$ g++ -g -O2 -o StdAfx.h.pch StdAfx.h $ g++ -g -O2 configfile.cpp StdAfx.h: In static member function `static size_t std::char_traits<wchar_t>::length(const wchar_t*)': StdAfx.h:15161: instantiated from `std::basic_string<_CharT, _Traits, _Alloc> std::operator+(const _CharT*, const std::basic_string<_CharT, _Traits, _Alloc>&) [with _CharT = wchar_t, _Traits = std::char_traits<wchar_t>, _Alloc = std::allocator<wchar_t>]' configfile.cpp:19: instantiated from here StdAfx.h:15161: internal compiler error: in gen_subprogram_die, at dwarf2out.c: 10752 Doesn't ICE when not using pch. I've attached the StdAfx.h file that includes libstdc++ headers and the StdAfx.ii file without includes. sed 's/^#.*$//' StdAfx.ii > StdAfx.h give a StdAfx.h file that also provokes the ICE. Release: gcc version 3.4 20030113 (experimental) Environment: i686-pc-linux-gnu How-To-Repeat: mkdir tmp cd tmp tar xfj gen_subprogram_die.tar.bz2 g++ -g -O2 -o StdAfx.h.pch StdAfx.h g++ -g -O2 configfile.cpp
State-Changed-From-To: open->feedback State-Changed-Why: I get a lot of weird error messages about "File number XX already used" from a snapshot from 2003-01-12, but no ICE. I don't get these error messages, no ICE, but several link failures from a recent snapshot (2003-01-30), but as far as I can see the functions that the compiler complains about are really not there (see below for output). Could you possibly retry and see whether this now works for you? Thanks W. test/x> ../../bin/gcc-2003-01-30/bin/c++ -g -O2 configfile.cpp /usr/lib/crt1.o: In function `_start': /usr/lib/crt1.o(.text+0x18): undefined reference to `main' /usr/bin/ld: Dwarf Error: Invalid or unhandled FORM value: 14. /usr/bin/ld: Dwarf Error: Invalid or unhandled FORM value: 14. /usr/bin/ld: Dwarf Error: Invalid or unhandled FORM value: 14. /tmp/ccaXYMqI.o: In function `CFEntryNotFoundException::CFEntryNotFoundException[not-in-charge]()': :5: undefined reference to `Exception::Exception[not-in-charge]()' /tmp/ccaXYMqI.o: In function `CFEntryNotFoundException::CFEntryNotFoundException[in-charge]()': :5: undefined reference to `Exception::Exception[not-in-charge]()' /tmp/ccaXYMqI.o: In function `CFParseException::CFParseException[not-in-charge](std::basic_string<wchar_t, std::char_traits<wchar_t>, std::allocator<wchar_t> > const&)': :11: undefined reference to `Exception::Exception[not-in-charge]()' /tmp/ccaXYMqI.o: In function `CFParseException::CFParseException[in-charge](std::basic_string<wchar_t, std::char_traits<wchar_t>, std::allocator<wchar_t> > const&)': :11: undefined reference to `Exception::Exception[not-in-charge]()' /tmp/ccaXYMqI.o: In function `ConfigFile::ConfigFile[not-in-charge](std::basic_string<char, std::char_traits<char>, std::allocator<char> > const&)': :18: undefined reference to `s2w(std::basic_string<char, std::char_traits<char>, std::allocator<char> > const&)' /tmp/ccaXYMqI.o: In function `ConfigFile::ConfigFile[in-charge](std::basic_string<char, std::char_traits<char>, std::allocator<char> > const&)': :18: undefined reference to `s2w(std::basic_string<char, std::char_traits<char>, std::allocator<char> > const&)' collect2: ld returned 1 exit status
From: Ritzert@t-online.de (Michael Ritzert) To: bangerth@dealii.org, gcc-bugs@gcc.gnu.org, gcc-prs@gcc.gnu.org, nobody@gcc.gnu.org, gcc-gnats@gcc.gnu.org Cc: Subject: Re: c++/9304: [pch] ICE in gen_subprogram_die Date: Mon, 3 Feb 2003 19:19:38 +0100 On the unstripped code (confidential) I now get [loads of warnings] configfile.cpp:46: internal compiler error: in write_unscoped_name, at cp/mangle.c:791 I will try to strip it down again so that this particular error remains as I find the time to do so. Michael
From: Wolfgang Bangerth <bangerth@ticam.utexas.edu> To: Michael Ritzert <Ritzert@t-online.de> Cc: gcc-gnats@gcc.gnu.org Subject: Re: c++/9304: [pch] ICE in gen_subprogram_die Date: Tue, 4 Feb 2003 08:38:14 -0600 (CST) > I will see what I am doing wrong and report back. I guess you do, but can you also say whether this is a recent snapshot, or from which date it stems? Maybe the fact that I don't see the problem comes from my using a newer (and fixed) version. Thanks W. ------------------------------------------------------------------------- Wolfgang Bangerth email: bangerth@ticam.utexas.edu www: http://www.ticam.utexas.edu/~bangerth/
