Re: [gentoo-user] Clang has gone walkabout
On Monday 10 Apr 2017 11:08:05 Neil Bothwick wrote: > On Mon, 10 Apr 2017 17:13:28 +0800, Andrew Lowe wrote: > > Do we have any clang users out there? I've had clang installed > > > > on my machine for ages and a simple "clang test.c" will result in an > > executable. I can even nearly build my whole machine using clang, so > > its up and running. I've now just updated clang, from a working 3.9.1 > > to a 4.0.0-r1 and clang has now disappeared. If I type in "clang > > --version", I get "command not found". "whereis clang" only gives me > > the library dir. Doing "ls -la /usr/bin/cla*" gives me "No such file or > > directory" > > Try "qlist clang" so see what is installed, "qlist clang | grep bin/" > should find the executables. > > qlist is part of portage-utils, which you probably already have. I seem to recall clang clashing recently with some package, which required the clang USE flag to be unset - was it llvm? I suspect as a result clang is no longer installed on my laptop. -- Regards, Mick signature.asc Description: This is a digitally signed message part.
[gentoo-user] Clang has gone walkabout
Hi all, Do we have any clang users out there? I've had clang installed on my machine for ages and a simple "clang test.c" will result in an executable. I can even nearly build my whole machine using clang, so its up and running. I've now just updated clang, from a working 3.9.1 to a 4.0.0-r1 and clang has now disappeared. If I type in "clang --version", I get "command not found". "whereis clang" only gives me the library dir. Doing "ls -la /usr/bin/cla*" gives me "No such file or directory" I've run the install several times. I've even uninstalled both clang and llvm and then reinstalled and still the same. The only thing that I can think of is that whilst doing the original update, for some reason my machine crashed during the clang install. This may have screwed something up. Has anyone managed to do the install/update and have a working latest clang? Thoughts/comments greatly appreciated, Andrew p.s. Looking in the /usr/lib64/clang/4.0.0 dir shows plenty of libraries in there.
Re: [gentoo-user] Clang has gone walkabout
On Mon, 10 Apr 2017 17:13:28 +0800, Andrew Lowe wrote: > Do we have any clang users out there? I've had clang installed > on my machine for ages and a simple "clang test.c" will result in an > executable. I can even nearly build my whole machine using clang, so > its up and running. I've now just updated clang, from a working 3.9.1 > to a 4.0.0-r1 and clang has now disappeared. If I type in "clang > --version", I get "command not found". "whereis clang" only gives me > the library dir. Doing "ls -la /usr/bin/cla*" gives me "No such file or > directory" Try "qlist clang" so see what is installed, "qlist clang | grep bin/" should find the executables. qlist is part of portage-utils, which you probably already have. -- Neil Bothwick Celery is not food. It is a member of the plywood family. pgpK93Ywpt_pS.pgp Description: OpenPGP digital signature
[gentoo-user] Re: Clang has gone walkabout
On 04/10/2017 12:13 PM, Andrew Lowe wrote: Hi all, Do we have any clang users out there? I've had clang installed on my machine for ages and a simple "clang test.c" will result in an executable. I can even nearly build my whole machine using clang, so its up and running. I've now just updated clang, from a working 3.9.1 to a 4.0.0-r1 and clang has now disappeared. If I type in "clang --version", I get "command not found". "whereis clang" only gives me the library dir. Doing "ls -la /usr/bin/cla*" gives me "No such file or directory" This is normal. You just need to logout of the system and in again for the updated environment. Some packages can modify the environment, so it's usually best to logout/login after every update.
Re: [gentoo-user] Clang has gone walkabout
On 10/04/17 18:08, Neil Bothwick wrote: On Mon, 10 Apr 2017 17:13:28 +0800, Andrew Lowe wrote: Do we have any clang users out there? I've had clang installed on my machine for ages and a simple "clang test.c" will result in an executable. I can even nearly build my whole machine using clang, so its up and running. I've now just updated clang, from a working 3.9.1 to a 4.0.0-r1 and clang has now disappeared. If I type in "clang --version", I get "command not found". "whereis clang" only gives me the library dir. Doing "ls -la /usr/bin/cla*" gives me "No such file or directory" Try "qlist clang" so see what is installed, "qlist clang | grep bin/" should find the executables. qlist is part of portage-utils, which you probably already have. Done as requested. There are 41 files found with clang in their name and they are all on the dir: /usr/lib/llvm/4/bin/ I'm no whiz bang sys-admin but that doesn't seem right to me. There is clang and clang++ and a whole lot of stuff sym linked to provide all the various permutations and combinations of names in there. But there is nothing in my path that points to that dir. I'll have to have a look at the ebuild to see if a symlink or something is not being applied. Any other thoughts appreciated, Andrew
Re: [gentoo-user] Clang has gone walkabout
On April 10, 2017 12:41:54 PM GMT+02:00, Andrew Lowe <a...@wht.com.au> wrote: >On 10/04/17 18:08, Neil Bothwick wrote: >> On Mon, 10 Apr 2017 17:13:28 +0800, Andrew Lowe wrote: >> >>> Do we have any clang users out there? I've had clang installed >>> on my machine for ages and a simple "clang test.c" will result in an >>> executable. I can even nearly build my whole machine using clang, so >>> its up and running. I've now just updated clang, from a working >3.9.1 >>> to a 4.0.0-r1 and clang has now disappeared. If I type in "clang >>> --version", I get "command not found". "whereis clang" only gives me >>> the library dir. Doing "ls -la /usr/bin/cla*" gives me "No such file >or >>> directory" >> >> Try "qlist clang" so see what is installed, "qlist clang | grep bin/" >> should find the executables. >> >> qlist is part of portage-utils, which you probably already have. >> >> > > Done as requested. There are 41 files found with clang in their name >and they are all on the dir: > > /usr/lib/llvm/4/bin/ > > I'm no whiz bang sys-admin but that doesn't seem right to me. There is > >clang and clang++ and a whole lot of stuff sym linked to provide all >the >various permutations and combinations of names in there. But there is >nothing in my path that points to that dir. I'll have to have a look at > >the ebuild to see if a symlink or something is not being applied. > > Any other thoughts appreciated, > > Andrew Try those and see if they respond correctly. If yes, add that dir to your PATH. -- Joost -- Sent from my Android device with K-9 Mail. Please excuse my brevity.
[gentoo-user] Comparatively slow clang startup vs. gcc - reasons?
(paging Michał Górny..) I've noticed that clang has gotten really slow, and while it does great analysis during the various compilation stages (which is fine), the startup itself is a major contributor to perceived slowness. Nothing demonstrates this better than running ./configure in a random project - here nghttp2: $time ./configure ./configure 5.83s user 1.10s system 100% cpu 6.866 total ./configure 5.73s user 1.04s system 101% cpu 6.669 total ./configure 5.74s user 1.04s system 101% cpu 6.671 total $time ./configure CC=clang CXX=clang++ ./configure CC=clang CXX=clang++ 21.43s user 2.56s system 99% cpu 24.011 total ./configure CC=clang CXX=clang++ 21.37s user 2.61s system 100% cpu 23.911 total ./configure CC=clang CXX=clang++ 21.56s user 2.51s system 100% cpu 23.995 total That's almost 3 times slower doing not much at all. A major contributor to this is the relatively large number of shared objects being loaded as a consequence of building llvm with shared objects, incurring large relocation costs compared to short-lived runs. gcc on the other hand consists of mostly-monolithic binaries, minus the few usual suspects like zlib etc. A brief run with "perf record" compiling helloworld.c and comparing the "perf report" output between gcc and clang confirms that clang spends the vast majority of its time in ld.so. Can we do something about this? I remember that llvm had optional static libs, which apparently were removed completely. Is there something that can be done with linker tricks (better relocation info?) when building llvm/clang to speed up the .so loading? curious, Holger
Re: [gentoo-user] Emerge trouble with firefox and thunderbird ...
