Created attachment 181419 [details] Fixes for llvm39 and llvm40 (ports tree patch)
Created attachment 181420 [details] ino64: enlarge ino_t to 64bit and several other ABI forward changes
I ask for the exp-run of the patch which extends ino_t to 64 bit, and also does related VFS ABI breaking changes. It is known that llvm* ports are broken due to asan doing dirty things, I attached the ports tree patch which fixes llvm39 and llvm40. During other attempts to test ports with the patch, it was once reported that devel/libclc is broken, but I was unable to reproduce the issue. Please see the file testing.txt which appear in the root of the patched src/ tree for details and motivation of the patch. IMPORTANT: after patching src, you must do (cd sys/kern && touch syscalls.master && make sysent) (cd sys/compat/freebsd && touch syscalls.master && make sysent) before rebuilding world and kernel.
Exp-run results on amd64: http://pb2.nyi.freebsd.org/build.html?mastername=headamd64PR218320-default&build=2017-04-05_20h35m39s New failures on amd64: + {"origin"=>"devel/libgtop", "phase"=>"build", "errortype"=>"clang"} + {"origin"=>"devel/llvm-devel", "phase"=>"build", "errortype"=>"clang"} + {"origin"=>"devel/llvm38", "phase"=>"build", "errortype"=>"clang"} + {"origin"=>"devel/radare2", "phase"=>"build", "errortype"=>"clang"} + {"origin"=>"emulators/qemu-user-static", "phase"=>"build", "errortype"=>"bad_C++_code"} + {"origin"=>"lang/ccl", "phase"=>"build", "errortype"=>"coredump"} + {"origin"=>"lang/dmd2", "phase"=>"build", "errortype"=>"coredump"} + {"origin"=>"lang/gcc5-aux", "phase"=>"build", "errortype"=>"???"} + {"origin"=>"lang/gcc6-aux", "phase"=>"build", "errortype"=>"???"} + {"origin"=>"lang/ghc", "phase"=>"configure", "errortype"=>"???"} + {"origin"=>"lang/ldc", "phase"=>"build", "errortype"=>"bad_C++_code"} + {"origin"=>"lang/modula3", "phase"=>"build", "errortype"=>"???"} + {"origin"=>"lang/rust", "phase"=>"build", "errortype"=>"???"} + {"origin"=>"lang/rust-nightly", "phase"=>"build", "errortype"=>"???"} + {"origin"=>"lang/twelf", "phase"=>"build", "errortype"=>"???"} + {"origin"=>"lang/urweb", "phase"=>"build", "errortype"=>"???"} + {"origin"=>"multimedia/webcamd", "phase"=>"build", "errortype"=>"???"} + {"origin"=>"net-mgmt/netdata", "phase"=>"build", "errortype"=>"clang"} + {"origin"=>"sysutils/py-psutil", "phase"=>"build", "errortype"=>"clang"} + {"origin"=>"sysutils/py-psutil121", "phase"=>"build", "errortype"=>"clang"} Around 650 new ports were skipped due to those failures. Failure logs: http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-04-05_20h35m39s/logs/errors/libgtop-2.32.0.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-04-05_20h35m39s/logs/errors/llvm-devel-5.0.d20170329_1.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-04-05_20h35m39s/logs/errors/llvm38-3.8.1_7.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-04-05_20h35m39s/logs/errors/radare2-1.0.2.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-04-05_20h35m39s/logs/errors/qemu-user-static-2.8.50.g20170307_2.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-04-05_20h35m39s/logs/errors/ccl-1.11.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-04-05_20h35m39s/logs/errors/dmd2-2.073.2.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-04-05_20h35m39s/logs/errors/gcc5-aux-20160603.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-04-05_20h35m39s/logs/errors/gcc6-aux-20170202_1.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-04-05_20h35m39s/logs/errors/ghc-7.10.2_2.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-04-05_20h35m39s/logs/errors/ldc-1.1.0.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-04-05_20h35m39s/logs/errors/modula3-5.10.0_4.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-04-05_20h35m39s/logs/errors/rust-1.16.0.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-04-05_20h35m39s/logs/errors/rust-nightly-1.17.0.20170212.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-04-05_20h35m39s/logs/errors/twelf-1.7.1.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-04-05_20h35m39s/logs/errors/urweb-20150520.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-04-05_20h35m39s/logs/errors/webcamd-4.8.0.4.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-04-05_20h35m39s/logs/errors/netdata-1.6.0.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-04-05_20h35m39s/logs/errors/py27-psutil-5.2.1.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-04-05_20h35m39s/logs/errors/py27-psutil121-1.2.1_2.log
Some of the errors appear to be Clang 4.0.0 issues or otherwise unrelated to ino64 -- e.g. from databases/mongodb: | In file included from src/mongo/db/db.cpp:47: | src/mongo/db/client.h:263:59: error: ordered comparison between pointer and zero ('mongo::Client *' and 'int') | inline bool haveClient() { return currentClient.get() > 0; } | ~~~~~~~~~~~~~~~~~~~ ^ ~
Exp-run results on i386: http://pb2.nyi.freebsd.org/build.html?mastername=headi386PR218320-default&build=2017-04-08_06h24m47s Baseline build is still ongoing (at http://beefy11.nyi.freebsd.org/build.html?mastername=head-i386-default&build=p437978_s316629 )so I will post new failures when i'm back in a week.