From: Ritzert@t-online.de (Michael Ritzert) To: bangerth@dealii.org, gcc-bugs@gcc.gnu.org, gcc-prs@gcc.gnu.org, nobody@gcc.gnu.org, gcc-gnats@gcc.gnu.org Cc: Subject: Re: c++/9304: [pch] ICE in gen_subprogram_die Date: Tue, 4 Feb 2003 12:03:36 +0100 Am Montag, 3. Februar 2003 19:19 schrieb Michael Ritzert: > On the unstripped code (confidential) I now get > > [loads of warnings] > configfile.cpp:46: internal compiler error: in write_unscoped_name, at > cp/mangle.c:791 Hmm. There were so many warning that I oversaw this at the very beginning: /usr/include/sys/cdefs.h:31:3: #error "You need a ISO C conforming compiler to use the glibc headers" with cdefs.h:31 being #if defined __GNUC__ && !defined __STDC__ # error "You need a ISO C conforming compiler to use the glibc headers" #endif I will see what I am doing wrong and report back. Michael
From: Ritzert@t-online.de (Michael Ritzert) To: Wolfgang Bangerth <bangerth@ticam.utexas.edu> Cc: gcc-gnats@gcc.gnu.org Subject: Re: c++/9304: [pch] ICE in gen_subprogram_die Date: Tue, 4 Feb 2003 19:38:50 +0100 On February 4th 2003 15:38 schrieb Wolfgang Bangerth: > > I will see what I am doing wrong and report back. > > I guess you do, but can you also say whether this is a recent snapshot, or > from which date it stems? Reading specs from /opt/gcc-HEAD/lib/gcc-lib/i686-pc-linux-gnu/3.4/specs Configured with: /home/ritzert/gcc/HEAD/gcc/configure --enable-threads=posix --prefix=/opt/gcc-HEAD Thread model: posix gcc version 3.4 20030202 (experimental) I looked into it briefly, but I can't find anything obvious I'm doing wrong. - I'm compiling the header and the actual source file with exactly the same flags. - When I don't use pch the error disappears. - When I try to reproduce with preprocessed sources where I replace the content of the header in question with #include <cdefs.h>, I can't reproduce it. So all I can say ATM is that "strange things happen"... I will compile today's CVS over night and test again. Michael
From: Ritzert@t-online.de (Michael Ritzert) To: Wolfgang Bangerth <bangerth@ticam.utexas.edu> Cc: gcc-gnats@gcc.gnu.org Subject: Re: c++/9304: [pch] ICE in gen_subprogram_die Date: Tue, 4 Feb 2003 19:53:32 +0100 To be 100% clear, I should add: - I'm not talking about the testcase in the bugreport, but about the original sourcefile the testcase was extracted from. - Compiling other sourcefiles against the very same PCH using the same Makefile works just fine. Michael
From: Ritzert@t-online.de (Michael Ritzert) To: Wolfgang Bangerth <bangerth@ticam.utexas.edu> Cc: gcc-gnats@gcc.gnu.org Subject: Re: c++/9304: [pch] ICE in gen_subprogram_die Date: Tue, 4 Feb 2003 22:45:51 +0100 Ok, some more info: StdAfx.h: empty(!) configfile.cpp: #include "StdAfx.h" #include <iostream> ============================================================================ i686-pc-linux-gnu 20030202 ----------------- ritzert@liz:~/NFS/test/pch$ c++ -g -O2 -fPIC -c StdAfx.h ritzert@liz:~/NFS/test/pch$ c++ -g -O2 -fPIC -c configfile.cpp In file included from /opt/gcc-HEAD/include/c++/3.4/iostream:46, from configfile.cpp:2: /opt/gcc-HEAD/include/c++/3.4/istream:767: internal compiler error: Segmentation fault Please submit a full bug report, with preprocessed source if appropriate. See <URL:http://www.gnu.org/software/gcc/bugs.html> for instructions. ritzert@liz:~/NFS/test/pch$ c++ -g -O2 -fPIC -pthread -c configfile.cpp In file included from /opt/gcc-HEAD/include/c++/3.4/iostream:46, from configfile.cpp:3: /opt/gcc-HEAD/include/c++/3.4/istream:767: internal compiler error: Segmentation fault Please submit a full bug report, with preprocessed source if appropriate. See <URL:http://www.gnu.org/software/gcc/bugs.html> for instructions. ============================================================================ i386-unknown-freebsd4.7 20030203: ----------------------- -bash-2.05b$ g++ -fPIC -c StdAfx.h -bash-2.05b$ g++ -fPIC -c configfile.cpp 2>&1 | head In file included from /home/ritzert/gcc-HEAD/install/include/c++/3.4/ios:50, from /home/ritzert/gcc-HEAD/install/include/c++/3.4/ostream:45, from /home/ritzert/gcc-HEAD/install/include/c++/3.4/iostream:45, from configfile.cpp:3: /home/ritzert/gcc-HEAD/install/include/c++/3.4/streambuf:-23783: internal compiler error: Segmentation fault Please submit a full bug report, with preprocessed source if appropriate. See <URL:http://www.gnu.org/software/gcc/bugs.html> for instructions. ============================================================================ alpha-unknown-linux-gnu 20030131 - works ----------------------- Michael
Responsible-Changed-From-To: unassigned->geoffk Responsible-Changed-Why: Author of PCH
State-Changed-From-To: feedback->open State-Changed-Why: I think we must leave this to the author of PCH. I still don't see the ICE on my system...
