sys-apps/dbus-1.2.12 fails with the following error: i686-pc-linux-gnu-gcc -O2 -march=native -pipe -fomit-frame-pointer -fno-ident -rdynamic -Wall -Wchar-subscripts -Wmissing-declarations -Wmissing-prototypes -Wnested-externs -Wpo inter-arith -Wcast-align -Wdeclaration-after-statement -fno-common -Wno-unused -Wno-sign-compare -Wno-pointer-sign -Wno-format -fno-strict-aliasing -Wl,-O1 -Wl,--as-needed -o db us-daemon-launch-helper activation-helper-bin.o config-loader-expat.o config-parser-common.o config-parser-trivial.o desktop-file.o utils.o activation-helper.o /usr/lib/libexpa t.so ../dbus/.libs/libdbus-convenience.a /usr/lib/gcc/i686-pc-linux-gnu/4.3.3/../../../../i686-pc-linux-gnu/bin/ld: -pie: unknown option /usr/lib/gcc/i686-pc-linux-gnu/4.3.3/../../../../i686-pc-linux-gnu/bin/ld: use the --help option for usage information collect2: ld returned 1 exit status make[2]: *** [dbus-daemon] Error 1 make[2]: *** Waiting for unfinished jobs.... make[2]: Leaving directory `/var/tmp/portage/sys-apps/dbus-1.2.12/work/dbus-1.2.12/bus' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/var/tmp/portage/sys-apps/dbus-1.2.12/work/dbus-1.2.12' make: *** [all] Error 2 Reproducible: Always Steps to Reproduce: 1. # USE=gold" emerge =binutils-2.19.51.0.4 2. # emerge =sys-apps/dbus-1.2.12 Actual Results: dbus fails to emerge. Expected Results: A. dbus should build successfully -or- B. binutils should come with a big warning that "gold" USE flag might cause problems.
are we seriously working towards supporting gold linker ?
(In reply to comment #1) > are we seriously working towards supporting gold linker ? Currently only the "patches accepted" kind of "working towards".
Ryan added <flag name='gold'>Use the new gold linker (written in C++) rather than the standard linker (this _will_ break things)</flag> to the metadata.xml for binutils sometime today. Seriously though, did you really expect everything to "just work(tm)" with the new experimental linker? You have to actively enable it, and one would hope you don't just arbitrarily turn on use flags because they sound shiny (oooh gold, must be good.) I'm not trying to be snarky - if you are enabling experimental features, you should know what you are doing. I am not suggesting you don't, I am just making a general observation. Ideally, this bug would have looked like this: When using the gold linker from binutils, sys-apps/dbus-1.2.12 fails with the following error: i686-pc-linux-gnu-gcc -O2 -march=native -pipe -fomit-frame-pointer -fno-ident -rdynamic -Wall -Wchar-subscripts -Wmissing-declarations -Wmissing-prototypes -Wnested-externs -Wpo inter-arith -Wcast-align -Wdeclaration-after-statement -fno-common -Wno-unused -Wno-sign-compare -Wno-pointer-sign -Wno-format -fno-strict-aliasing -Wl,-O1 -Wl,--as-needed -o db us-daemon-launch-helper activation-helper-bin.o config-loader-expat.o config-parser-common.o config-parser-trivial.o desktop-file.o utils.o activation-helper.o /usr/lib/libexpa t.so ../dbus/.libs/libdbus-convenience.a /usr/lib/gcc/i686-pc-linux-gnu/4.3.3/../../../../i686-pc-linux-gnu/bin/ld: -pie: unknown option /usr/lib/gcc/i686-pc-linux-gnu/4.3.3/../../../../i686-pc-linux-gnu/bin/ld: use the --help option for usage information collect2: ld returned 1 exit status Attached is the patch and a patch to the ebuild. Patch has been submitted upstream at http://fdobugs/bugid
no, it's not packages that need fixing, it's the linker. there is no "patch" that can be provided.
...to dbus that is. you know what i mean. :P
(In reply to comment #3) > Ryan added <flag name='gold'>Use the new gold linker (written in C++) rather > than the standard linker (this _will_ break things)</flag> to the metadata.xml > for binutils sometime today. Thanks! > Seriously though, did you really expect > everything to "just work(tm)" with the new experimental linker? You have to > actively enable it, and one would hope you don't just arbitrarily turn on use > flags because they sound shiny (oooh gold, must be good.) Actually, Steev, I did do some research. The only thing I was surprised about was that considering that it was released little more than a month ago, there was nothing in the USE-flag description saying that it was experimental. Am I upset that dbus does not build? No. Did I expect things to break? I did. I've considered the outcomes: Best case - I save a few minutes from my 300-package update; worst case - everything breaks and I reinstall as amd64. I built around 50 packages with the new linker until I hit something that I did not want to skip, and then I switched back to ld. > I'm not trying to be snarky - if you are enabling experimental features, you > should know what you are doing. I am not suggesting you don't, I am just > making a general observation. > > Ideally, this bug would have looked like this: > > When using the gold linker from binutils, sys-apps/dbus-1.2.12 fails with the > following error: > > ... > > Attached is the patch and a patch to the ebuild. > > Patch has been submitted upstream at http://fdobugs/bugid Ideally, yes, that would be great. However, I am not am neither a Gentoo nor a Binutils developer, I do not intend to become one anytime soon. As a user, I encountered problems, and reported them. (I reported 4 related to the gold linker.) If you feel you can find them all by yourself, then feel free to ignore my reports.
upstream has added support for -pie. it'll trickle down into newer versions of binutils over time.