o -c vtkImageLogic.ii
Fix: Use -mno-sse.
State-Changed-From-To: open->analyzed State-Changed-Why: Confirmed. Here's a much reduced testcase: ---------------------------- void foo() { float trueValue; float* outSI; if (*outSI) *outSI = trueValue; else *outSI = 0; } ---------------------------- It crashed the C and C++ front ends of 3.2/3/4 branches alike: g/a> /home/bangerth/bin/gcc-3.4-pre/bin/gcc -march=pentium3 -O -c x.c x.c: In function `foo': x.c:9: error: could not split insn (insn 34 11 35 0x4001839c (set (reg:SF 21 exmm0 [62]) (if_then_else:SF (ne (mem:SF (reg/v/f:SI 0 eax [59]) [0 S4 A32]) (reg:SF 21 exmm0 [61])) (reg/v:SF 22 exmm1 [58]) (const_double:SF 0 [0x0] 0.0 [0x0.0p+0]))) 495 {*sse_movsfcc_const0_3} (insn_list 11 (nil)) (expr_list:REG_DEAD (reg/v:SF 22 exmm1 [58]) (expr_list:REG_EQUIV (mem:SF (reg/v/f:SI 0 eax [59]) [0 S4 A32]) (nil)))) x.c:9: internal compiler error: in final_scan_insn, at final.c:2495 Please submit a full bug report, with preprocessed source if appropriate. See <URL:http://www.gnu.org/software/gcc/bugs.html> for instructions. W.
From: Dara Hazeghi <dhazeghi@yahoo.com> To: mutz@kde.org, gcc-gnats@gcc.gnu.org Cc: Subject: Re: optimization/9331: ICE in final_scan_finsn Date: Fri, 9 May 2003 22:33:00 -0700 http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit- trail&database=gcc&pr=9331 Hello, I can confirm that this bug occurs on gcc 3.2.3, but on gcc 3.3 branch and mainline (20030509) I get no ICE in the testcase with -O3 -msse. Would it be possible for the submitter of this bug to verify whether the bug still exists for them? Dara
From: Dara Hazeghi <dhazeghi@yahoo.com> To: mutz@kde.org, gcc-gnats@gcc.gnu.org Cc: Subject: Re: optimization/9331: ICE in final_scan_finsn Date: Mon, 12 May 2003 10:17:42 -0700 http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit- trail&database=gcc&pr=9331 Hello, well judging as how this can't be reproduced on either active gcc branch (3.2 is now closed), I'd vote to close this bug, but if you have a reason to keep it open, feel free to say so. Dara
State-Changed-From-To: analyzed->feedback State-Changed-Why: See Dara's question.
State-Changed-From-To: feedback->closed State-Changed-Why: Fixed in 3.3 and 3.4
From: Marc Mutz <mutz@kde.org> To: gcc-prs@gcc.gnu.org, mutz@kde.org, gcc-bugs@gcc.gnu.org, gcc-gnats@gcc.gnu.org, nobody@gcc.gnu.org Cc: Subject: Re: optimization/9331: ICE in final_scan_finsn Date: Mon, 12 May 2003 16:59:05 +0200 <snip> > Would it be possible for the submitter of this bug to verify whether > the bug still exists for them? <snip> I'm sorry, but I don't have a spare system to play with newer gcc versions. I know this is not what you expect from a bug reporter (I won't, too). http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=9331 -- I am Bush of USA. You will be pacified. Resistance is futile.
