Re: Fedora mass rebuild 2017

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Marek Polacek píše v Út 07. 02. 2017 v 22:32 +0100:
> It's been a tradition now that every January we rebuild all the
> Fedora packages
> with the upcoming GCC, to reveal as many bugs as possible before we
> release
> the new version.  This year is no different.
> 
> There were 18811 packages overall (last year we had 17741 packages).
> 17263 built fine with the new GCC (mostly gcc-7.0.0-0.1.fc26.src.rpm
> but I also
> used a newer version from rawhide).  1350 failed with both GCC 6 and
> GCC 7,
> so I ignored these.  This left us with ~198 packages that had to be
> analyzed,
> a number which, fortunately, was smaller than last year, when we'd
> had 577 FTBFS to
> investigate.

MongoDB is not failing in your rebuild. However latest minor upgrade
(v3.4.2) with gcc 7 is failing.
I've figured out, that is optimization issue (with -O0 it is fine, with
-01 it is failing) and it is ONLY ON ppc64le (other architectures are
fine).

I know which file is cause this issue, but I can't figure out with
which optimization. I get what changes between O0 and O1 in output of
'gcc -O0/1 -Q --help=optimizers' and compiled with -O0 and flags of
individual optimizations (which are added by O1), but failure does not
appear.

What else differs in -O1? How should I debug it?

> 
> The Fedora packages I rebuilt were grabbed on Jan 12.
> 
> It's all fairly good, and nothing particularly stands out.  There is
> the usual
> batch of new warnings (-Wimplicit-fallthrough and -Wformat-truncation 
> causing
> the biggest churn), some C++ FE changes (especially the "Fix type-
> dependence
> and the current instantiation" changes made the compiler to reject
> invalid code
> that had previously been accepted, plus invalid conversions with '\0'
> are now
> rejected), some libstdc++ changes (header dependency changes), as
> well as some
> mangling changes.
> 
> We found several GCC bugs, most of which have already been
> fixed.  Furthermore,
> Fortran ABI has changed in GCC 7.  Anyway, there's nothing that
> particularly
> worries me (except perhaps the -fprintf-return-value stuff that might
> still
> cause some wrong codes).
> 
> Shortlog appended for the people who want to get an overview of the
> details.
> 
> As usual, there will be a "porting to" document to ease the
> transition to the
> new GCC.  We already have https://gcc.gnu.org/gcc-7/porting_to.html,
> even though
> this document is still somewhat in flux.
> 
> I'd like to thank Jakub Jelinek and Jonathan Wakely, the guys I roped
> in to help
> me with all this undertaking.
> 
> GCC bugs
> ~~~~~~~~
> The following is a list of bugs we've found so far in the compiler
> and the C++
> library during the mass rebuild:
> 
> schroot-1.6.5-16.fc24.src.rpm
> 	ICE in cxx_incomplete_type_diagnostic
> 	https://gcc.gnu.org/PR78690
> 	not fixed yet
> 
> python-plyvel-0.9-7.fc26.src.rpm
> 	error: invalid rhs for gimple memory store
> 	https://gcc.gnu.org/PR79232
> 	fixed in gcc-7.0.1-0.4.fc26
> 
> mathicgb-1.0-6.20160202.gitbb268df.fc26.src.rpm
> 	ICE in tsubst_copy
> 	https://gcc.gnu.org/PR79253
> 	fixed upstream and in gcc-7.0.1-0.5.fc26
> 
> cvc4-1.4-11.fc25.src.rpm
> 	TLS model wrong for static data members
> 	https://gcc.gnu.org/PR79288
> 	fixed in gcc-7.0.1-0.4.fc26
> 
> golang-1.7.4-1.fc26.src.rpm
> 	DWARF info for typeof of C function with no args and no
> prototype
> 	is empty pointer
> 	https://gcc.gnu.org/PR79289
> 	fixed upstream and in gcc-7.0.1-0.4.fc26