New failures on i386: + {"origin"=>"devel/libgtop", "phase"=>"build", "errortype"=>"clang"} + {"origin"=>"devel/llvm-devel", "phase"=>"build", "errortype"=>"clang"} + {"origin"=>"devel/llvm38", "phase"=>"build", "errortype"=>"clang"} + {"origin"=>"devel/radare2", "phase"=>"build", "errortype"=>"clang"} + {"origin"=>"emulators/open-vm-tools", "phase"=>"build", "errortype"=>"clang_werror"} + {"origin"=>"emulators/open-vm-tools-nox11", "phase"=>"build", "errortype"=>"clang_werror"} + {"origin"=>"games/kiki", "phase"=>"build", "errortype"=>"bad_C++_code"} + {"origin"=>"lang/ccl", "phase"=>"build", "errortype"=>"coredump"} + {"origin"=>"lang/dmd2", "phase"=>"build", "errortype"=>"coredump"} + {"origin"=>"lang/gcc5-aux", "phase"=>"build", "errortype"=>"???"} + {"origin"=>"lang/gcc6-aux", "phase"=>"build", "errortype"=>"???"} + {"origin"=>"lang/ghc", "phase"=>"configure", "errortype"=>"???"} + {"origin"=>"lang/ldc", "phase"=>"build", "errortype"=>"bad_C++_code"} + {"origin"=>"lang/rust", "phase"=>"build", "errortype"=>"???"} + {"origin"=>"lang/rust-nightly", "phase"=>"build", "errortype"=>"???"} + {"origin"=>"lang/twelf", "phase"=>"build", "errortype"=>"???"} + {"origin"=>"lang/urweb", "phase"=>"build", "errortype"=>"???"} + {"origin"=>"multimedia/webcamd", "phase"=>"build", "errortype"=>"???"} + {"origin"=>"net-mgmt/netdata", "phase"=>"build", "errortype"=>"clang"} + {"origin"=>"sysutils/py-psutil", "phase"=>"build", "errortype"=>"clang"} + {"origin"=>"sysutils/py-psutil121", "phase"=>"build", "errortype"=>"clang"} Failure logs on i386: http://pb2.nyi.freebsd.org/data/headi386PR218320-default/2017-04-08_06h24m47s/logs/errors/libgtop-2.32.0.log http://pb2.nyi.freebsd.org/data/headi386PR218320-default/2017-04-08_06h24m47s/logs/errors/llvm-devel-5.0.d20170329_1.log http://pb2.nyi.freebsd.org/data/headi386PR218320-default/2017-04-08_06h24m47s/logs/errors/llvm38-3.8.1_7.log http://pb2.nyi.freebsd.org/data/headi386PR218320-default/2017-04-08_06h24m47s/logs/errors/radare2-1.0.2.log http://pb2.nyi.freebsd.org/data/headi386PR218320-default/2017-04-08_06h24m47s/logs/errors/open-vm-tools-10.1.0_1,2.log http://pb2.nyi.freebsd.org/data/headi386PR218320-default/2017-04-08_06h24m47s/logs/errors/open-vm-tools-nox11-10.1.0_1,2.log http://pb2.nyi.freebsd.org/data/headi386PR218320-default/2017-04-08_06h24m47s/logs/errors/kiki-1.0.2_18.log http://pb2.nyi.freebsd.org/data/headi386PR218320-default/2017-04-08_06h24m47s/logs/errors/ccl-1.11.log http://pb2.nyi.freebsd.org/data/headi386PR218320-default/2017-04-08_06h24m47s/logs/errors/dmd2-2.073.2.log http://pb2.nyi.freebsd.org/data/headi386PR218320-default/2017-04-08_06h24m47s/logs/errors/gcc5-aux-20160603.log http://pb2.nyi.freebsd.org/data/headi386PR218320-default/2017-04-08_06h24m47s/logs/errors/gcc6-aux-20170202_1.log http://pb2.nyi.freebsd.org/data/headi386PR218320-default/2017-04-08_06h24m47s/logs/errors/ghc-7.10.2_2.log http://pb2.nyi.freebsd.org/data/headi386PR218320-default/2017-04-08_06h24m47s/logs/errors/ldc-1.1.0.log http://pb2.nyi.freebsd.org/data/headi386PR218320-default/2017-04-08_06h24m47s/logs/errors/rust-1.16.0.log http://pb2.nyi.freebsd.org/data/headi386PR218320-default/2017-04-08_06h24m47s/logs/errors/rust-nightly-1.17.0.20170212.log http://pb2.nyi.freebsd.org/data/headi386PR218320-default/2017-04-08_06h24m47s/logs/errors/twelf-1.7.1.log http://pb2.nyi.freebsd.org/data/headi386PR218320-default/2017-04-08_06h24m47s/logs/errors/urweb-20150520.log http://pb2.nyi.freebsd.org/data/headi386PR218320-default/2017-04-08_06h24m47s/logs/errors/webcamd-4.8.0.4.log http://pb2.nyi.freebsd.org/data/headi386PR218320-default/2017-04-08_06h24m47s/logs/errors/netdata-1.6.0.log http://pb2.nyi.freebsd.org/data/headi386PR218320-default/2017-04-08_06h24m47s/logs/errors/py27-psutil-5.2.1.log http://pb2.nyi.freebsd.org/data/headi386PR218320-default/2017-04-08_06h24m47s/logs/errors/py27-psutil121-1.2.1_2.log Around 670 ports were skipped due to those new failures.
