Created attachment 547756 [details] build log file Hi, I am having issues with latest evolution data server instalation. The stack from log file is as following: [ 57%] ^[[32m^[[1mLinking CXX shared library libebackend-1.2.so^[[0m cd /var/tmp/portage/gnome-extra/evolution-data-server-3.26.6/work/evolution-data-server-3.26.6_build/src/libebackend && /usr/bin/cmake -E cmake_link_scrip$ /usr/bin/x86_64-pc-linux-gnu-g++ -fPIC -march=sandybridge -mmmx -mno-3dnow -msse -msse2 -msse3 -mssse3 -mno-sse4a -mcx16 -msahf -mno-movbe -maes -m$ Traceback (most recent call last): File "/usr/bin/g-ir-scanner", line 66, in <module> sys.exit(scanner_main(sys.argv)) File "/usr/lib64/gobject-introspection/giscanner/scannermain.py", line 529, in scanner_main transformer = create_transformer(namespace, options) File "/usr/lib64/gobject-introspection/giscanner/scannermain.py", line 401, in create_transformer transformer.register_include_uninstalled(include_path) File "/usr/lib64/gobject-introspection/giscanner/transformer.py", line 153, in register_include_uninstalled self._parse_include(include_path, uninstalled=True) File "/usr/lib64/gobject-introspection/giscanner/transformer.py", line 224, in _parse_include parser.parse(filename) File "/usr/lib64/gobject-introspection/giscanner/girparser.py", line 63, in parse tree = parse(filename) File "/usr/lib64/python3.5/xml/etree/ElementTree.py", line 1195, in parse tree.parse(source, parser) File "/usr/lib64/python3.5/xml/etree/ElementTree.py", line 596, in parse self._root = parser._parse_whole(source) xml.etree.ElementTree.ParseError: no element found: line 9888, column 6 make[2]: *** [src/addressbook/libebook-contacts/CMakeFiles/gir-typelibs-EBookContacts_1_2_gir.dir/build.make:87: src/addressbook/libebook-contacts/EBookCo$ make[2]: Leaving directory '/var/tmp/portage/gnome-extra/evolution-data-server-3.26.6/work/evolution-data-server-3.26.6_build' make[1]: *** [CMakeFiles/Makefile2:2367: src/addressbook/libebook-contacts/CMakeFiles/gir-typelibs-EBookContacts_1_2_gir.dir/all] Error 2 make[1]: *** Waiting for unfinished jobs.... eix output: [U] gnome-extra/evolution-data-server Available versions: 3.24.7(0/60)^t (~)3.26.6(0/60)^t {api-doc-extras berkdb +gnome-online-accounts google +gtk gtk-doc +introspection ipv6 kerberos ldap test vala +weather} Installed versions: 3.24.7^t(16:04:11 6.4.2018)(berkdb gnome-online-accounts google gtk introspection ipv6 kerberos ldap weather -api-doc-extras -test -vala) Homepage: https://wiki.gnome.org/Apps/Evolution Description: Evolution groupware backend Could this be something with Python 3.5 itself? Attached are full log and emerge --info output. This package is related to commit: https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=23e3bef716c6c4a7020f3755874adb7faa6964c4
Created attachment 547758 [details] emerge --info Added output of #emerge --info
I don't think this is the cause, but can you please try with a better CFLAGS? I understand that you are just writing out -march=native, but you shouldn't be including all the -mno-* flags, as that overrides any build system added -m* flags (where it deals actual CPU support of them at runtime properly). Maybe just CFLAGS="-march=native -O2 -pipe" just to try, and if it still fails, we can investigate further. Another problem could be a temporary python eclasses problem, but that was fixed quite some days ago, so if you have synced, shouldn't be that either.
And if it still fails with simpler CFLAGS and synced tree, please try to find the file it complains line 9888 column 6 about and look into that or attach here
That is due to distcc installation. In the Gentoo Distcc guide there is referenced command to use instead of march=native: gcc -v -E -x c -march=native -mtune=native - < /dev/null 2>&1 | grep cc1 | perl -pe 's/^.* - //g;' Which on my dual Xeon system generates CFLAGS as you see them...
I know, as I said. That specific guide however is wrong - -mno* should generally be filtered out, or it breaks e.g. gimp. Please try -march=native without distcc just to see if this is the problem or not (or maybe just using distcc could be the trigger? but based on emerge --info you don't seem to use it right now)
Exactly I just have CFLAGS configured, but at moment disabled. I will try to filter all mno options out and give first shot.
(In reply to Ladislav Jech from comment #6) > I will try to filter all mno options out and give first shot. I asked you to try with just "-march=native -O2 -pipe" right now, not with -mno options filtered out. That's what I suggest going forward globally.
I understand :-) I first tried to filter out all -mno options which resulted to: CFLAGS="-march=sandybridge -mmmx -msse -msse2 -msse3 -mssse3 -mcx16 \ -msahf -maes -mpclmul -mpopcnt \ -mavx -msse4.2 \ -msse4.1 \ -mfxsr -mxsave -mxsaveopt \ --param l1-cache-size=32 --param l1-cache-line-size=64 --param l2-cache-size=20480 -mtune=sandybridge -O2 -pipe" And now the compilation worked just fine, I will make a comment to Distcc wiki to author about this behavior. Thank you guys for quick help on this!
https://wiki.gentoo.org/wiki/Distcc#-march.3Dnative already links to https://blogs.gentoo.org/mgorny/2014/06/23/inlining-marchnative-for-distcc/ which tells more specifics and that -mno is bad, but for some reason the script indeed still ignores all that. That in turn links to a stackoverflow page, which has a couple different versions how to ignore -mno; one is: gcc -march=native -v -E - < /dev/null 2>&1 | grep cc1 | perl -pe 's/ -mno-\S+//g; s/^.* - //g;'
Superior! Thank you, I need to study this in more details...
