View previous topic :: View next topic |
Author |
Message |
mv Watchman
Joined: 20 Apr 2005 Posts: 6747
|
Posted: Mon Jul 20, 2015 5:48 pm Post subject: My experience: gcc-5 usable by now |
|
|
According to my experience switching to gcc-5 is what can now be done even on production systems:
I am running all of my systems (amd64 and x86) compiled completely with gcc-5.1 for quite a while now; the last problem I had was resolved with the release of gcc-5.2: If qtwebkit is compiled with gcc-5.2 (instead of gcc-5.1), skype will no longer crash.
For comparison: My systems are rather complete desktops, including libreoffice, full texlive (2014 and now even 2015), and a lot of math applications. (Be aware that I use neither gnome nor kde; however, in the German forum full kde (4 and 5) were reported to work flawlessly with gcc-5.1, too. I suppose - though I didn't try - that gnome shouldn't be a problem, either).
Things to be aware of: - Compile the whole system with gcc-5: Compiling only part of the system calls for trouble due to the changed ABI. This is the most important point. If you observe this and run carefully a full emerge -e @world (observing the points below) almost nothing can go wrong. eix-installed-after -b ... is your friend...
- Some packages need to be ~amd64 or ~x86 to compile.
- Very few packages need in addition a patch to compile; however, all patches I needed are contained in the gentoo bugzilla and can all be found over the gcc-5 tracker bug (most of these patches are meanwhile in the gentoo repository anyway).
- I used flto only where it compiled out-of-the-box (I did not attempt to patch binutils or switch to gold in case of problems). Besides compilation issues, I had only one problem with flto on gcc-5 so far: If icu is compiled with flto then libreoffice will not compile on x86 (it does on amd64, though).
- I did not try graphite, because I had bad experience with it (subtle crashs and instabilities) from previous gcc versions.
|
|
Back to top |
|
|
VoidMage Watchman
Joined: 14 Oct 2006 Posts: 6196
|
Posted: Mon Jul 20, 2015 6:46 pm Post subject: |
|
|
If we're talking about the same binutils patch, it's already included in binutils 2.25, you just need a symlink then.
Still, any chance you've tested wine ? I known it was miscompiled by 5.1...
In fact, it seems that there's still a bug that affects it, that's fixed on the 5 branch, but not yet in 5.2 (unless gcc maintainers imported the fix). |
|
Back to top |
|
|
mv Watchman
Joined: 20 Apr 2005 Posts: 6747
|
Posted: Tue Jul 21, 2015 9:46 am Post subject: |
|
|
VoidMage wrote: | Still, any chance you've tested wine |
I only compiled it (the stable version), but probably did not use it since years... |
|
Back to top |
|
|
mv Watchman
Joined: 20 Apr 2005 Posts: 6747
|
Posted: Tue Jul 21, 2015 9:55 am Post subject: |
|
|
VoidMage wrote: | If we're talking about the same binutils patch, it's already included in binutils 2.25, you just need a symlink then. |
Are you speaking about the linker plugin?
If it really is true that it is built by default (I did not check), why is the symlink not created by default?
It should'n hurt to install this symlink, since AFAIK, it is ignored unless -fuse-linker-plugin is passed to the compiler/linker. |
|
Back to top |
|
|
nlsa8z6zoz7lyih3ap Guru
Joined: 25 Sep 2007 Posts: 388 Location: Canada
|
Posted: Mon Jul 27, 2015 10:04 pm Post subject: |
|
|
Inspired by this topic I have suucessfully upgraded to gcc-5.2 with kde. The result does seem to be stable enough for production use.
Here are some points about this upgrade.
(1) In order to compile kde I upgraded to "~amd64". Not everything in kde-meta compiles, but everything in my working desktop selected with sets of my choice does and, as far as I can see, works just fine.
(2) xemacs and wine did not compile with gcc-5.2 so I recompiled them with gcc-4.9.3 and they seem to work just fine on this system.
(3) the upgrade from gcc-4.9.3 to gcc-5.2 was not seamless the way that it was from 4.7 to 4.9 . However, with a little persistance I was able to succeed.
I always enjoy trying things like this. Except for the fun of it, I don't think that the effort was worth it in any practical sense, so I don't really recommend it to anyone else yet, except to point out that if you want to try it the result works very well indeed. (I tried this with gcc-5.0 and had no success whatsoever.)