> 
> texlive-2016-30.20160520.fc26.src.rpm
> tlog-2-1.fc25.src.rpm
> 	wrong code at -O2 and -fprintf-return-value
> 	https://gcc.gnu.org/PR79327
> 	fixed upstream and in gcc-7.0.1-0.6.fc26
> 
> perl-Prima-1.50-1.fc26.src.rpm
> 	-fprintf-return-value doesn't handle flexible-like array
> members properly
> 	https://gcc.gnu.org/PR79352
> 	fixed upstream and in gcc-7.0.1-0.6.fc26
> 
> clucene09-0.9.21b-16.fc24.src.rpm
> qt5-qttools-5.7.1-4.fc26.src.rpm
> 	g++ rejects valid code with error: looser throw specifier
> 	https://gcc.gnu.org/PR79393 class="Apple-tab-span"
> style="white-space:pre">	
> 
> Failures due to new warnings and -Werror
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> efivar-30-4.fc26.src.rpm
> 	a new -Walloc-size-larger-than warning
> 
> bind-dyndb-ldap-11.0-1.fc26.src.rpm
> efibootmgr-14-3.fc26.src.rpm
> fwupdate-8-2.fc26.src.rpm
> 	a new -Wduplicate-decl-specifier warning
> 
> mstflint-4.4.0-1.12.gd1edd58.1.fc26.src.rpm
> 	a new -Wformat-length warning
> 
> blobwars-1.19-13.fc24.src.rpm
> czmq-4.0.2-1.fc26.src.rpm
> elfutils-0.168-1.fc26.src.rpm
> fcoe-utils-1.0.30-5.git91c0c8c.fc24.src.rpm
> glib2-2.51.0-2.fc26.src.rpm
> glibc-2.24.90-26.fc26.src.rpm
> ipv6calc-0.99.2-17.fc26.src.rpm
> isomd5sum-1.1.0-4.fc26.src.rpm
> ixpdimm_sw-01.00.00.2144-1.fc26.src.rpm
> libhid-0.2.17-21.fc25.src.rpm
> libpsm2-10.2.2-2.fc25.src.rpm
> libyui-3.2.8-1.fc26.src.rpm
> logrotate-3.11.0-2.fc26.src.rpm
> memcached-1.4.33-1.fc26.src.rpm
> nfs-ganesha-2.4.1-2.fc26.src.rpm
> nodm-0.12-1.fc25.src.rpm
> openvas-libraries-8.0.8-2.fc26.src.rpm
> pacemaker-1.1.16-1.fc26.src.rpm
> sbd-1.2.1-3.fc24.1.src.rpm
> Simple-Fuzzer-0.7.1-1.src.rpm
> usbip-4.5-1.fc25.src.rpm
> varnish-5.0.0-1.fc26.src.rpm
> vsftpd-3.0.3-4.fc26.src.rpm
> xrootd-4.5.0-1.fc26.src.rpm
> 	a new -Wformat-truncation
> 
> ArpON-3.0-2.fc25.src.rpm
> clevis-2-1.fc26.src.rpm
> coreboot-utils-4.5-2.fc26.src.rpm
> crash-7.1.7-1.fc26.src.rpm
> gmqcc-0.3.5-10.fc26.src.rpm
> hidrd-0.2.0-3.20160712git1abf7033.fc25.src.rpm
> libs3-2.0-0.5.20161104gita052a00.fc26.src.rpm
> libvma-8.0.1-1.fc25.src.rpm
> lldpad-1.0.1-4.git036e314.fc24.src.rpm
> luksmeta-3-1.fc26.src.rpm
> mdadm-3.4-3.fc26.src.rpm
> mingw-w64-tools-3.1.999-0.8.trunk.git430863.20140530.fc24.src.rpm
> msgpack-1.4.1-1.fc25.src.rpm
> percona-xtrabackup-2.2.9-4.fc24.src.rpm
> pesign-0.112-5.fc26.src.rpm
> pidgin-sipe-1.21.1-1.fc25.src.rpm
> qpid-cpp-1.35.0-1.fc26.src.rpm
> qpid-qmf-1.35.0-1.fc26.src.rpm
> quassel-irssi-0-2.20161120gitcbd9bd7.fc26.src.rpm
> sway-0.11-4.fc26.src.rpm
> 	a new -Wimplicit-fallthrough warning
> 
> cdrkit-1.1.11-32.fc26.src.rpm
> 	a new -Wint-in-bool-contex warning
> 
> Invalid C++
> ~~~~~~~~~~~~~~~~~~~~~~
> Add64-1.2.2-10.fc24.src.rpm
> berusky2-0.10-11.fc24.src.rpm
> easystroke-0.6.0-15.fc24.src.rpm
> guitarix-0.35.0-3.fc26.src.rpm
> inkscape-0.92.0-11.fc26.src.rpm
> jack-audio-connection-kit-1.9.10-5.fc24.src.rpm
> kate4-4.14.3-16.fc24.src.rpm
> milkytracker-0.90.86-3.fc25.src.rpm
> muse-2.2.1-2.fc24.src.rpm
> rawtherapee-4.2.1234-2.20161105gitb766110.fc26.src.rpm
> svxlink-15.11-1.fc24.src.rpm
> vdr-2.2.0-10.fc25.src.rpm
> 	error: call of overloaded 'abs(...)' is ambiguous
> 	All overloads of the abs function are now declared by including
> either
> 	of <cstdlib> or <cmath>, and the custom overloads conflict with
> one of
> 	the additional overloads in the standard
> headers.  Additionally,
> 	calling abs with an argument of unsigned type is now ill-formed 
> after
> 	inclusion of any standard abs overload. 