On Sat, Mar 09, 2024 at 02:45:02PM +, Peter Humphrey wrote > >>> Running pre-merge checks for sys-libs/compiler-rt-18.1.0 > * Building using a compiler other than clang may result in broken atomics > * library. Enable USE=clang unless you have a very good reason not to. According to https://packages.gentoo.org/packages/sys-libs/compiler-rt "sys-libs/compiler-rt" is a "Compiler runtime library for clang (built-in part)" so like... dohhh. Use clang to support clang. The real question is what else, besides clang and its libraries, are you building that requires clang? > Does the compiler-rt ebuild override USE in make.conf? You can build it with USE="-clang", but that defeats the entire purpose of building compiler-rt. -- Roses are red Roses are blue Depending on their velocity Relative to you
[gentoo-user] ebuild for gcc-independent llvm and clang for linux
Hello everybody, I'm going to build and install llvm and clang packages in my Linux PC as 1. Binaries of llvm and claag installed finally were compiled by clang. 2. The binaries of 1. depend on libc++.so instead of libstdc++.so. 3. libc++.so was installed by portage. 4. The binaries of llvm, clang, and libc++.so are not linked with libgcc.
[gentoo-user] Is there a reason why LLVM/Clang ebuilds don't support "mutislot"?
Other distros like Ubuntu support the installation of multiple versions of LLVM/Clang side by side. One of the things Clang is really good at is support for the most recently approved upcoming features of the C++17 standard. The best support for testing such features is with the latest sys-devel/llvm-. However if I want to compile Mesa against a stable version LLVM/Clang as well, I don't get that option.
Re: [gentoo-user] Emerge trouble with firefox and thunderbird ...
On Sun, Mar 10, 2024 at 08:43:46PM +0100, ralfconn wrote > Given the warning message reported by Peter ("Enable USE=clang unless > you have a very good reason not to.") That message comes from sys-libs/compiler-rt which is a dedicated runtime lib for clang. It makes sense to use clang if you're building the clang toolchain... just like using gcc if you're building the gcc toolchain. -- Roses are red Roses are blue Depending on their velocity Relative to you
Re: [gentoo-user] Re: Mammoth emerge ...
. Chromium switched to 'clang++ v5.x' as its primary compiler. Why? The Chromium devs are using 'c++' features supported in gcc v8+. . So ... first compile run is with 'gcc' ... then Chromium is re-compiled with 'clang++'. That is what I am seeing ( console && log wise ). 2 Compile runs ... twice the time. . No gold linker setup on my system. Just how is 'clang++' supposed to work with 'ld.bfd'? . As far as I can tell, all optimization depending on '-march= / -mtune= ' is still discarded, as well. ( clang / clang ++, does not seem to accept the '-march= / -mtune= / -O2 / -pipe' switches either. ) . Corbin
Re: [gentoo-user] Emerge trouble with firefox and thunderbird ...
On Saturday, 9 March 2024 12:49:33 GMT Walter Dnes wrote: > I have "-clang" in USE in make.conf and no problems resulting from it. > clang seems to be another "solution in search of a problem" along the > lines of rust and cups and systemd and hatbuzz, etc, which keep trying > to worm their way into everybody's linux system. When I tried USE=-clang emerge -uaDvN @world, I got this: [...] >>> Running pre-merge checks for sys-libs/compiler-rt-18.1.0 * Building using a compiler other than clang may result in broken atomics * library. Enable USE=clang unless you have a very good reason not to. Does the compiler-rt ebuild override USE in make.conf? -- Regards, Peter.
Re: [gentoo-user] Emerge trouble with firefox and thunderbird ...
Il 10/03/24 15:08, Peter Humphrey ha scritto: On Sunday, 10 March 2024 07:17:27 GMT Walter Dnes wrote: So there are at least 2 people who've found out that Firefox can and *MUST* be built with USE="-clang". Ah. I'll change my USE flag straight away. Thanks Walter. This got me wondering, here I don't have clang in my make.config. The main packages using it are: [I] www-client/firefox Installed versions: 123.0.1(rapid)(21:16:32 03/06/24)(X clang ... [I] mail-client/thunderbird Installed versions: 115.8.1(21:14:19 03/07/24)(X clang ... [I] app-office/libreoffice Installed versions: 7.6.5.2^t(21:42:53 03/06/24)(... -clang ... libreoffice sees it unset because it is not present in the global uses. firefox and thunderbird instead set it in the ebuild: $ grep clang /var/db/repos/gentoo/www-client/firefox/firefox-123.0.ebuild IUSE="+clang cpu_flags_arm_neon dbus debug eme-free hardened hwaccel" $ grep clang /var/db/repos/gentoo/mail-client/thunderbird/thunderbird-115.8.1.ebuild IUSE="+clang cpu_flags_arm_neon dbus debug eme-free hardened hwaccel" Both packages have no issues here with +clang. Given the warning message reported by Peter ("Enable USE=clang unless you have a very good reason not to.") and the fact that gentoo developers decided to switch it on specifically for these packages it would probably be a better idea to file a bugzilla rather than forcing the use of GCC, which might fix the issue now but lead to problems later. As usual, YMMV. raf
Re: [gentoo-user] Clang has gone walkabout
On 10/04/17 18:57, J. Roeleveld wrote: On April 10, 2017 12:41:54 PM GMT+02:00, Andrew Lowe <a...@wht.com.au> wrote: On 10/04/17 18:08, Neil Bothwick wrote: On Mon, 10 Apr 2017 17:13:28 +0800, Andrew Lowe wrote: Do we have any clang users out there? I've had clang installed on my machine for ages and a simple "clang test.c" will result in an executable. I can even nearly build my whole machine using clang, so its up and running. I've now just updated clang, from a working 3.9.1 to a 4.0.0-r1 and clang has now disappeared. If I type in "clang --version", I get "command not found". "whereis clang" only gives me the library dir. Doing "ls -la /usr/bin/cla*" gives me "No such file or directory" Try "qlist clang" so see what is installed, "qlist clang | grep bin/" should find the executables. qlist is part of portage-utils, which you probably already have. Done as requested. There are 41 files found with clang in their name and they are all on the dir: /usr/lib/llvm/4/bin/ I'm no whiz bang sys-admin but that doesn't seem right to me. There is clang and clang++ and a whole lot of stuff sym linked to provide all the various permutations and combinations of names in there. But there is nothing in my path that points to that dir. I'll have to have a look at the ebuild to see if a symlink or something is not being applied. Any other thoughts appreciated, Andrew Try those and see if they respond correctly. If yes, add that dir to your PATH. -- Joost They work as expected and I can add the dir to the path with no problems, I'm more concerned about why I have add the path - is the ebuild screwed up in some way? What is the portage/ebuild doco like? Is it well documented or are there gaping holes that lead to frustration - my level of understanding of coding is 25 years of C/C++ coding on CAD systems & engineering applications and even though I run a Gentoo box as my default machine, I've never had the need to get into bash scripting - but might. Andrew
Re: [gentoo-user] Clang has gone walkabout