(4) I always have two working partitions of "/" so it will be painless to reutrn to gcc-4.9.3 and "amd64" now that I have had my fun.
I used no special patches or overlays. |
|
Back to top |
|
|
nlsa8z6zoz7lyih3ap Guru
Joined: 25 Sep 2007 Posts: 388 Location: Canada
|
Posted: Tue Jul 28, 2015 2:49 pm Post subject: |
|
|
My system with gcc-5.2 as described above hung when shutting down from kde and required a hard reset. Perhaps this is because traces of gcc-4.9 remained on the system.
Thus, in my case, gcc-5.2 is not yet ready for production use. |
|
Back to top |
|
|
Buffoon Veteran
Joined: 17 Jun 2015 Posts: 1369 Location: EU or US
|
Posted: Tue Jul 28, 2015 3:35 pm Post subject: |
|
|
emerge -e @system, 35 packages didn't build with gcc-5.2.0. |
|
Back to top |
|
|
mv Watchman
Joined: 20 Apr 2005 Posts: 6747
|
Posted: Tue Jul 28, 2015 5:15 pm Post subject: |
|
|
nlsa8z6zoz7lyih3ap wrote: | hung when shutting down from kde and required a hard reset |
This doesn't sound like a compiler problem at all: Even a segfault or something similar should not require a hard reset.
Quote: | Perhaps this is because traces of gcc-4.9 remained on the system. |
If this is the case, further discussion is not necessary: It is a hard requirement that really everything was recompiled with gcc-5. Otherwise, due to the changed ABI, you will experience all sort of trouble. |
|
Back to top |
|
|
mv Watchman
Joined: 20 Apr 2005 Posts: 6747
|
Posted: Tue Jul 28, 2015 5:26 pm Post subject: |
|
|
Buffoon wrote: | emerge -e @system, 35 packages didn't build with gcc-5.2.0. |
Given that number, I would be surprised if the reason is not a wrong emerge order. (OTOH, any single build failure can also cause any subsequent packages to fail, and then you might have chaos of mixed versoins).
emerge -e @system is perhaps not the right strategy. I usually upgrade gcc as follows. I have a set @toolchain which consists of Code: | sys-libs/timezone-data
sys-devel/binutils
sys-devel/binutils-config
sys-devel/gcc
sys-devel/gcc-config
sys-kernel/linux-headers
sys-libs/glibc
sys-libs/zlib
dev-libs/gmp
dev-libs/mpfr
dev-libs/mpc
dev-libs/isl |
If I have enough computer time, I do Quote: | emerge -1O @toolchain | three times (if I have less computer time, I make sure to emerge linux-headers first and gcc last of the second time, and then no third time should be necessary). After this, an should run through. If it is does not, fix immediately that package which fails (possibly by upgrading to unstable version or using a patch from bugzilla) and repeat the last command. |
|
Back to top |
|
|
Buffoon Veteran
Joined: 17 Jun 2015 Posts: 1369 Location: EU or US
|
Posted: Tue Jul 28, 2015 5:35 pm Post subject: |
|
|
Thanks, will create that toolchain set. Is isl still used by gcc-5.2.0? |
|
Back to top |
|
|
nlsa8z6zoz7lyih3ap Guru
Joined: 25 Sep 2007 Posts: 388 Location: Canada
|
Posted: Tue Jul 28, 2015 5:35 pm Post subject: |
|
|
Quote: | If this is the case, further discussion is not necessary: It is a hard requirement that really everything was recompiled with gcc-5. Otherwise, due to the changed ABI, you will experience all sort of trouble. |
I agree. I had to leave gcc-4.9 on the system in order to compile wine.
The point of my post was to share my experience that I was able to run stably with my select kde packages from "~amd64" and gcc-5.9.2.