> 
> capnproto-0.5.3-2.fc25.src.rpm
> 	error: could not convert template argument 'b' from 'bool' to
> 'capnp::Kind'
> 	Presumably invalid C++.
> 
> atlascpp-0.6.3-6.fc24.src.rpm
> glmark2-2014.03-6.fc25.src.rpm
> ladish-2-13.3.gitfcb16ae.fc24.src.rpm
> libinvm-cim-1.0.0.1043-1.fc26.src.rpm
> libvoikko-3.8-2.fc25.src.rpm
> openoffice.org-diafilter-1.7.5-1.fc24.src.rpm
> pulseview-0.3.0-1.fc24.src.rpm
> 	error: dynamic exception specifications are deprecated in
> C++11; use 'noexcept' instead
> 	In C++17, throw(int) is no longer valid.
> 
> asgp-1.0.18-6.fc26.src.rpm
> bear-0.7.0-0.11.20161230git781ec80.fc26.src.rpm
> crawl-0.19.1-1.fc26.src.rpm
> libclaw-1.7.4-16.fc26.src.rpm
> plee-the-bear-0.7.0-10.fc26.src.rpm
> steghide-0.5.1-28.fc26.src.rpm
> 	error: expected primary-expression before ... token
> 	Invalid C++ code; it needs to be "position.template
> cast_value_type_to".
> 	See https://womble.decadent.org.uk/c++/template-faq.html#disamb
> iguation.
> 
> normaliz-2.12.2-8.fc24.src.rpm
> 	error: expression cannot be used as a function
> 	Invalid C++, Jon opened:
> 	https://bugzilla.redhat.com/show_bug.cgi?id=1417678
> class="Apple-tab-span" style="white-space:pre">	
> 
> apt-cacher-ng-0.9.0-1.fc24.src.rpm
> clementine-1.3.1-4.fc26.src.rpm
> conky-1.10.4-1.fc26.src.rpm
> dwgrep-0.2-4.fc24.src.rpm
> fflas-ffpack-2.2.2-3.fc26.src.rpm
> glogg-1.1.2-1.fc26.src.rpm
> gpgme-1.8.0-7.fc26.src.rpm
> gtengine-3.5-1.fc26.src.rpm
> lldb-3.9.0-3.fc26.src.rpm
> pingus-0.7.6-20.fc24.src.rpm
> plasma-desktop-5.8.5-2.fc26.src.rpm
> worker-3.3.3-6.fc24.src.rpm
> 	Quoting from porting_to:
> 	Several C++ Standard Library headers have been changed to no
> longer
> 	include the <functional> header. As such, C++ programs that
> used components
> 	defined in <functional> without explicitly including that
> header will no
> 	longer compile.
> 
> 	Previously components such as std::bind and std::function were
> implicitly
> 	defined after including unrelated headers such as <memory>,
> <futex>, <mutex>,
> 	and <regex>. Correct code should #include <functional> to
> define them. 
> 
> 	Also see https://bugzilla.redhat.com/show_bug.cgi?id=1417383
> 	
> bison-3.0.4-4.fc24.src.rpm
> 	This package contains code that is trying to add some classes
> to std::,
> 	that is undefined behavior.
> 
> Coin2-2.5.0-28.fc26.src.rpm
> Coin3-3.1.3-18.fc26.src.rpm
> dbus-c++-0.9.0-11.fc24.src.rpm
> festival-1.96-32.fc24.src.rpm
> gambas3-3.9.2-2.fc26.src.rpm
> htdig-3.2.0-0.25.b6.fc24.src.rpm
> kyotocabinet-1.2.76-11.fc26.src.rpm
> mup-6.5-1.fc26.src.rpm
> percolator-3.01-1.fc26.src.rpm
> pki-core-10.3.5-10.fc26.src.rpm
> scim-1.4.17-1.fc25.src.rpm
> scim-bridge-0.4.16-22.fc24.src.rpm
> speed-dreams-2.2.1-2.fc26.src.rpm
> torcs-1.3.7-1.fc25.src.rpm
> 	error: invalid conversion from 'char' to 'char*'
> 	s = '\0'
> 	This code is invalid in C++11 and GCC doesn't compile it
> anymore (pointer
> 	conversions are only allowed for an integral-literal and not a
> character
> 	literal).  In C++03 and in C the compiler warns (-Wpointer-
> compare).