On 17-04-10 at 20:48, Andrew Lowe wrote: > On 10/04/17 18:57, J. Roeleveld wrote: > > On April 10, 2017 12:41:54 PM GMT+02:00, Andrew Lowe <a...@wht.com.au> > > wrote: > >> On 10/04/17 18:08, Neil Bothwick wrote: > >>> On Mon, 10 Apr 2017 17:13:28 +0800, Andrew Lowe wrote: > >>> > >>>> Do we have any clang users out there? I've had clang installed > >>>> on my machine for ages and a simple "clang test.c" will result in an > >>>> executable. I can even nearly build my whole machine using clang, so > >>>> its up and running. I've now just updated clang, from a working > >> 3.9.1 > >>>> to a 4.0.0-r1 and clang has now disappeared. If I type in "clang > >>>> --version", I get "command not found". "whereis clang" only gives me > >>>> the library dir. Doing "ls -la /usr/bin/cla*" gives me "No such file > >> or > >>>> directory" > >>> > >>> Try "qlist clang" so see what is installed, "qlist clang | grep bin/" > >>> should find the executables. > >>> > >>> qlist is part of portage-utils, which you probably already have. > >>> > >>> > >> > >>Done as requested. There are 41 files found with clang in their name > >> and they are all on the dir: > >> > >>/usr/lib/llvm/4/bin/ > >> > >>I'm no whiz bang sys-admin but that doesn't seem right to me. There is > >> > >> clang and clang++ and a whole lot of stuff sym linked to provide all > >> the > >> various permutations and combinations of names in there. But there is > >> nothing in my path that points to that dir. I'll have to have a look at > >> > >> the ebuild to see if a symlink or something is not being applied. > >> > >>Any other thoughts appreciated, > >> > >>Andrew > > > > Try those and see if they respond correctly. > > If yes, add that dir to your PATH. > > > > -- > > Joost > > > > They work as expected and I can add the dir to the path with no > problems, I'm more concerned about why I have add the path - is the > ebuild screwed up in some way? > > What is the portage/ebuild doco like? Is it well documented or are > there gaping holes that lead to frustration - my level of understanding > of coding is 25 years of C/C++ coding on CAD systems & engineering > applications and even though I run a Gentoo box as my default machine, > I've never had the need to get into bash scripting - but might. > > Andrew > Try running `env-update && source /etc/profile'. Your path should be extended by /etc/profile.env which is generated from /etc/env.d/10llvm-9995. -- Simon Thelen
Re: [gentoo-user] Clang has gone walkabout
On 10/04/17 20:58, Simon Thelen wrote: On 17-04-10 at 20:48, Andrew Lowe wrote: On 10/04/17 18:57, J. Roeleveld wrote: On April 10, 2017 12:41:54 PM GMT+02:00, Andrew Lowe <a...@wht.com.au> wrote: On 10/04/17 18:08, Neil Bothwick wrote: On Mon, 10 Apr 2017 17:13:28 +0800, Andrew Lowe wrote: Do we have any clang users out there? I've had clang installed on my machine for ages and a simple "clang test.c" will result in an executable. I can even nearly build my whole machine using clang, so its up and running. I've now just updated clang, from a working 3.9.1 to a 4.0.0-r1 and clang has now disappeared. If I type in "clang --version", I get "command not found". "whereis clang" only gives me the library dir. Doing "ls -la /usr/bin/cla*" gives me "No such file or directory" Try "qlist clang" so see what is installed, "qlist clang | grep bin/" should find the executables. qlist is part of portage-utils, which you probably already have. Done as requested. There are 41 files found with clang in their name and they are all on the dir: /usr/lib/llvm/4/bin/ I'm no whiz bang sys-admin but that doesn't seem right to me. There is clang and clang++ and a whole lot of stuff sym linked to provide all the various permutations and combinations of names in there. But there is nothing in my path that points to that dir. I'll have to have a look at the ebuild to see if a symlink or something is not being applied. Any other thoughts appreciated, Andrew Try those and see if they respond correctly. If yes, add that dir to your PATH. -- Joost They work as expected and I can add the dir to the path with no problems, I'm more concerned about why I have add the path - is the ebuild screwed up in some way? What is the portage/ebuild doco like? Is it well documented or are there gaping holes that lead to frustration - my level of understanding of coding is 25 years of C/C++ coding on CAD systems & engineering applications and even though I run a Gentoo box as my default machine, I've never had the need to get into bash scripting - but might. Andrew Try running `env-update && source /etc/profile'. Your path should be extended by /etc/profile.env which is generated from /etc/env.d/10llvm-9995. SUCCESS!! Thanks
[gentoo-user] Re: llvm compile error
Before this gets out of hand.. On Fri, 24 Jan 2014 19:44:38 +0100, Silvio Siefke wrote: llvm-3.4 is not hardmasked. It is marked ~x86 which is something very different. Correct, i mean Clang in version 3.4 is hardmasked. I want install clang alone llvm i not need. I think mesa use llvm too but i has not set the use flag. 1) clang-3.4 is not hardmasked either, so you need to set ~x86 keyword for that package as well as for llvm - because: 2) clang is only a metapackage, without any contents on its own. The compiler binaries are part of the llvm package. You simply cannot have clang without llvm. 3) FreeBSD indeed uses clang as default compiler, but uses 3.3 since 3.4 still has quite a few new and exciting bugs. Hope that clears things up. -h
Re: [gentoo-user] clang <<==>> gcc ?
Rasmus.thomsen <rasmus.thom...@protonmail.com> schrieb am So., 30. Apr. 2017, 12:19: > Hello, > > it's entirely possible to replace gcc for clang for *most* packages, > however some will not build currently and will require you to set up a > package.env file with entries for those packages (like described on clang's > wiki entry). Clang usually compiles faster than GCC does, but produces > slower binaries (at least for me). Also, clang offers flto=thin, which > doesn't require as much ram as gcc's lto > > Regards, > Rasmus > > > Sent from ProtonMail mobile > > > > Original Message > On 30 Apr 2017, 12:11, < tu...@posteo.de> wrote: > > > Hi, > > before I do a lot of reconfiguring, recompiling and finally > do the same thing again in the opposite direction: > > What are the experiences to replace gcc with clang for either > only userland tools or the whole system (with haveing gcc as > fallback)? > Is it worth the effort? > What are the benefits and the drawbacks? > > Thanks for any input in advance! > Cheers > Meino > > Hi, There is a tracking bug about what does not compile with clang. https://bugs.gentoo.org/show_bug.cgi?id=408963 >
Re: [gentoo-user] Failed to emerge www-client/firefox-68.4.1, Log file:
On Thu, Jan 16, 2020 at 4:46 PM Daniel Frey wrote: > > On 1/16/20 10:40 AM, Valmor de Almeida wrote: > > Hello list, > > I have been trying for a while getting firefox emerged; no luck. > > Inputs appreciated. > > Thanks, > > -- > > Valmor > > > > * sys-devel/clang:9 is missing! Cannot use LLVM slot 9 ... > > Do you have this installed? > > > What's the output of `equery list llvm clang`? > > Is it possible llvm updated but clang did not? The message sounds like > they depend on each other. > > Dan > Here it is: -> equery list llvm clang * Searching for llvm ... [IP-] [ ] sys-devel/llvm-8.0.1:8 * Searching for clang ... [IP-] [ ] sys-devel/clang-8.0.1:8 if slot 9 means version 9, it is not installed and I don't know why it does not get pulled in by portage. Thanks, -- Valmor
Re: [gentoo-user] Failed to emerge www-client/firefox-68.4.1, Log file:
On 1/16/20 10:40 AM, Valmor de Almeida wrote: Hello list, I have been trying for a while getting firefox emerged; no luck. Inputs appreciated. Thanks, -- Valmor * sys-devel/clang:9 is missing! Cannot use LLVM slot 9 ... Do you have this installed? What's the output of `equery list llvm clang`? Is it possible llvm updated but clang did not? The message sounds like they depend on each other. Dan
[gentoo-user] Re: ebuild for gcc-independent llvm and clang for linux
Sorry, I mistook the operation of Gmail. I will send with full-written article again. 2013/12/26 やまぐちたかゆき tyamaguchi.gen...@gmail.com: Hello everybody, I'm going to build and install llvm and clang packages in my Linux PC as 1. Binaries of llvm and claag installed finally were compiled by clang. 2. The binaries of 1. depend on libc++.so instead of libstdc++.so. 3. libc++.so was installed by portage. 4. The binaries of llvm, clang, and libc++.so are not linked with libgcc.
[gentoo-user] Re: sys-devel/llvm and LLVM_TARGETS
On 2022-03-12, Nikos Chantziaras wrote: > On 12/03/2022 18:03, Grant Edwards wrote: >> On 2022-03-12, Nikos Chantziaras wrote: >>> On 12/03/2022 10:43, Dale wrote: >>>> https://bugs.gentoo.org/767700 >>>> >>>> Is that the one? It mentions the target but I don't quite understand >>>> the why. The biggest thing, will this break something if I let it do >>>> it? >>> >>> No. Unlike GCC, LLVM/Clang is always a cross-compiler. >> >> You can't use LLVM/Clang to compile for the host on which it's >> running? > > Why not? Because "LLVM/Clang is always a cross compiler". A cross compiler is a compiler that compiles for a target architecture/OS different than that of the host on which it is running. Therefore, LLVM/Clang always compiles for a target architecture/OS different than that of the host on which it is running. -- Grant
[gentoo-user] clang <<==>> gcc ?