The problem didn't arise until I reinstalled gcc-4.9 to compile wine. These problems will all go away, I suspect, when everything that I need compiles with gcc-5 . |
|
Back to top |
|
|
schorsch_76 Guru
Joined: 19 Jun 2012 Posts: 450
|
Posted: Tue Jul 28, 2015 5:36 pm Post subject: |
|
|
@mv: Thanks for that thread! In deed i was thinking to switch to gcc 5.2 too! Now a backup is running on my laptop and i prepare to switch to it _________________ // valid again: I forgot about the git access. Now 1.2GB big. Start: 2015-06-25
git daily portage tree
Web: https://github.com/schorsch1976/portage
git clone https://github.com/schorsch1976/portage |
|
Back to top |
|
|
mv Watchman
Joined: 20 Apr 2005 Posts: 6747
|
Posted: Tue Jul 28, 2015 5:40 pm Post subject: |
|
|
Buffoon wrote: | Thanks, will create that toolchain set. Is isl still used by gcc-5.2.0? |
I don't understand this question. gcc did never use any portage set. The set is just a convenient way to safe you from typing the whole list of packages manually. Of course, you can alternatively write a script (or just type the list of packages). |
|
Back to top |
|
|
Buffoon Veteran
Joined: 17 Jun 2015 Posts: 1369 Location: EU or US
|
Posted: Tue Jul 28, 2015 5:43 pm Post subject: |
|
|
mv wrote: | Buffoon wrote: | Thanks, will create that toolchain set. Is isl still used by gcc-5.2.0? |
I don't understand this question. gcc did never use any portage set. The set is just a convenient way to safe you from typing the whole list of packages manually. Of course, you can alternatively write a script (or just type the list of packages). |
Yes I understand creating a custom set is for convenience. I saw isl in your set and was wondering if isl is still needed in a 5.2.0 system. |
|
Back to top |
|
|
mv Watchman
Joined: 20 Apr 2005 Posts: 6747
|
Posted: Tue Jul 28, 2015 6:33 pm Post subject: |
|
|
Buffoon wrote: | if isl is still needed in a 5.2.0 system. |
Sorry, I had missread the word "isl" for "it" in the original question.
If you emerge gcc-5.2 with USE=graphite, it is needed. Although I don't use graphite, I think it doesn't hurt to keep the USE-flag on. |
|
Back to top |
|
|
Buffoon Veteran
Joined: 17 Jun 2015 Posts: 1369 Location: EU or US
|
Posted: Tue Jul 28, 2015 6:42 pm Post subject: |
|
|
Thanks! My bad, should have written dev-libs/isl. |
|
Back to top |
|
|
Ormaaj Guru
Joined: 28 Jan 2008 Posts: 319
|
Posted: Wed Jul 29, 2015 6:27 am Post subject: |
|
|
Graphite seems usable for the most part. I just rebuilt everything on one system with gcc-5.2 and graphite. I noticed a marked increase in the number of graphite-related failures compared to 4.9.3 (up from virtually zero to around 50).
Code: | # uname -a
Linux buildbox 4.1.3 #13 SMP PREEMPT Wed Jul 29 00:17:58 CDT 2015 x86_64 Intel(R) Core(TM) i7-5960X CPU @ 3.00GHz GenuineIntel GNU/Linux
# eix -I --only-names | wc -l
2405 |
Here's the full list of failures (including graphite fails):
Code: | declare -a fails='([0]="media-gfx/gimp-texturize-plugin" [1]="sci-libs/amd" [2]="app-shells/posh" [3]="media-libs/celt:0.5.1" [4]="sys-apps/v86d" \
[5]="sci-libs/camd" [6]="dev-util/cdecl" [7]="x11-libs/wxGTK:2.8" [8]="games-misc/gtklife" [9]="dev-java/antlr:3" [10]="games-puzzle/ltris" \
[11]="net-ftp/atftp" [12]="dev-vcs/cvs2svn" [13]="app-portage/layman-dbtools" [14]="net-misc/knemo:4" [15]="x11-apps/mesa-progs" \
[16]="dev-java/xml-xmlbeans:1" [17]="dev-dotnet/nini" [18]="dev-dotnet/dbus-sharp" [19]="dev-dotnet/nant" [20]="dev-dotnet/mono-addins" \
