https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63684
--- Comment #16 from Sasha Levin ---
The build works fine if I'm using a different obj directory.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63684
--- Comment #15 from Phil Muldoon ---
Can you try building out of tree to see if it fixes your case? IE:
mkdir obj
cd obj
../trunk/configure .
(replace trunk with the source directory if you are using git)
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63684
--- Comment #14 from Jakub Jelinek ---
That is weird, because actually the objdir subdir of srcdir case is what works
fine and is daily tested, and objdir == srcdir case is the problematic case
that often doesn't work.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63684
--- Comment #13 from Sasha Levin ---
>From "Installing GCC: Configuration"
(https://gcc.gnu.org/install/configure.html):
"building where srcdir == objdir should still work, but doesn't get extensive
testing"
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63684
Jakub Jelinek changed:
What|Removed |Added
CC||jakub at gcc dot gnu.org
--- Comment #12
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63684
--- Comment #11 from Sasha Levin ---
That's correct. I'm just building in-tree and using 'git clean -df' to reset
everything back to a pristine checkout state when I need to.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63684
--- Comment #10 from Phil Muldoon ---
Sorry I mean my "build is now complete"
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63684
--- Comment #9 from Phil Muldoon ---
My build is not complete with the parameters specified in the PR:
../trunk/configure --prefix=/home/pmuldoon/gdb_compile_gcc/ --with-arch=native
--with-tune=native --enable-languages=c --enable-lto --enable-g
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63684
--- Comment #8 from Phil Muldoon ---
I am still waiting for my build to complete the full-stages. Meanwhile does
--disable-libcc1 allow you to continue with your work?
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63684
--- Comment #7 from Sasha Levin ---
The built was from a clean checkout.
auto-host.h seems to be correctly created during the build. I see this after
the build failure:
$ stat host-x86_64-unknown-linux-gnu/gcc/auto-host.h
File: `host-x86_64
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63684
--- Comment #6 from Phil Muldoon ---
I am already past that point in my build. auto-host.h is an automatically
generated file, so I am quite puzzled why in your case it has not found it. In
my case, it is in obj/gcc/auto-host.h. I built with y
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63684
--- Comment #5 from Sasha Levin ---
It happened very early. I've attached the full build log for reference.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63684
--- Comment #4 from Sasha Levin ---
Created attachment 33846
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=33846&action=edit
Build log
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63684
--- Comment #3 from Phil Muldoon ---
I am trying to replicate locally. Do you recall at which stage (1, 2 .. etc)
during the build process this occurred at?
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63684
--- Comment #2 from Sasha Levin ---
Building on x86, config options are:
./configure --prefix=/home/sasha/gcc-inst/ --with-arch=native
--with-tune=native --enable-languages=c --enable-lto --enable-gold
--disable-libstdcxx
e --with-tune=native -
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63684
--- Comment #1 from Andrew Pinski ---
What target? How options did you configure GCC with?
16 matches
Mail list logo