Hi, before I do a lot of reconfiguring, recompiling and finally do the same thing again in the opposite direction: What are the experiences to replace gcc with clang for either only userland tools or the whole system (with haveing gcc as fallback)? Is it worth the effort? What are the benefits and the drawbacks? Thanks for any input in advance! Cheers Meino
Re: [gentoo-user] Re: sys-devel/llvm and LLVM_TARGETS
On Sunday, 13 March 2022 08:03:04 GMT Wols Lists wrote: > To the best of my knowledge, gcc can't currently handle Rust, so yes > LLVM is needed. (Not Clang, because it isn't C :-) (Although Firefox > probably also uses loads of C, so Clang would be needed for that.) And that's why we have dev-lang/rust on hand, no? -- Regards, Peter.
Re: [gentoo-user] Emerge trouble with firefox and thunderbird ...
On Saturday, 9 March 2024 19:37:40 GMT Walter Dnes wrote: > On Sat, Mar 09, 2024 at 02:45:02PM +, Peter Humphr > The real question is what else, besides clang and its libraries, are you > building that requires clang? Firefox. -- Regards, Peter.
[gentoo-user] Re: ebuild for gcc-independent llvm and clang for linux
Sorry for earlier mis-send. I send the article again. I'm going to build and install llvm and clang packages in my Linux PC as 1. Binaries of llvm and claag installed finally were compiled by clang. 2. The binaries of 1. depend on libc++.so instead of libstdc++.so. 3. libc++.so was installed by portage. 4. The binaries of llvm, clang, and libc++.so are not linked with libgcc. If linkage with libraries like libgcc is necessary, compiler_rt is linked. Checking the ebuilds of llvm, I found sys-libs/libcxx in DEPEND. But from the comment of revision, append libcxx in dependency is for freebsd. I want to ask I. Building packages described above can be created using ebuilds in main repository in linux? II. if I. is not possible, any ebuild of llvm and clang exists anywhere? If ebuilds that fill II. are not exist, I will to try create such ebuilds.
Re: [gentoo-user] mesa-12.0.1 fails to emerge
On 02/07/2017 05:02 PM, Mick wrote: > How could I go beyond this point? > > = > /var/tmp/portage/media- > libs/mesa-12.0.1/work/mesa-12.0.1/src/gallium/state_track > ers/clover/llvm/invocation.cpp:212:75: error: no matching function for call > to > ‘ > clang::CompilerInvocation::setLangDefaults(clang::LangOptions&, > clang::InputKind > , llvm::Triple, clang::LangStandard::Kind)’ > = > > Extract from compile log attached. > > I am running sys-devel/clang-runtime-3.9.1 like so: > > Installed versions: 3.9.1(21:25:30 07/02/17)(openmp -libcxx ABI_MIPS="- > n32 -n64 -o32" ABI_PPC="-32 -64" ABI_S390="-32 -64" ABI_X86="64 -32 -x32") > --- If you would please, post if the USE flag "opencl" is set in "make.conf", or in "package.use" for any packages. Corbin
Re: [gentoo-user] mesa-12.0.1 fails to emerge
On Tuesday 07 Feb 2017 18:24:06 Corbin Bird wrote: > On 02/07/2017 05:02 PM, Mick wrote: > > How could I go beyond this point? > > > > = > > /var/tmp/portage/media- > > libs/mesa-12.0.1/work/mesa-12.0.1/src/gallium/state_track > > ers/clover/llvm/invocation.cpp:212:75: error: no matching function for > > call to ‘ > > clang::CompilerInvocation::setLangDefaults(clang::LangOptions&, > > clang::InputKind > > , llvm::Triple, clang::LangStandard::Kind)’ > > = > > > > Extract from compile log attached. > > > > I am running sys-devel/clang-runtime-3.9.1 like so: > > Installed versions: 3.9.1(21:25:30 07/02/17)(openmp -libcxx > > ABI_MIPS="- > > > > n32 -n64 -o32" ABI_PPC="-32 -64" ABI_S390="-32 -64" ABI_X86="64 -32 -x32") > > --- > If you would please, post if the USE flag "opencl" is set in > "make.conf", or in "package.use" for any packages. > > > Corbin Yes! It had sneaked in make.conf. I removed it and am re-emerging now. Thanks. :-) -- Regards, Mick signature.asc Description: This is a digitally signed message part.
Re: [gentoo-user] clang <<==>> gcc ?
Hello, it's entirely possible to replace gcc for clang for *most* packages, however some will not build currently and will require you to set up a package.env file with entries for those packages (like described on clang's wiki entry). Clang usually compiles faster than GCC does, but produces slower binaries (at least for me). Also, clang offers flto=thin, which doesn't require as much ram as gcc's lto Regards, Rasmus Sent from ProtonMail mobile Original Message On 30 Apr 2017, 12:11, wrote: Hi, before I do a lot of reconfiguring, recompiling and finally do the same thing again in the opposite direction: What are the experiences to replace gcc with clang for either only userland tools or the whole system (with haveing gcc as fallback)? Is it worth the effort? What are the benefits and the drawbacks? Thanks for any input in advance! Cheers Meino
Re: [gentoo-user] Re: Mammoth emerge ...
On Thursday, 3 May 2018 18:00:45 BST Corbin Bird wrote: > . > Chromium switched to 'clang++ v5.x' as its primary compiler. > Why? > The Chromium devs are using 'c++' features supported in gcc v8+. > . > So ... first compile run is with 'gcc' ... then Chromium is re-compiled > with 'clang++'. > That is what I am seeing ( console && log wise ). > 2 Compile runs ... twice the time. > . > No gold linker setup on my system. > Just how is 'clang++' supposed to work with 'ld.bfd'? > . > As far as I can tell, all optimization depending on '-march= / -mtune= ' > is still discarded, as well. > ( clang / clang ++, does not seem to accept the '-march= / -mtune= / -O2 > / -pipe' switches either. ) > . > Corbin Thanks Corbin, it makes sense. Back in November build time jumped from 8 to 20 hours on my system. It's only gone further downhill since. :-( I will give USE="jumbo-build" a spin later to see what improvement I may get. -- Regards, Mick signature.asc Description: This is a digitally signed message part.