Created attachment 182332 [details] ino64: enlarge ino_t to 64bit and several other ABI forward changes, v2
Created attachment 182333 [details] Ports tree patch (llvm, ghc, rust and some minor ports)
Updated patch set. Hopefully, the big ports blocking many others are fixed. Still not sure what to do with small (leaf) ports.
lang/ghc fails: build depends are not installed yet during patch phase so ${CC} is not available http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-05-06_15h54m27s/logs/errors/ghc-7.10.2_2.log Maybe the "${CC} ${CFLAGS} .." part has to be moved to pre-configure target?
Is there a patch for multimedia/webcamd ? 61 ports (including gnome3 and pidgin) are skipped because of it right now.
Created attachment 182355 [details] Ports tree patch (llvm, ghc, rust, webcamd and some minor ports) Reorganized ghc patch to compile the hack at pre-configure stage. Added patch for webcamd.
Exp-run results on i386: http://pb2.nyi.freebsd.org/build.html?mastername=headamd64PR218320-default&build=2017-05-11_05h06m10s New failures on i386: + {"origin"=>"devel/llvm-devel", "phase"=>"build", "errortype"=>"clang"} + {"origin"=>"devel/llvm38", "phase"=>"build", "errortype"=>"clang"} + {"origin"=>"devel/radare2", "phase"=>"build", "errortype"=>"clang"} + {"origin"=>"emulators/qemu-user-static", "phase"=>"build", "errortype"=>"bad_C++_code"} + {"origin"=>"lang/ccl", "phase"=>"build", "errortype"=>"coredump"} + {"origin"=>"lang/dmd2", "phase"=>"build", "errortype"=>"coredump"} + {"origin"=>"lang/gcc5-aux", "phase"=>"build", "errortype"=>"???"} + {"origin"=>"lang/gcc6-aux", "phase"=>"build", "errortype"=>"???"} + {"origin"=>"lang/ldc", "phase"=>"build", "errortype"=>"bad_C++_code"} + {"origin"=>"lang/modula3", "phase"=>"build", "errortype"=>"???"} + {"origin"=>"lang/rust-nightly", "phase"=>"extract", "errortype"=>"???"} + {"origin"=>"lang/twelf", "phase"=>"build", "errortype"=>"???"} + {"origin"=>"lang/urweb", "phase"=>"build", "errortype"=>"???"} + {"origin"=>"sysutils/py-psutil121", "phase"=>"build", "errortype"=>"clang"} Failure logs on i386: http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-05-11_05h06m10s/logs/errors/llvm-devel-5.0.d20170329_2.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-05-11_05h06m10s/logs/errors/llvm38-3.8.1_8.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-05-11_05h06m10s/logs/errors/radare2-1.0.2.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-05-11_05h06m10s/logs/errors/qemu-user-static-2.8.50.g20170307_2.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-05-11_05h06m10s/logs/errors/ccl-1.11.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-05-11_05h06m10s/logs/errors/dmd2-2.073.2.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-05-11_05h06m10s/logs/errors/gcc5-aux-20160603.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-05-11_05h06m10s/logs/errors/gcc6-aux-20170202_1.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-05-11_05h06m10s/logs/errors/ldc-1.1.0.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-05-11_05h06m10s/logs/errors/modula3-5.10.0_4.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-05-11_05h06m10s/logs/errors/rust-nightly-1.17.0.20170212.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-05-11_05h06m10s/logs/errors/twelf-1.7.1.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-05-11_05h06m10s/logs/errors/urweb-20150520.log http://pb2.nyi.freebsd.org/data/headamd64PR218320-default/2017-05-11_05h06m10s/logs/errors/py27-psutil121-1.2.1_2.log
The results above were on amd64
(In reply to Antoine Brodin from comment #13) So this looks quite good, in the sense that all ports and dependencies which were supposed to be fixed, are fixed. Am I right ? What is the portmgr group opinion on the current state ? Do you require that I fixed all remaining ports, or do you agree that low-profile ports fixes can be delegated to the ports maintainers ? Thank you for all the efforts.
Created attachment 182641 [details] Update libgtop patch with __FreeBSD_version guards The libgtop patch from the big ports patch with __FreeBSD_version code. Only thing needed is the correct value to be filled in when the ino64 work is committed.
For reference, ino64 is __FreeBSD_version 1200031
Is there anything further that needs to be done here?
I believe gcc6-aux is still broken on i386?
The code has landed, long ago, if there still are leftovers, a new PR should be open dedicated to it