> 
> vegastrike-0.5.1-27.r1.fc24.src.rpm
> 	error: request for member ... in ...
> 	Looks like '.' is used instead of '->', so invalid C++.
> 
> apvlv-0.1.4-11.fc25.src.rpm
> arm-none-eabi-gcc-cs-6.2.0-1.fc26.src.rpm
> aspell-0.60.6.1-14.fc26.src.rpm
> avr-gcc-6.2.0-1.fc26.src.rpm
> dyninst-9.3.0-1.fc26.src.rpm
> edk2-20161105git3b25ca8-1.fc26.src.rpm
> frysk-0.4-54.fc26.src.rpm
> gearmand-1.1.14-1.fc26.src.rpm
> glog-0.3.3-12.fc25.src.rpm
> gpsim-0.29.0-1.fc24.src.rpm
> hylafax+-5.5.8-1.fc24.src.rpm
> kdelibs-4.14.28-1.fc26.src.rpm
> libmemcached-1.0.18-7.fc24.src.rpm
> libmp4v2-2.1.0-0.6.trunkREV507.fc24.src.rpm
> mingw-gcc-6.3.0-1.fc26.src.rpm
> mozjs17-17.0.0-16.fc26.src.rpm
> mozjs24-24.2.0-11.fc26.src.rpm
> ode-0.14-2.fc24.src.rpm
> oorexx-4.2.0-3.fc26.src.rpm
> primer3-2.3.6-6.fc24.src.rpm
> shigofumi-0.6-4.fc25.src.rpm
> snoopy-2.2.6-3.fc24.src.rpm
> vdr-live-0.3.0-24.20150213git6ea279a.fc26.src.rpm
> xmoto-0.5.11-9.fc24.src.rpm
> 	error: ISO C++ forbids comparison between pointer and integer
> 	You can no longer write "ptr == '\0'" as it's invalid C++11.
> 
> atanks-6.5-1.fc26.src.rpm
> blitz-0.10-10.fc24.src.rpm
> nodejs-6.9.4-1.fc26.src.rpm
> OpenEXR-2.2.0-6.fc26.src.rpm
> qt5-qtbase-5.7.1-12.fc26.src.rpm
> qt5-qtwebengine-5.7.1-4.fc26.src.rpm
> v8-314-3.14.5.10-3.fc25.src.rpm
> z3-4.5.0-1.fc26.src.rpm
> 	error: invalid use of incomplete type
> 	This code is ill-formed and G++ now rejects such code.
> 
> k3d-0.8.0.6-2.fc26.src.rpm
> 	error: no match for 'operator<<' (operand types are
> 'std::ostream
> 	{aka std::basic_ostream<char>}' and 'const double')
> 	Missing <ostream>.
> 
> codeblocks-16.01-2.fc26.src.rpm
> cyphesis-0.6.2-9.fc24.src.rpm
> 	there are no arguments to .. that depend on a template
> parameter,
> 	so a declaration of ... must be available
> 	This code is ill-formed and G++ now rejects such code.
> 
> ETL-0.04.22-1.fc26.src.rpm
> libffado-2.3.0-1.fc26.src.rpm
> meshlab-1.3.2-10.fc25.src.rpm
> sflphone-1.4.1-20.fc26.src.rpm
> 	error: no matching function for call to ...
> 	Invalid code.
> 
> mongo-cxx-driver-1.1.2-3.fc26.src.rpm
> 	Unknown failure.  Fails in a shim test, so perhaps some
> 	libstdc++ changes?
> 
> gmm-5.0-1.fc24.src.rpm
> 	Package exceeded -ftemplate-depth limit.
> 
> cppcheck-1.75-1.fc26.src.rpm
> pdns-4.0.1-3.fc26.src.rpm
> pdns-recursor-4.0.3-2.fc26.src.rpm
> 	error: types may not be defined in a for-range-declaration
> 	Invalid code: "The decl-specifier-seq shall not define a
> 	class or enumeration."
> 	(This used to be a warning.)
> 
> openigtlink-0-0.2.20151015gitccb2438.fc24.src.rpm
> 	error: 'void*' is not a pointer-to-object type
> 	Invalid code.
> 
> Other
> ~~~~~
> ruby-2.3.3-61.fc26.src.rpm
> 	This was a GC bug in Ruby; fixed.
> 	https://bugzilla.redhat.com/show_bug.cgi?id=1417590
> 
> florist-2011-19.fc24.src.rpm
> matreshka-0.7.0-5.fc24.src.rpm
> 	Error: no native compiler found for language 'ada', default
> runtime
> 	This is gcc-gnat related.  Didn't investigate further.
