Installing GCC 13.1.0

[Jackalope]

I've just noticed that my 32-bit Debian virtual machine doesn't boot after I've updated VirtualBox to the version 7.0.6. I get the following error:

I've tried to re-install Debian from CD, but the installer also crashes with the same error message. What should I do?

Undo the thing that you did before it broke (back out your VirtualBox update).

In anticipation of your next question, if you want to know how to fix the kernel panic your OS vendor ando virtualization provider are over there. ---->

If it were me, I'd start by reading VirtualBox release notes for the new version and every version you skipped and see if anything of use is in them. If not, go back to my first suggestion.

[Nay_Sayer]

This feels like a good time to relate a little of my life story to you FA.

When I went to college it was because I wanted to be the first in my immediate family to do so.

So I picked psychology because how the human brain works fascinates me. Now some areas I did very well like 'interpersonal communication' but in others like ' Statistics' I was terrible. I knew right then I had hit my limit and I was ok with that, It Didn't hurt to try.

The point here is if I had wasted time pursuing it I may never have found the job I do now, Which I really enjoy. I'm not telling you to give up but instead, maybe refocus your drive elsewhere so you can flourish.

[FlatAssembler]

It tells you the problem:

"recompile with -fPIC"

Google what that means.  I've always used it in my UNIX or LINUX compiles.  It doesn't make sense not to use it.

I've tried to do the following before running `configure`:

Code:

# Downgrade GCC to the one that comes with Debian.
export CC=/usr/bin/gcc
export CXX=/usr/bin/g++
# Tell "configure" to always add "-fPIC" when invoking GCC
export CFLAGS="-fPIC"
export CXXFLAGS=$CFLAGS

I still get exactly the same error. And I think I understand why:
1. The error occurs at the Stage 2 of the compilation of GCC. That means the GCC being invoked isn't the GCC you have already installed on your system, but the newly-built GCC 13.1. It's the `xgcc`, the minimal GCC used to compile the rest of GCC. It shouldn't matter which version of GCC was used to produce `xgcc` in the Stage 1 (unless you assume there is some weird wrong-code bug).
2. Take a look at the invocation that occurs right before the error:

Code:

libtool: link:  /home/teo/gcc-objdir/./gcc/xgcc -shared-libgcc -B/home/teo/gcc-objdir/./gcc -nostdinc++ -L/home/teo/gcc-objdir/x86_64-pc-linux-gnu/libstdc++-v3/src -L/home/teo/gcc-objdir/x86_64-pc-linux-gnu/libstdc++-v3/src/.libs -L/home/teo/gcc-objdir/x86_64-pc-linux-gnu/libstdc++-v3/libsupc++/.libs -B/usr/local/x86_64-pc-linux-gnu/bin/ -B/usr/local/x86_64-pc-linux-gnu/lib/ -isystem /usr/local/x86_64-pc-linux-gnu/include -isystem /usr/local/x86_64-pc-linux-gnu/sys-include     -fPIC -DPIC -shared -nostdlib /usr/lib/x86_64-linux-gnu/crti.o /home/teo/gcc-objdir/./gcc/crtbeginS.o  .libs/hwasan_allocation_functions.o .libs/hwasan_allocator.o .libs/hwasan.o .libs/hwasan_dynamic_shadow.o .libs/hwasan_exceptions.o .libs/hwasan_fuchsia.o .libs/hwasan_globals.o .libs/hwasan_interceptors.o .libs/hwasan_interceptors_vfork.o .libs/hwasan_linux.o .libs/hwasan_memintrinsics.o .libs/hwasan_new_delete.o .libs/hwasan_poisoning.o .libs/hwasan_report.o .libs/hwasan_setjmp_aarch64.o .libs/hwasan_setjmp_x86_64.o .libs/hwasan_tag_mismatch_aarch64.o .libs/hwasan_thread.o .libs/hwasan_thread_list.o .libs/hwasan_type_test.o  -Wl,--whole-archive ../sanitizer_common/.libs/libsanitizer_common.a ../interception/.libs/libinterception.a ../libbacktrace/.libs/libsanitizer_libbacktrace.a -Wl,--no-whole-archive  -Wl,-rpath -Wl,/home/teo/gcc-objdir/x86_64-pc-linux-gnu/libstdc++-v3/src/.libs -Wl,-rpath -Wl,/usr/local/lib/../lib64 -L/home/teo/gcc-objdir/x86_64-pc-linux-gnu/libstdc++-v3/src/.libs -L/home/teo/gcc-objdir/x86_64-pc-linux-gnu/libstdc++-v3/src -L/home/teo/gcc-objdir/x86_64-pc-linux-gnu/libstdc++-v3/libsupc++/.libs -ldl -lrt -lpthread ../../libstdc++-v3/src/.libs/libstdc++.so -lm -L/home/teo/gcc-objdir/./gcc -L/lib/x86_64-linux-gnu -L/lib/../lib64 -L/usr/lib/x86_64-linux-gnu -lc -lgcc_s /home/teo/gcc-objdir/./gcc/crtendS.o /usr/lib/x86_64-linux-gnu/crtn.o  -mshstk   -Wl,-soname -Wl,libhwasan.so.0 -o .libs/libhwasan.so.0.0.0
/usr/bin/ld: .libs/hwasan.o: relocation R_X86_64_PC32 against undefined symbol `__ehdr_start' can not be used when making a shared object; recompile with -fPIC