Re: [gentoo-user] Re: llvm compile error
Hello, On Thu, 23 Jan 2014 14:26:25 -0800 walt w41...@gmail.com wrote: Hi Silvio. You're installing an ~x86 package on an x86 machine, I think. Are you doing that on purpose? I don't know if that is causing the error, but it sometimes causes problems. Have i something not understand, i make emerge -av clang and become gentoomobile siefke # emerge -av clang These are the packages that would be merged, in order: Calculating dependencies... done! [ebuild N~] sys-devel/llvm-3.4:0/3.4 USE=clang libffi ncurses python static-analyzer xml -debug -doc -gold -multitarget -ocaml {-test} -udis86 PYTHON_TARGETS=python2_7 -pypy2_0 -python2_6 VIDEO_CARDS=-radeon 0 kB [ebuild N~] sys-devel/clang-3.4-r100:0/3.4 USE=python static-analyzer -debug -multitarget 0 kB Total: 2 packages (2 new), Size of downloads: 0 kB The following keyword changes are necessary to proceed: (see package.accept_keywords in the portage(5) man page for more details) # required by sys-devel/clang-3.4-r100 # required by clang (argument) =sys-devel/llvm-3.4 ~x86 # required by sys-devel/llvm-3.4[clang] =sys-devel/clang-3.4-r100 ~x86 Use --autounmask-write to write changes to config files (honoring CONFIG_PROTECT). Carefully examine the list of proposed changes, paying special attention to mask or keyword changes that may expose experimental or unstable packages. Then emerge --autounmask-write -av clang and after dispatch-conf i run emerge -av clang and become the error. I not understand what you mean? -- Found PythonInterp: /var/tmp/portage/sys-devel/llvm-3.4/temp/python2.7/bin/python (found version 2.7.5) -- Constructing LLVMBuild project information CMake Error at CMakeLists.txt:332 (message): Unexpected failure executing llvm-build: Usage: llvm-build [options] llvm-build: error: invalid native target: 'X86' (not in project) Thank you for help Nice Day
Re: [gentoo-user] Firefox crashes on some www-pages on a newer Gentoo system
> > > I would do something like 'emerge -1 xorg-server xorg-drivers > > @x11-module-rebuild mesa llvm clang' then restart X and try again. > > Thank you for your reply. > > Initially, I understood the above recomendation as the suggestion to > rebuild the packages mentioned above with different use flags. > > Do you think that rebuilding them with the same use flags may help? > It can, for example, if the major version of the kernel has changed but you havent rebuild xorg since (I only recall having this issue once). If you use newuse and changed-deps with emerge its probably less likely to find other issues. > The said problem appeared just since the intallation of the new > Gentoo system in January-February 2018 and not since changing > the major version of gcc this spring. > > P.S. clang is not installed on my Gentoo system at all. > > I've checked the dependencies on my system, and firefox is pulling that in for me, but checking the ebuilds you can see it becomes a dependency from v60 onwards; /usr/portage/www-client/firefox $ ls files/ firefox-52.9.0.ebuild Manifest firefox-52.6.0.ebuild firefox-60.1.0.ebuild metadata.xml firefox-52.8.0.ebuild firefox-61.0-r1.ebuild /usr/portage/www-client/firefox $ grep clang * grep: files: Is a directory firefox-60.1.0.ebuild:>=sys-devel/clang-4.0.1 firefox-60.1.0.ebuild:has_version "sys-devel/clang:${LLVM_SLOT}" firefox-61.0-r1.ebuild:>=sys-devel/clang-4.0.1 firefox-61.0-r1.ebuild:has_version "sys-devel/clang:${LLVM_SLOT}" /usr/portage/www-client/firefox $
Re: [gentoo-user] Firefox crashes on some www-pages on a newer Gentoo system
2018-08-02 3:16 GMT+03:00 Adam Carter : >> > I would do something like 'emerge -1 xorg-server xorg-drivers >> > @x11-module-rebuild mesa llvm clang' then restart X and try again. >> >> Thank you for your reply. >> >> Initially, I understood the above recomendation as the suggestion to >> rebuild the packages mentioned above with different use flags. >> >> Do you think that rebuilding them with the same use flags may help? > > > It can, for example, if the major version of the kernel has changed but you > havent rebuild xorg since (I only recall having this issue once). Ok, thank you. I will try to do it in a two weeks. (Currently, is still too hot to start so massive recompilation.) > If you use newuse and changed-deps with emerge its probably less likely to > find other issues. > >> >> The said problem appeared just since the intallation of the new >> Gentoo system in January-February 2018 and not since changing >> the major version of gcc this spring. >> >> P.S. clang is not installed on my Gentoo system at all. >> > > I've checked the dependencies on my system, and firefox is pulling that in > for me, but checking the ebuilds you can see it becomes a dependency from > v60 onwards; > > /usr/portage/www-client/firefox $ ls > files/ firefox-52.9.0.ebuild Manifest > firefox-52.6.0.ebuild firefox-60.1.0.ebuild metadata.xml > firefox-52.8.0.ebuild firefox-61.0-r1.ebuild > /usr/portage/www-client/firefox $ grep clang * > grep: files: Is a directory > firefox-60.1.0.ebuild:>=sys-devel/clang-4.0.1 > firefox-60.1.0.ebuild:has_version "sys-devel/clang:${LLVM_SLOT}" > firefox-61.0-r1.ebuild:>=sys-devel/clang-4.0.1 > firefox-61.0-r1.ebuild:has_version "sys-devel/clang:${LLVM_SLOT}" > /usr/portage/www-client/firefox $ Ok, thank you. I still have FF version 5.8.0 (64 bit). It is the latest stable version on amd64.
Re: [gentoo-user] Re: sys-devel/llvm and LLVM_TARGETS
Peter Humphrey wrote: > On Sunday, 13 March 2022 08:03:04 GMT Wols Lists wrote: > >> To the best of my knowledge, gcc can't currently handle Rust, so yes >> LLVM is needed. (Not Clang, because it isn't C :-) (Although Firefox >> probably also uses loads of C, so Clang would be needed for that.) > And that's why we have dev-lang/rust on hand, no? > So, it takes rust, gcc, clang and maybe some other stuff to make Firefox??? Jeez, that sounds complicated. o_O Sounds like a recipe for a soup or something. ROFL I might add, I did finish the updates and everything compiled and works fine. So, it is safe to have all those options enabled. Weird but safe. ;-) Dale :-) :-)
Re: [gentoo-user] Re: sys-devel/llvm and LLVM_TARGETS
On 12/03/2022 17:36, Dale wrote: I've sort of read about llvm and clang and I seem to recall things like Firefox needing them or something. I've just watched firefox emerging (yes I know, paint drying and all that :-), and there's loads of Rust code in there. To the best of my knowledge, gcc can't currently handle Rust, so yes LLVM is needed. (Not Clang, because it isn't C :-) (Although Firefox probably also uses loads of C, so Clang would be needed for that.) Cheers, Wol
Re: [gentoo-user] emerge failes due to no valid source for pythonexec-2.2 when I have pythonexec-2.4 installed???
On Sun, 6 Feb 2022 at 21:52, Steven Lembark wrote: > emerge: there are no ebuilds built with USE flags to satisfy > ">=dev-lang/python-exec-2:2/2=[python_targets_python3_6]". > !!! One of the following packages is required to complete your request: > - dev-lang/python-exec-2.4.8::gentoo (Missing IUSE: python_targets_python3_6) > (dependency required by "sys-devel/clang-9.0.1::gentoo" [installed]) > (dependency required by "@selected" [set]) > (dependency required by "@world" [argument]) Why do you have clang 9 installed? It is no longer in the repo, and apparently requires a python-exec compiled against python 3.6. Unless you have a particular need for it, the solution seems to be to remove clang 9. Cheers, Arve
Re: [gentoo-user] Thunderbird build failure ..
Cal, On Sunday, 2023-01-01 13:21:34 -0800, you wrote: > ... > You're right, it looks like the Thunderbird ebuild has a clang USE > turned on by default; On my rig "clang" and "llvm" used about 90 minutes each to build and af- terwards often caused rebuilds for other packages which, too, took their time. So eventually I added "-clang -llvm" to my global "USE" variable, thus not only saving plenty of compile time but aparently also avoiding plen- ty of problems ... Sincerely, Rainer PS: Happy New Year to all list members :-)
Re: [gentoo-user] Emerge trouble with firefox and thunderbird ...
On Sat, Mar 09, 2024 at 08:04:06AM +, Wols Lists wrote > For anyone else who hits this sort of problem, I did an > > USE=-clang emerge --update @world > > (firefox and thunderbird were the only programs I thought this would > touch), and it worked. I have "-clang" in USE in make.conf and no problems resulting from it. clang seems to be another "solution in search of a problem" along the lines of rust and cups and systemd and hatbuzz, etc, which keep trying to worm their way into everybody's linux system. -- Roses are red Roses are blue Depending on their velocity Relative to you
Re: [gentoo-user] Changing compilers
On 19/03/2012 10:18 PM, Andrew Lowe wrote: On 03/19/12 22:02, Michael Mol wrote: On Mon, Mar 19, 2012 at 8:32 AM, Mark Knechtmarkkne...@gmail.com wrote: On Sun, Mar 18, 2012 at 10:26 PM, Andrew Lowea...@wht.com.au wrote: Hi all, [snip] ... ... [snip] which currently ranks at 151 in the top 500 list :) It's amazing how fast this list changes, 6 months ago, this machine was at 107 and 6 months before that 87. Andrew Just in closing on this subject, thanks to those who responded, have a quick look at this page from the llvm/clang project: http://clang.llvm.org/diagnostics.html For my non FEA/CFD programming, I don't care if clang is 5 - 10% slower than gcc, the diagnostic output that clang produces looks to be spectacular in comparison to gcc and will be enough for me to dump gcc and shift to clang. Andrew
Re: [gentoo-user] Is there a reason why LLVM/Clang ebuilds don't support "mutislot"?