From: Ritzert@t-online.de (Michael Ritzert) To: Wolfgang Bangerth <bangerth@ticam.utexas.edu> Cc: gcc-gnats@gcc.gnu.org Subject: Re: c++/9304: [pch] ICE in gen_subprogram_die Date: Tue, 4 Feb 2003 23:20:06 +0100 The bootstrap of gcc HEAD just finished and now I get for the testcase in the bugreport: ritzert@liz:~/tmp$ g++ -v Reading specs from /opt/gcc-HEAD/lib/gcc-lib/i686-pc-linux-gnu/3.4/specs Configured with: /home/ritzert/gcc/HEAD/gcc/configure --enable-threads=posix --prefix=/opt/gcc-HEAD Thread model: posix gcc version 3.4 20030204 (experimental) ritzert@liz:~/tmp$ g++ -g -O2 -o StdAfx.h.pch StdAfx.h ritzert@liz:~/tmp$ g++ -g -O2 -o configfile.lo configfile.cpp /opt/gcc-HEAD/include/c++/3.4/bits/basic_string.tcc: In static member function `static size_t std::char_traits<wchar_t>::length(const wchar_t*)': /opt/gcc-HEAD/include/c++/3.4/bits/basic_string.tcc:622: instantiated from `std::basic_string<_CharT, _Traits, _Alloc> std::operator+(const _CharT*, const std::basic_string<_CharT, _Traits, _Alloc>&) [with _CharT = wchar_t, _Traits = std::char_traits<wchar_t>, _Alloc = std::allocator<wchar_t>]' configfile.cpp:19: instantiated from here /opt/gcc-HEAD/include/c++/3.4/bits/basic_string.tcc:622: internal compiler error: in gen_subprogram_die, at dwarf2out.c:10784 Please submit a full bug report, with preprocessed source if appropriate. See <URL:http://www.gnu.org/software/gcc/bugs.html> for instructions. Seems I plain forgot to test the initial testcase earlier... Michael PS: On FreeBSD 4.7, I get for the testcase with s/wstring/string/, s/L"/"/: [...] configfile.cpp:15: instantiated from here /home/ritzert/gcc-HEAD/install/include/c++/3.4/bits/stl_construct.h:125: internal compiler error: Segmentation fault Please submit a full bug report, with preprocessed source if appropriate. See <URL:http://www.gnu.org/software/gcc/bugs.html> for instructions. Looks like yet another different error...
State-Changed-From-To: open->suspended State-Changed-Why: I don't have access to hardware to reproduce this problem; Darwin has neither wide character support nor DWARF2. Perhaps the bug will turn up for someone else later.
From: Michael Ritzert <michael.ritzert@realtech.de> To: gcc-gnats@gcc.gnu.org, geoffk@gcc.gnu.org Cc: gcc-bugs@gcc.gnu.org Subject: Re: c++/9304: [pch] ICE in gen_subprogram_die Date: Wed, 5 Feb 2003 11:13:30 +0100 Important info: The error on the testcase in the bugreport only appears o= n=20 Debian GNU/Linux unstable, but not on SuSE Linux 7.3 (using identical=20 versions of gcc). So if you can't reproduce it, it is likely due to a=20 different environment. If somebody wants to try and reproduce the error, just ask me in private = mail=20 to produce another preprocessed testcase. For some reason I can no longer= =20 compile the one in the bugreport (StdAfx.h:18891: error:=20 `_M_c_locale_codecvt' has not been declared) - propably because of the=20 stricter parser. Michael
From: Geoffrey Keating <geoffk@apple.com> To: gcc-gnats@gcc.gnu.org, geoffk@gcc.gnu.org, ritzert@t-online.de, gcc-bugs@gcc.gnu.org, gcc-prs@gcc.gnu.org Cc: Subject: Re: c++/9304: [pch] ICE in gen_subprogram_die Date: Tue, 22 Apr 2003 10:28:16 -0700 Does this bug still happen with top-of-tree? There have been several DWARF-and-PCH-related changes since this bug was filed. -- Geoff Keating <geoffk@apple.com>
State-Changed-From-To: suspended->feedback State-Changed-Why: Waiting for confirmation the bug still exists.
State-Changed-From-To: feedback->closed State-Changed-Why: From: Ritzert@t-online.de Date: Wed Apr 23, 2003 3:30:23 AM US/Pacific Hi, I was just about to report this bug closed. I had a testcase for the report in my tree that did not fail tonight. I guess this patch fixed it: 2003-04-22 Geoffrey Keating <geoffk@apple.com> Loren James Rittle <ljrittle@acm.org> * dwarf2out.c (fde_table_allocated): Mark with GTY. [...] Thanks, Michael