[21]="app-editors/qxmledit" [22]="dev-dotnet/dbus-sharp-glib" [23]="dev-dotnet/notify-sharp" [24]="dev-dotnet/gtk-sharp:3" \
[25]="sys-boot/grub:0" [26]="app-office/dia" [27]="kde-apps/libkdcraw:4" [28]="kde-apps/kolf:4" [29]="kde-apps/kpat:4" \
[30]="dev-util/valgrind" [31]="dev-libs/klibc" [32]="dev-util/mono-debugger" [33]="dev-util/emilpro" [34]="x11-misc/fraqtive" \
[35]="app-shells/ksh" [36]="media-plugins/caps-plugins" [37]="media-libs/libmikmod" [38]="dev-libs/protobuf" [39]="net-misc/nx" \
[40]="sci-libs/scipy" [41]="sys-fs/mtd-utils" [42]="dev-python/matplotlib" [43]="net-libs/libtirpc" [44]="net-misc/mosh" \
[45]="app-text/multitail" [46]="app-text/trang" [47]="app-text/jing" [48]="gnome-base/libgnomecanvas" [49]="www-client/phantomjs" \
[50]="dev-python/wsgiintercept" [51]="dev-qt/qthelp:4" [52]="media-libs/glu" [53]="media-libs/a52dec" [54]="media-libs/opus" \
[55]="media-libs/gavl" [56]="dev-lang/tcl:0" [57]="media-libs/schroedinger" [58]="media-libs/giflib" [59]="sys-libs/ncurses" \
[60]="dev-lang/lua:0" [61]="media-libs/aalib" [62]="media-libs/libbs2b" [63]="media-libs/libvpx" [64]="app-text/dvisvgm" \
[65]="dev-lang/tk:0" [66]="sys-apps/groff" [67]="dev-libs/libclc" [68]="dev-dotnet/libgdiplus" [69]="media-libs/openjpeg:0" \
[70]="media-libs/libwebp" [71]="media-gfx/graphviz" [72]="media-libs/phonon-gstreamer" [73]="media-libs/phonon" [74]="dev-qt/qtgui:4" \
[75]="sys-boot/grub:2" [76]="media-libs/x264" [77]="media-libs/opencv" [78]="media-libs/freeglut" [79]="x11-libs/libXi" \
[80]="sys-devel/gdb" [81]="media-gfx/fontforge" [82]="x11-misc/xscreensaver" [83]="net-analyzer/nmap" [84]="app-shells/mksh" \
[85]="dev-qt/qtwidgets" [86]="media-gfx/sane-backends" [87]="sci-mathematics/octave" [88]="app-text/cuneiform" [89]="app-text/ghostscript-gpl" \
[90]="sys-fs/udev" [91]="x11-base/xorg-server" [92]="net-analyzer/sslsniff" [93]="sci-libs/gdal" [94]="app-text/texlive-core" \
[95]="dev-java/icedtea:8" [96]="dev-games/simgear" [97]="games-simulation/flightgear")' |
And the list after rebuilding that set without graphite:
Code: | declare -a fails2='([0]="media-gfx/gimp-texturize-plugin" [1]="app-shells/posh" [2]="sys-apps/v86d" [3]="dev-util/cdecl" [4]="dev-java/antlr:3" \
[5]="games-puzzle/ltris" [6]="net-ftp/atftp" [7]="dev-vcs/cvs2svn" [8]="app-portage/layman-dbtools" [9]="net-misc/knemo:4" \
[10]="x11-apps/mesa-progs" [11]="dev-java/xml-xmlbeans:1" [12]="dev-dotnet/nini" [13]="dev-dotnet/dbus-sharp" [14]="dev-dotnet/nant" \
[15]="dev-dotnet/mono-addins" [16]="app-editors/qxmledit" [17]="dev-dotnet/dbus-sharp-glib" [18]="dev-dotnet/notify-sharp" [19]="dev-dotnet/gtk-sharp:3" \
[20]="sys-boot/grub:0" [21]="kde-apps/libkdcraw:4" [22]="dev-util/valgrind" [23]="dev-libs/klibc" [24]="dev-util/mono-debugger" \
[25]="dev-util/emilpro" [26]="x11-misc/fraqtive" [27]="app-shells/ksh" [28]="dev-libs/protobuf" [29]="sys-fs/mtd-utils" \
[30]="net-libs/libtirpc" [31]="net-misc/mosh" [32]="app-text/trang" [33]="app-text/jing" [34]="www-client/phantomjs" \
[35]="dev-python/wsgiintercept" [36]="media-libs/gavl" [37]="app-text/dvisvgm" [38]="dev-libs/libclc" [39]="media-libs/phonon-gstreamer" \
[40]="media-libs/phonon" [41]="sys-boot/grub:2" [42]="media-libs/opencv" [43]="app-shells/mksh" [44]="sys-fs/udev" [45]="net-analyzer/sslsniff")' |
Overall nothing critical in there. sys-fs/udev builds fine, that was an unrelated problem. Not all of those are even related to gcc-5.2.