This also makes it difficult to use GHC's LLVM backend, since its compatible versions usually lag behind the current one.
Re: [gentoo-user] Emerge trouble with firefox and thunderbird ...
Il 10/03/24 23:44, Walter Dnes ha scritto: On Sun, Mar 10, 2024 at 08:43:46PM +0100, ralfconn wrote Given the warning message reported by Peter ("Enable USE=clang unless you have a very good reason not to.") That message comes from sys-libs/compiler-rt which is a dedicated runtime lib for clang. It makes sense to use clang if you're building the clang toolchain... just like using gcc if you're building the gcc toolchain. Correct, my mistake. My point was that it's worth filing a bug since it's possible that GCC support will be dropped upstream. But, the only 'news' I found on this regard [1] is quite old and it looks like GCC is still being supported. raf [1] https://blog.mozilla.org/nfroyd/2018/05/29/when-implementation-monoculture-right-thing/
[gentoo-user] llvm clang ABI confusion
Hello, My previous post suggest that every file under /etc/portage can be parsed for flags and directives. I removed all of the archives and that cleanup of /etc/portage/ is in progress. This question is unrelated, as best as I can tell. Ok, so I just went to install something and got this :: emerge -uDNvtp =dev-java/sbt-0.13.10_rc2 [ebuild U ~] dev-java/sbt-0.13.10_rc2::gentoo [0.13.8::gentoo] USE="-binary" 210,199 KiB [ebuild U ~] dev-lang/scala-2.11.7-r1:2.11/2.11.7::gentoo [2.11.4-r1:2.11/2.11.4::gentoo] USE="source -binary -doc -emacs" 0 KiB [ebuild U ~] dev-lang/scala-2.10.6:2.10/2.10.6::gentoo [2.10.4-r2:2.10/2.10.4::gentoo] USE="source -binary -doc -emacs" 82,629 KiB Total: 3 packages (3 upgrades), Size of downloads: 292,827 KiB WARNING: One or more updates/rebuilds have been skipped due to a dependency conflict: sys-devel/llvm:0 sys-devel/llvm-3.5.0:0/3.5::gentoo, ebuild scheduled for merge conflicts with sys-devel/llvm-3.5.0[clang(-),-debug,python,static- analyzer,abi_x86_32(-),abi_x86_64(-)] required by (sys-devel/clang-3.5.0-r100:0/3.5::gentoo, installed) I have both installed:: [I] sys-devel/llvm Installed versions: 3.5.0(06:15:37 PM 01/25/2016)(clang libffi ncurses python static-analyzer -debug -doc -gold -libedit -multitarget -ocaml -test -xml ABI_MIPS="-n32 -n64 -o32" ABI_PPC="-32 -64" ABI_S390="-32 -64" ABI_X86="32 64 -x32" KERNEL="-Darwin -FreeBSD" PYTHON_TARGETS="python2_7 -pypy" VIDEO_CARDS="radeon") and [I] sys-devel/clang Installed versions: 3.5.0-r100(12:07:02 PM 01/17/2016)(python static-analyzer -debug -multitarget ABI_MIPS="-n32 -n64 -o32" ABI_PPC="-32 -64" ABI_S390="-32 -64" ABI_X86="32 64 -x32") in my /etc/portage/package.use/000.abi I have this entry:: */* -abi_x86_32 both "egrep -r clang /etc/portage/" and "egrep -r clang /etc/portage/" return empty. Commenting out the above referenced entry in did not fix the conflict. I still want to ensure I have no 32 bit packages on this system. Is that my problem that clang/llvm require 32 bit sources? If so, how do I fix this and keep 32 bit packages off the system? Surely I am confused as to resolution options. Suggestions? James
[gentoo-user] Re: llvm clang ABI confusion
Alan McKinnon gmail.com> writes: > 1. Relax your stance and accept that some software out there that you > might want is 32 bits > 2. Refuse to have 32 bits, so give up on llvm and clang. Find something > else and move on. > You must pick one of those two. There is no magic hidden solution that > magically lets you have something and also not have it. OK, (1) can be moved to another system or in a container for those experiments. I'll choose door number (2). thx. James
[gentoo-user] Re: sys-devel/llvm and LLVM_TARGETS
On 2022-03-12, Nikos Chantziaras wrote: > On 12/03/2022 10:43, Dale wrote: >> https://bugs.gentoo.org/767700 >> >> Is that the one? It mentions the target but I don't quite understand >> the why. The biggest thing, will this break something if I let it do >> it? > > No. Unlike GCC, LLVM/Clang is always a cross-compiler. You can't use LLVM/Clang to compile for the host on which it's running? -- Grant
[gentoo-user] Re: sys-devel/llvm and LLVM_TARGETS
On 12/03/2022 18:03, Grant Edwards wrote: On 2022-03-12, Nikos Chantziaras wrote: On 12/03/2022 10:43, Dale wrote: https://bugs.gentoo.org/767700 Is that the one? It mentions the target but I don't quite understand the why. The biggest thing, will this break something if I let it do it? No. Unlike GCC, LLVM/Clang is always a cross-compiler. You can't use LLVM/Clang to compile for the host on which it's running? Why not?
Re: [gentoo-user] kdevelop broken (llvm slot issue)
On 19/08/18 11:21, Alexander Puchmayr wrote: > > This issue is covered by bug https://bugs.gentoo.org/651658, which is open > since March 2018 and no progress since also March 2018. > > It seems as if multiple slots of llvm cause the problems. mesa pulls in llvm: > 5, while other programs pull in llvm:6 (via clang:6) > > Does anyone have an idea how to get a working kdevelop again? I think you have to not have multiple LLVM/Clang installations, unfortunately. That's what the bug indicates. I am not having issues with KDevelop with Clang support and I have everything on the latest version: LLVM 6.0.1-r1 libffi ncurses Clang 6.0.1 +static-analyzer LLVM_TARGETS="AMDGPU BPF NVPTX X86" KDevelop 5.2.3 gdbui hex plasma qmake welcomepage kDevelop-php 5.2.3 kdevelop-python 5.2.3 Mesa 18.* classic dri3 egl gallium gbm gles2 llvm wayland Most of the above are defaults. I would just ensure everything is built against one version of Clang/LLVM and get rid of the other versions from the machine. If you really don't need the Clang features (if you are using KDevelop for non-C/C++), you can disable it at runtime: /usr/bin/env KDEV_DISABLE_PLUGINS=kdevclangsupport kdevelop %u You have to kill all KDevelop instances completely for this to work. I have this is in my menu because KDevelop gets dumb with QML JS vs JavaScript for me, making KDevelop nearly impossible to use with Node projects. -- Andrew signature.asc Description: OpenPGP digital signature
[gentoo-user] kdevelop broken (llvm slot issue)