> 
> cryptlib-3.4.3-8.fc26.src.rpm
> cryptobone-1.1.1-2.fc26.src.rpm
> 	error: CPU you selected does not support x86-64 instruction set
> 	-mcpu=pentium in the options list.
> 
> diffutils-3.5-1.fc26.src.rpm
> 	A bug in the package.  I opened a BZ with a patch:
> 	https://bugzilla.redhat.com/show_bug.cgi?id=1419536
> 
> libbson-1.5.1-1.fc26.src.rpm
> 	-fwrapv helped, so I bet this is a bug in the package.  I
> suggest
> 	to use -fsanitize=undefined to see where the overflow happens.
> 
> libosmocore-0.9.0-4.2015110~t916423ef.fc24.src.rpm
> 	-fgnu89-inline helped.  See what I wrote in
> 	https://gcc.gnu.org/gcc-5/porting_to.html
> 
> ecl-16.1.2-1.fc26.src.rpm
> 	This failed because there was an insanely huge struct and the
> new
> 	diagnostics (fix-it hints) was trying to print it all.  I
> killed
> 	the build.
> 
> coin-or-Dip-0.91.2-6.fc25.src.rpm
> R-DynDoc-1.52.0-1.fc26.src.rpm
> R-pls-2.4.3-5.fc24.src.rpm
> 	Missing libgfortran 7.
> 
> bastet-0.43.1-11.fc24.src.rpm
> ocaml-cil-1.7.3-27.fc26.src.rpm
> ocp-0.1.22-0.10.git849cc42.fc26.src.rpm
> openms-2.1.0-3.fc26.src.rpm
> openttd-1.6.1-2.fc26.src.rpm
> 	These are not ready for a new gcc version (gcc -dumpversion
> says '7'
> 	and not e.g. '6.3.1') and the packages fail to cope with that.
> 
> perl-HTTP-Proxy-0.304-3.fc25.src.rpm
> perl-OpenGL-0.6704-5.fc25.src.rpm
> 	Two Perl failures that I wasn't able to analyze, but since they
> 	don't use gcc at all, I presume it's not gcc's fault.
> 
> vavoom-1.33-19.fc26.src.rpm
> 	For some unknown reason a piece of a Makefile isn't performed,
> 	resulting in a failure.
> 
> arc-gui-clients-0.4.6-9.fc25.src.rpm
> archimedes-2.0.1-5.fc24.src.rpm
> audacity-2.1.3-0.7.20161109git53a5c93.fc26.src.rpm
> bwm-ng-0.6.1-2.fc24.src.rpm
> ember-0.7.2-15.fc25.src.rpm
> extremetuxracer-0.7.4-1.fc26.src.rpm
> funguloids-1.06-18.fc24.src.rpm
> glite-lbjp-common-gsoap-plugin-3.2.12-10.fc26.src.rpm
> gnokii-0.6.31-14.fc24.src.rpm
> gobby-0.4.13-10.fc24.src.rpm
> linbox-1.4.2-5.fc26.src.rpm
> marsshooter-0.7.6-2.fc26.src.rpm
> mgetty-1.1.37-1.fc26.src.rpm
> mkvtoolnix-9.6.0-1.fc26.src.rpm
> openalchemist-0.4-19.fc24.src.rpm
> powerline-2.5-4.fc26.src.rpm
> sobby-0.4.7-12.fc24.src.rpm
> tunneler-1.1.1-17.fc24.src.rpm
> 	Undefined references.  Some of them might be due to C99/GNU89
> inline
> 	semantics, but some of them are result of the mangling changes,
> see
> 	porting_to.html.
> 
> pspp-0.10.2-1.fc26.src.rpm
> 	A bug in the package, see
> 	https://bugzilla.redhat.com/show_bug.cgi?id=1419174
> 
> 	Marek
> _______________________________________________
> devel mailing list -- devel@xxxxxxxxxxxxxxxxxxxxxxx
> To unsubscribe send an email to devel-leave@xxxxxxxxxxxxxxxxxxxxxxx
_______________________________________________
devel mailing list -- devel@xxxxxxxxxxxxxxxxxxxxxxx
To unsubscribe send an email to devel-leave@xxxxxxxxxxxxxxxxxxxxxxx




[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Index of Archives]     [Fedora Announce]     [Fedora Kernel]     [Fedora Testing]     [Fedora Formulas]     [Fedora PHP Devel]     [Kernel Development]     [Fedora Legacy]     [Fedora Maintainers]     [Fedora Desktop]     [PAM]     [Red Hat Development]     [Gimp]     [Yosemite News]
  Powered by Linux