I strongly suggest people not use emerge -e, because portage's resume support is way too terrible. You'll never get through the full list because portage wrongly considers your resume list "invalid" at random even if every single dependency is satisfied. AFAICT it's impossible to recover the resume list so you'll have to either start over or use qlop to guess at what's left to do. Instead just have eix output all installed packages and have portage build each in a loop individually.
Line breaks added by NeddySeagoon |
|
Back to top |
|
|
schorsch_76 Guru
Joined: 19 Jun 2012 Posts: 450
|
Posted: Wed Jul 29, 2015 5:10 pm Post subject: |
|
|
Here at my laptop the following packages failed to build:
Code: |
* The following 13 packages have failed to build, install, or execute
* postinst:
*
* (x11-apps/sessreg-1.1.0:0/0::gentoo, ebuild scheduled for merge), Log file:
* '/var/tmp/portage/x11-apps/sessreg-1.1.0/temp/build.log'
* (dev-libs/libgpg-error-1.13:0/0::gentoo, ebuild scheduled for merge), Log file:
* '/var/tmp/portage/dev-libs/libgpg-error-1.13/temp/build.log'
* (sys-libs/ncurses-5.9-r3:5/5::gentoo, ebuild scheduled for merge), Log file:
* '/var/tmp/portage/sys-libs/ncurses-5.9-r3/temp/build.log'
* (net-analyzer/bmon-3.6-r1:0/0::gentoo, ebuild scheduled for merge), Log file:
* '/var/tmp/portage/net-analyzer/bmon-3.6-r1/temp/build.log'
* (sys-block/thin-provisioning-tools-0.3.2-r1:0/0::gentoo, ebuild scheduled for merge), Log file:
* '/var/tmp/portage/sys-block/thin-provisioning-tools-0.3.2-r1/temp/build.log'
* (sys-devel/llvm-3.5.0:0/3.5::gentoo, ebuild scheduled for merge), Log file:
* '/var/tmp/portage/sys-devel/llvm-3.5.0/temp/build.log'
* (x11-base/xorg-server-1.16.4:0/1.16.1::gentoo, ebuild scheduled for merge), Log file:
* '/var/tmp/portage/x11-base/xorg-server-1.16.4/temp/build.log'
* (app-crypt/pinentry-0.9.0:0/0::gentoo, ebuild scheduled for merge), Log file:
* '/var/tmp/portage/app-crypt/pinentry-0.9.0/temp/build.log'
* (www-client/firefox-31.8.0:0/0::gentoo, ebuild scheduled for merge), Log file:
* '/var/tmp/portage/www-client/firefox-31.8.0/temp/build.log'
* (app-editors/emacs-24.4-r4:24/24::gentoo, ebuild scheduled for merge), Log file:
* '/var/tmp/portage/app-editors/emacs-24.4-r4/temp/build.log'
* (mail-client/thunderbird-31.8.0:0/0::gentoo, ebuild scheduled for merge), Log file:
* '/var/tmp/portage/mail-client/thunderbird-31.8.0/temp/build.log'
* (media-gfx/inkscape-0.48.5-r1:0/0::gentoo, ebuild scheduled for merge), Log file:
* '/var/tmp/portage/media-gfx/inkscape-0.48.5-r1/temp/build.log'
* (sci-electronics/kicad-20130518:0/0::gentoo, ebuild scheduled for merge), Log file:
* '/var/tmp/portage/sci-electronics/kicad-20130518/temp/build.log'
*
|
It is mostly a stable system, so a lot of patches are missing. The question is, go the way of ~amd64 or go the way of manual user-patches .... _________________ // valid again: I forgot about the git access. Now 1.2GB big. Start: 2015-06-25
git daily portage tree
Web: https://github.com/schorsch1976/portage
git clone https://github.com/schorsch1976/portage |
|
Back to top |
|
|
mv Watchman
Joined: 20 Apr 2005 Posts: 6747
|
Posted: Wed Jul 29, 2015 6:06 pm Post subject: |
|
|
Ormaaj wrote: | Graphite seems usable for the most part. |
Compilation with graphite had never been a problem. I had experienced several crashes and wrong behaviour with the compiled programs with <gcc-5 without any observable speed increase. |
|
Back to top |
|
|
mv Watchman
Joined: 20 Apr 2005 Posts: 6747
|
Posted: Wed Jul 29, 2015 6:20 pm Post subject: |
|
|
schorsch_76 wrote: | x11-apps/sessreg-1.1.0:0/0::gentoo | Compiled without problems here. Quote: | dev-libs/libgpg-error-1.13 | Version 1.19 compiled without problems here. Quote: | sys-libs/ncurses-5.9-r3 | Version 5.9-r4 compiled without problems Quote: | net-analyzer/bmon-3.6-r1
sys-block/thin-provisioning-tools-0.3.2-r1
mail-client/thunderbird-31.8.0
media-gfx/inkscape-0.48.5-r1
sci-electronics/kicad-20130518 | I did not try to compile these packages - I do not use them Quote: | sys-devel/llvm-3.5.0 | 3.6.1 compiled without problems here. Quote: | x11-base/xorg-server-1.16.4 | I tried only 1.17.1-r1 which needed a patch from bugzilla. Perhaps 1.17.2 already contains that patch (I did not try). Quote: | app-crypt/pinentry-0.9.0 | 0.9.0-r3 compiled without problems here. Quote: | www-client/firefox-31.8.0 | 39.0 compiled without problems here. Quote: | app-editors/emacs-24.4-r4 | 24.5 compiled without problems here. |
|
Back to top |
|
|
schorsch_76 Guru
Joined: 19 Jun 2012 Posts: 450
|
Posted: Wed Jul 29, 2015 7:30 pm Post subject: |
|
|
I got the following with patches building with gcc-5.2.0.