On 19/08/18 11:21, Alexander Puchmayr wrote: > > This issue is covered by bug https://bugs.gentoo.org/651658, which is open > since March 2018 and no progress since also March 2018. > > It seems as if multiple slots of llvm cause the problems. mesa pulls in llvm: > 5, while other programs pull in llvm:6 (via clang:6) > > Does anyone have an idea how to get a working kdevelop again? I think you have to not have multiple LLVM/Clang installations, unfortunately. That's what the bug indicates. I am not having issues with KDevelop with Clang support and I have everything on the latest version: LLVM 6.0.1-r1 Clang 6.0.1 KDevelop 5.2.3 gdbui hex plasma qmake welcomepage kDevelop-php 5.2.3 kdevelop-python 5.2.3 I would just ensure everything is built against one version of Clang/LLVM and get rid of the other versions from the machine. If you really don't need the Clang features (if you are using KDevelop for non-C/C++), you can disable it at runtime: /usr/bin/env KDEV_DISABLE_PLUGINS=kdevclangsupport kdevelop %u I have this is in my menu because KDevelop gets dumb with QML JS vs JavaScript for me, making KDevelop nearly impossible to use with Node projec -- Andrew signature.asc Description: OpenPGP digital signature
Re: [gentoo-user] llvm clang ABI confusion
On 10/02/2016 23:52, James wrote: > Hello, > > My previous post suggest that every file under /etc/portage > > can be parsed for flags and directives. I removed all of the archives > > and that cleanup of /etc/portage/ is in progress. This question > > is unrelated, as best as I can tell. > > > Ok, so I just went to install something and got this :: > > emerge -uDNvtp =dev-java/sbt-0.13.10_rc2 > > [ebuild U ~] dev-java/sbt-0.13.10_rc2::gentoo [0.13.8::gentoo] > > USE="-binary" 210,199 KiB > > [ebuild U ~] dev-lang/scala-2.11.7-r1:2.11/2.11.7::gentoo > > [2.11.4-r1:2.11/2.11.4::gentoo] USE="source -binary -doc -emacs" 0 KiB > > [ebuild U ~] dev-lang/scala-2.10.6:2.10/2.10.6::gentoo > > [2.10.4-r2:2.10/2.10.4::gentoo] USE="source -binary -doc -emacs" 82,629 KiB > > Total: 3 packages (3 upgrades), Size of downloads: 292,827 KiB > > WARNING: One or more updates/rebuilds have been skipped due to a > dependency conflict: > > sys-devel/llvm:0 > > sys-devel/llvm-3.5.0:0/3.5::gentoo, ebuild scheduled for merge > conflicts with > > sys-devel/llvm-3.5.0[clang(-),-debug,python,static- > analyzer,abi_x86_32(-),abi_x86_64(-)] > > required by (sys-devel/clang-3.5.0-r100:0/3.5::gentoo, installed) > > I have both installed:: [I] sys-devel/llvm > > Installed versions: 3.5.0(06:15:37 PM 01/25/2016)(clang libffi ncurses > > python > static-analyzer -debug -doc -gold -libedit -multitarget -ocaml -test -xml > > ABI_MIPS="-n32 -n64 -o32" ABI_PPC="-32 -64" ABI_S390="-32 -64" > > ABI_X86="32 64 -x32" KERNEL="-Darwin -FreeBSD" PYTHON_TARGETS="python2_7 > -pypy" > VIDEO_CARDS="radeon") > > and [I] sys-devel/clang > Installed versions: 3.5.0-r100(12:07:02 PM 01/17/2016)(python > > static-analyzer -debug -multitarget ABI_MIPS="-n32 -n64 -o32" > > ABI_PPC="-32 -64" ABI_S390="-32 -64" ABI_X86="32 64 -x32") > > > in my /etc/portage/package.use/000.abi I have this > entry:: */* -abi_x86_32 > > > both "egrep -r clang /etc/portage/" and "egrep -r clang /etc/portage/" > return empty. Commenting out the above referenced entry in did not fix the > conflict. > > I still want to ensure I have no 32 bit packages on this system. > Is that my problem that clang/llvm require 32 bit sources? If so, how do I > fix this and keep 32 bit packages off the system? Surely I am confused as to > resolution options. The conflict is very simple. The ebuild says "must have 32 bits!" You said "no zero zip nada 32 bits!' The system says "screw this, you figure it out dude, I give up" Solutions: There are 2, and they are mutually exclusive 1. Relax your stance and accept that some software out there that you might want is 32 bits 2. Refuse to have 32 bits, so give up on llvm and clang. Find something else and move on. You must pick one of those two. There is no magic hidden solution that magically lets you have something and also not have it. -- Alan McKinnon alan.mckin...@gmail.com
Re: [gentoo-user] llvm compile error
Hello, On Fri, 24 Jan 2014 09:00:26 +0200 Alan McKinnon alan.mckin...@gmail.com wrote: llvm-3.4 is not hardmasked. It is marked ~x86 which is something very different. Correct, i mean Clang in version 3.4 is hardmasked. I want install clang alone llvm i not need. I think mesa use llvm too but i has not set the use flag. https://packages.gentoo.org/package/sys-devel/clang The mistake you made is using --autounmask-write That feature writes local package.unmask entries to satisfy dependencies. It's a very blunt tool, it blindly keywords whatever it thinks it needs to and when it goes wrong, it goes very wrong quickly. The tool write package.accept_keywords, a package.unmask i have not on system. Yes its a shit tool, normal i make echo package ~x86 package.a.. What should do? Should i compile without portage? Should i not use? When i want emerge clang is masked, ok why? FreeBSD use it and say is stable. FreeBSD maybe not the reference on earth but the BSD's make a good job. When i saw all versions of Clang is masked. You are running x86. If you want a package that is marked ~x86 then you need to take very careful note of everything that must be keyworded to build that package. If you want something basic like llvm that will cause many other packages to be upgrade with it, then you need to be especially careful. I use much x86 packages and has never problem. For example i use calibre without x86 flag i must use calibre in version 1.2. but the version 1.20 works stable and fine and i can use with my tablet, because i read much when im on the road. https://packages.gentoo.org/package/app-text/calibre So i think for some packages is accept between risk and compromiss. I find better x86 flag better as install software in /usr without portage. Thank you for help Nice Day Silvio
Re: [gentoo-user] llvm compile error
On 24/01/2014 20:44, Silvio Siefke wrote: Hello, On Fri, 24 Jan 2014 09:00:26 +0200 Alan McKinnon alan.mckin...@gmail.com wrote: llvm-3.4 is not hardmasked. It is marked ~x86 which is something very different. Correct, i mean Clang in version 3.4 is hardmasked. I want install clang alone llvm i not need. I think mesa use llvm too but i has not set the use flag. https://packages.gentoo.org/package/sys-devel/clang The mistake you made is using --autounmask-write That feature writes local package.unmask entries to satisfy dependencies. It's a very blunt tool, it blindly keywords whatever it thinks it needs to and when it goes wrong, it goes very wrong quickly. The tool write package.accept_keywords, a package.unmask i have not on system. Yes its a shit tool, normal i make echo package ~x86 package.a.. What should do? Should i compile without portage? Should i not use? When i want emerge clang is masked, ok why? FreeBSD use it and say is stable. FreeBSD maybe not the reference on earth but the BSD's make a good job. When i saw all versions of Clang is masked. No, just stop using automated tools to unmask/keyword everything based just on depends. Do it yourself, then you know what you unamsked/keyworded. Nobody suggested you stop using portage, I only said to stop hitting the system with a big hammer to get things to build. -- Alan McKinnon alan.mckin...@gmail.com