Subject: Bug 9331 CVSROOT: /cvs/gcc Module name: gcc Changes by: mark@gcc.gnu.org 2004-07-11 16:48:29 Modified files: libjava : ChangeLog libjava/java/net: URLStreamHandler.java Log message: Reported by Roman Kennke <roman@ontographics.com> (bug #9331) * java/net/URLStreamHandler.java (parseURL): When url file part doesn't contain a '/' just ignore context. Patches: http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/libjava/ChangeLog.diff?cvsroot=gcc&r1=1.2920&r2=1.2921 http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/libjava/java/net/URLStreamHandler.java.diff?cvsroot=gcc&r1=1.30&r2=1.31
Created attachment 7708 [details] preprocessed source from an x86_64 system
I'm having a problem very similar to this one on an x86_64 system, but only with 'g++ -m64', and not with -O0. Here are the details: > gcc --version gcc (GCC) 3.2.3 20030502 (Red Hat Linux 3.2.3-34) > uname -a Linux dq002.viz.acl.lanl.gov 2.4.21-15.EL #1 SMP Thu Apr 22 00:09:47 EDT 2004 x86_64 x86_64 x86_64 GNU/Linux /usr/bin/g++ -o vtkImageLogic.o -Wno-deprecated -O2 -I/usr/local/paraview/Dashboards/ MyTests/dq/zRel18-LinuxDQ-gcc-none-Release -I/usr/local/paraview/Dashboards/MyTests/dq/ zRel18-LinuxDQ-gcc-none-Release/VTK -I/usr/local/paraview/Dashboards/MyTests/dq/zRel18- LinuxDQ-gcc-none-Release/VTK/Utilities -I/usr/local/paraview/Dashboards/MyTests/ParaView-1-8/ VTK/Parallel -I/usr/local/paraview/Dashboards/MyTests/ParaView-1-8/VTK/Hybrid -I/usr/local/ paraview/Dashboards/MyTests/ParaView-1-8/VTK/Patented -I/usr/local/paraview/Dashboards/ MyTests/ParaView-1-8/VTK/Rendering -I/usr/local/paraview/Dashboards/MyTests/ParaView-1-8/ VTK/IO -I/usr/local/paraview/Dashboards/MyTests/ParaView-1-8/VTK/Imaging -I/usr/local/ paraview/Dashboards/MyTests/ParaView-1-8/VTK/Graphics -I/usr/local/paraview/Dashboards/ MyTests/ParaView-1-8/VTK/GenericFiltering -I/usr/local/paraview/Dashboards/MyTests/ParaView-1 -8/VTK/Filtering -I/usr/local/paraview/Dashboards/MyTests/ParaView-1-8/VTK/Common -I/usr/ local/paraview/Dashboards/MyTests/ParaView-1-8/VTK/Utilities -I/usr/local/paraview/Dashboards/ MyTests/ParaView-1-8/VTK/Common/Testing/Cxx -I/usr/local/paraview/Dashboards/MyTests/dq/ zRel18-LinuxDQ-gcc-none-Release/VTK/Utilities/DICOMParser -I/usr/local/paraview/Dashboards/ MyTests/ParaView-1-8/VTK/Utilities/DICOMParser -I/usr/local/paraview/Dashboards/MyTests/dq/ zRel18-LinuxDQ-gcc-none-Release/VTK/Utilities/freetype/include -I/usr/local/paraview/Dashboards/ MyTests/ParaView-1-8/VTK/Utilities/freetype/include -I/usr/local/paraview/Dashboards/MyTests/dq/ zRel18-LinuxDQ-gcc-none-Release/VTK/Utilities/freetype -I/usr/local/paraview/Dashboards/ MyTests/ParaView-1-8/VTK/Utilities/ftgl/src -I/usr/local/paraview/Dashboards/MyTests/dq/zRel18- LinuxDQ-gcc-none-Release/VTK/Utilities/ftgl -I/usr/share/doc/NVIDIA_GLX-1.0/include -I/usr/ X11R6/include -DVTK_PYTHON_BUILD -DVTK_LEAN_AND_MEAN -c /usr/local/paraview/Dashboards/ MyTests/ParaView-1-8/VTK/Imaging/vtkImageLogic.cxx /usr/local/paraview/Dashboards/MyTests/ParaView-1-8/VTK/Imaging/vtkImageLogic.cxx: In function `void vtkImageLogicExecute1(vtkImageLogic*, vtkImageData*, vtkImageData*, int*, int, T*) [with T = VTK_TT]': /usr/local/paraview/Dashboards/MyTests/ParaView-1-8/VTK/Imaging/vtkImageLogic.cxx:242: instantiated from here /usr/local/paraview/Dashboards/MyTests/ParaView-1-8/VTK/Imaging/vtkImageLogic.cxx:90: could not split insn (insn:TI 465 512 468 (set (reg:DF 21 rxmm0 [102]) (if_then_else:DF (ne (mem:DF (reg/v/f:DI 1 rdx [79]) [25 S8 A64]) (reg:DF 21 rxmm0 [102])) (reg:DF 23 rxmm2) (const_double:DF 0 [0x0] 0 [0x0] 0 [0x0]))) 660 {*sse_movdfcc_const0_3} (insn_list 515 (insn_list 512 (nil))) (expr_list:REG_DEAD (reg:DF 23 rxmm2) (expr_list:REG_EQUIV (mem:DF (reg/v/f:DI 0 rax [82]) [25 S8 A64]) (nil)))) /usr/local/paraview/Dashboards/MyTests/ParaView-1-8/VTK/Imaging/vtkImageLogic.cxx:90: Internal compiler error in final_scan_insn, at final.c:2624 Please submit a full bug report, with preprocessed source if appropriate. See <URL:http://bugzilla.redhat.com/bugzilla/> for instructions.