# preprocessor patch "-P"
x11-apps/sessreg-1.1.0:0/0::gentoo #546742
dev-libs/libgpg-error-1.13::gentoo #555820
# out of memory error / tmpfs
# fix: cleanup /var/tmp/portage
mail-client/thunderbird-31.8.0
media-gfx/inkscape-0.48.5-r1
The other packages i need to find out the reasons. I got notes and read through the gcc porting guide.
Edit: todays portage update fixed for me
- sys-block/thin-provisioning-tools as it upgraded to 0.4.x _________________ // valid again: I forgot about the git access. Now 1.2GB big. Start: 2015-06-25
git daily portage tree
Web: https://github.com/schorsch1976/portage
git clone https://github.com/schorsch1976/portage |
|
Back to top |
|
|
mv Watchman
Joined: 20 Apr 2005 Posts: 6747
|
Posted: Wed Jul 29, 2015 9:52 pm Post subject: |
|
|
schorsch_76 wrote: | x11-apps/sessreg-1.1.0 |
Whoops, you are right. I overlooked that I had added a patch here.
Quote: | # out of memory error / tmpfs
# fix: cleanup /var/tmp/portage |
This is not gcc-5 related. This is the disadvantage if you put /var/tmp on a tmpfs. |
|
Back to top |
|
|
schorsch_76 Guru
Joined: 19 Jun 2012 Posts: 450
|
Posted: Thu Jul 30, 2015 7:10 am Post subject: |
|
|
@mv: Yes of course, not gcc specific, but the reason was the tmpfs.
Now i am down to 4 packages which dont build with gcc-5.2.
bmon
llvm
kicad
xorg-server
llvm: I dont want to ~arch it, because mesa blocks <3.6 llvm. So i Keep it until a new mesa arrives.
The other tools i will add a project specific conf to use gcc 4.9 until they arrive at a version which builds under gcc 5.2
about xorg-server ... i am not sure if i want to go to ~arch. Right now i have under 10 ~arch packages....
right now the whole "world" gets rebuild to check all again. _________________ // valid again: I forgot about the git access. Now 1.2GB big. Start: 2015-06-25
git daily portage tree
Web: https://github.com/schorsch1976/portage
git clone https://github.com/schorsch1976/portage |
|
Back to top |
|
|
mv Watchman
Joined: 20 Apr 2005 Posts: 6747
|
Posted: Thu Jul 30, 2015 7:22 am Post subject: |
|
|
schorsch_76 wrote: | llvm: I dont want to ~arch it, because mesa blocks <3.6 llvm |
None of the newer (~arch) versions of mesa has this blocker (but I use USE=-llvm anyway for mesa). |
|
Back to top |
|
|
jean-michel.smith n00b
Joined: 04 Oct 2014 Posts: 5
|
Posted: Fri Jul 31, 2015 2:26 pm Post subject: For those needing java/icedtea |
|
|
For those who need java with gcc-5.2.0 (I do for certain remote consols), icedtea-7.2.6.1 from the java overlay compiles and runs fine in a gcc-5.2.0 environment. This was the only blocker for my desktop. |
|
Back to top |
|
|
|