The GNU Make does pass the `-fPIC` to the `xgcc`, it's just apparently being ignored.

[HappySkeptic]

-fPIC must be invoked on creation of the .o files (i.e. hwasan.o) , not just at the link stage.

[FlatAssembler]

-fPIC must be invoked on creation of the .o files (i.e. hwasan.o) , not just at the link stage.

Well, there doesn't seem to be a simple solution. Telling `configure` to use a known-good compiler (that comes with Debian) or to add `-fPIC` to every invocation to the compiler doesn't solve the problem, in fact, it doesn't even address it (the error stays the same).
The error seems to occur in the new part of GCC related to static analysis, which it is basically copying from LLVM. Is there a way to tell `configure` I don't want a static analyzer, that I just want the compiler? I mainly want to know whether my C++ programs compile using GCC 13.1.0, I don't need a static analyzer for that.

On Windows, I have Docker installed. I guess I will to install the GCC 13.1.0 image from the Docker hub and experiment with GCC 13.1.0 using Docker. After all, that's how I came to know about GCC 13.1.0, by the GitLab CI/CD of my AEC-to-WebAssembly compiler failing, and GitLab uses GCC through Docker. I once used to use the latest GCC through Docker. It sucks because sharing files between Docker containers and your host machine is difficult, but I guess I will have to.

I am currently trying to build GCC 13.1.0 on CygWin, I think it has already gone to the Stage 3 (so, that error with `-fPIC` doesn't occur on CygWin, but maybe some other will).

[FlatAssembler]

Just managed to install GCC 13.1.0 on CygWin, but I cannot reproduce the error that GitLab claims occurs when compiling my compiler.

[arewethereyet]

Wow, thanks for the update!

[FlatAssembler]

This feels like a good time to relate a little of my life story to you FA.

When I went to college it was because I wanted to be the first in my immediate family to do so.

So I picked psychology because how the human brain works fascinates me. Now some areas I did very well like 'interpersonal communication' but in others like ' Statistics' I was terrible.  I knew right then I had hit my limit and I was ok with that, It Didn't hurt to try.

The point here is if I had wasted time pursuing it I may never have found the job I do now, Which I really enjoy.  I'm not telling you to give up but instead, maybe refocus your drive elsewhere so you can flourish.

Then both of us were killed by mathematics, right? You were killed by the statistics, while I was killed by Signals and Systems and its successor Control Engineering.

I am refocusing my drive elsewhere. I am publishing papers about the names of places.

[BrianSoddingBoru4]

This feels like a good time to relate a little of my life story to you FA.

When I went to college it was because I wanted to be the first in my immediate family to do so.

So I picked psychology because how the human brain works fascinates me. Now some areas I did very well like 'interpersonal communication' but in others like ' Statistics' I was terrible.  I knew right then I had hit my limit and I was ok with that, It Didn't hurt to try.

The point here is if I had wasted time pursuing it I may never have found the job I do now, Which I really enjoy.  I'm not telling you to give up but instead, maybe refocus your drive elsewhere so you can flourish.

Then both of us were killed by mathematics, right? You were killed by the statistics, while I was killed by Signals and Systems and its successor Control Engineering.

I am refocusing my drive elsewhere. I am publishing papers about the names of places.

You should do Muckanaghederdauhaulia next.

Boru

[FlatAssembler]

Then both of us were killed by mathematics, right? You were killed by the statistics, while I was killed by Signals and Systems and its successor Control Engineering.

I am refocusing my drive elsewhere. I am publishing papers about the names of places.

You should do Muckanaghederdauhaulia next.

Boru

Which journal in Croatia do you think would publish a paper about an Irish placename? Especially a morphologically transparent one, as, ccording to Wikepedia, it means "pig-marsh between two sea inlets". I have no idea what to write about it.