Re: [gentoo-user] mesa-12.0.1 fails to emerge
On 02/08/2017 12:47 AM, Mick wrote: > On Tuesday 07 Feb 2017 18:24:06 Corbin Bird wrote: >> On 02/07/2017 05:02 PM, Mick wrote: >>> How could I go beyond this point? >>> >>> = >>> /var/tmp/portage/media- >>> libs/mesa-12.0.1/work/mesa-12.0.1/src/gallium/state_track >>> ers/clover/llvm/invocation.cpp:212:75: error: no matching function for >>> call to ‘ >>> clang::CompilerInvocation::setLangDefaults(clang::LangOptions&, >>> clang::InputKind >>> , llvm::Triple, clang::LangStandard::Kind)’ >>> = >>> >>> Extract from compile log attached. >>> >>> I am running sys-devel/clang-runtime-3.9.1 like so: >>> Installed versions: 3.9.1(21:25:30 07/02/17)(openmp -libcxx >>> ABI_MIPS="- >>> >>> n32 -n64 -o32" ABI_PPC="-32 -64" ABI_S390="-32 -64" ABI_X86="64 -32 -x32") >> --- >> If you would please, post if the USE flag "opencl" is set in >> "make.conf", or in "package.use" for any packages. >> >> >> Corbin > Yes! It had sneaked in make.conf. I removed it and am re-emerging now. > > Thanks. :-) --- Your Welcome. Easy way to find the def of that USE flag ( it gave away the error ) : ~ # equery u media-libs/mesa + + opencl : Enable the Clover Gallium OpenCL state tracker. Easy way to check what else currently installed that may be using that USE flag : ~ # equery h opencl * Searching for USE flag opencl ... [IP-] [ ] media-gfx/imagemagick-6.9.7.4:0/6.9.7.4 [IP-] [ ] media-libs/mesa-17.0.0_rc2:0 [IP-] [ ] media-libs/x264-0.0.20160712:0/148 [IP-] [ ] sys-apps/hwloc-1.11.2:0/5 Corbin
[gentoo-user] mesa-12.0.1 fails to emerge
How could I go beyond this point? /var/tmp/portage/media- libs/mesa-12.0.1/work/mesa-12.0.1/src/gallium/state_track ers/clover/llvm/invocation.cpp:212:75: error: no matching function for call to ‘ clang::CompilerInvocation::setLangDefaults(clang::LangOptions&, clang::InputKind , llvm::Triple, clang::LangStandard::Kind)’ Extract from compile log attached. I am running sys-devel/clang-runtime-3.9.1 like so: Installed versions: 3.9.1(21:25:30 07/02/17)(openmp -libcxx ABI_MIPS="- n32 -n64 -o32" ABI_PPC="-32 -64" ABI_S390="-32 -64" ABI_X86="64 -32 -x32") -- Regards, Micklibtool: compile: x86_64-pc-linux-gnu-g++ -DPACKAGE_NAME=\"Mesa\" -DPACKAGE_TAR NAME=\"mesa\" -DPACKAGE_VERSION=\"12.0.1\" "-DPACKAGE_STRING=\"Mesa 12.0.1\"" "- DPACKAGE_BUGREPORT=\"https://bugs.freedesktop.org/enter_bug.cgi?product=Mesa\"; -DPACKAGE_URL=\"\" -DPACKAGE=\"mesa\" -DVERSION=\"12.0.1\" -DSTDC_HEADERS=1 -DHA VE_SYS_TYPES_H=1 -DHAVE_SYS_STAT_H=1 -DHAVE_STDLIB_H=1 -DHAVE_STRING_H=1 -DHAVE_ MEMORY_H=1 -DHAVE_STRINGS_H=1 -DHAVE_INTTYPES_H=1 -DHAVE_STDINT_H=1 -DHAVE_UNIST D_H=1 -DHAVE_DLFCN_H=1 -DLT_OBJDIR=\".libs/\" -DYYTEXT_POINTER=1 -DHAVE___BUILTI N_BSWAP32=1 -DHAVE___BUILTIN_BSWAP64=1 -DHAVE___BUILTIN_CLZ=1 -DHAVE___BUILTIN_C LZLL=1 -DHAVE___BUILTIN_CTZ=1 -DHAVE___BUILTIN_EXPECT=1 -DHAVE___BUILTIN_FFS=1 - DHAVE___BUILTIN_FFSLL=1 -DHAVE___BUILTIN_POPCOUNT=1 -DHAVE___BUILTIN_POPCOUNTLL1 -DHAVE___BUILTIN_UNREACHABLE=1 -DHAVE_FUNC_ATTRIBUTE_CONST=1 -DHAVE_FUNC_ATTRI BUTE_FLATTEN=1 -DHAVE_FUNC_ATTRIBUTE_FORMAT=1 -DHAVE_FUNC_ATTRIBUTE_MALLOC=1 -DH AVE_FUNC_ATTRIBUTE_PACKED=1 -DHAVE_FUNC_ATTRIBUTE_PURE=1 -DHAVE_FUNC_ATTRIBUTE_R ETURNS_NONNULL=1 -DHAVE_FUNC_ATTRIBUTE_UNUSED=1 -DHAVE_FUNC_ATTRIBUTE_WARN_UNUSE D_RESULT=1 -DHAVE_FUNC_ATTRIBUTE_WEAK=1 -DHAVE_DLADDR=1 -DHAVE_CLOCK_GETTIME=1 - DHAVE_PTHREAD=1 -DHAVE_SHA1_IN_LIBNETTLE=1 -I. -I/var/tmp/portage/media-libs/mes a-12.0.1/work/mesa-12.0.1/src/gallium/state_trackers/clover -I/var/tmp/portage/m edia-libs/mesa-12.0.1/work/mesa-12.0.1/include -I/var/tmp/portage/media-libs/mes a-12.0.1/work/mesa-12.0.1/src -I/var/tmp/portage/media-libs/mesa-12.0.1/work/mes a-12.0.1/src/gallium/include -I/var/tmp/portage/media-libs/mesa-12.0.1/work/mesa -12.0.1/src/gallium/drivers -I/var/tmp/portage/media-libs/mesa-12.0.1/work/mesa- 12.0.1/src/gallium/auxiliary -I/var/tmp/portage/media-libs/mesa-12.0.1/work/mesa -12.0.1/src/gallium/winsys -I../../../../src -I/var/tmp/portage/media-libs/mesa- 12.0.1/work/mesa-12.0.1/src/gallium/state_trackers/clover -std=c++11 -fvisibilit y=hidden -I/usr/include -std=c++11 -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACR OS -D__STDC_LIMIT_MACROS -D__STDC_LIMIT_MACROS -D__STDC_CONSTANT_MACROS -D_GNU_S OURCE -DUSE_SSE41 -DNDEBUG -DTEXTURE_FLOAT_ENABLED -DUSE_X86_64_ASM -DHAVE_XLOCA LE_H -DHAVE_SYS_SYSCTL_H -DHAVE_STRTOF -DHAVE_MKOSTEMP -DHAVE_DLOPEN -DHAVE_POSI X_MEMALIGN -DHAVE_LIBDRM -DHAVE_SHA1 -DGLX_USE_DRM -DHAVE_LIBUDEV -DGLX_INDIRECT _RENDERING -DGLX_DIRECT_RENDERING -DGLX_USE_TLS -DHAVE_ALIAS -DHAVE_DRI3 -DHAVE_ MINCORE -DHAVE_ST_VDPAU -DHAVE_LLVM=0x0309 -DMESA_LLVM_VERSION_PATCH=1 -DLIBCLC_ INCLUDEDIR=\"/usr/include/\" -DLIBCLC_LIBEXECDIR=\"/usr/lib/clc/\" -DCLANG_RESOU RCE_DIR=\"/usr/lib/clang/3.9.1\" -march=native -O2 -pipe -Wall -fno-strict-alias ing -fno-builtin-memcmp -c /var/tmp/portage/media-libs/mesa-12.0.1/work/mesa-12. 0.1/src/gallium/state_trackers/clover/llvm/invocation.cpp -fPIC -DPIC -o llvm/. libs/libclllvm_la-invocation.o /var/tmp/portage/media-libs/mesa-12.0.1/work/mesa-12.0.1/src/gallium/state_track ers/clover/llvm/invocation.cpp: In function ‘llvm::Module* {anonymous}::compile_ llvm(llvm::LLVMContext&, const string&, const header_map&, const string&, const string&, const string&, const string&, unsigned int (&)[7], unsigned int&, std:: string&)’: /var/tmp/portage/media-libs/mesa-12.0.1/work/mesa-12.0.1/src/gallium/state_track ers/clover/llvm/invocation.cpp:212:75: error: no matching function for call to ‘ clang::CompilerInvocation::setLangDefaults(clang::LangOptions&, clang::InputKind , llvm::Triple, clang::LangStandard::Kind)’ clang::LangStandard::lang_opencl11); ^ /var/tmp/portage/media-libs/mesa-12.0.1/work/mesa-12.0.1/src/gallium/state_track ers/clover/llvm/invocation.cpp:212:75: note: candidate is: In file included from /usr/include/clang/Frontend/CompilerInstance.h:17:0, from /var/tmp/portage/media-libs/mesa-12.0.1/work/mesa-12.0.1/s rc/gallium/state_trackers/clover/llvm/invocation.cpp:25: /usr/include/clang/Frontend/CompilerInvocation.h:160:15: note: static v[41/1956] ::CompilerIn
[gentoo-user] Firefox 64 binaries - LLVM6+PGO+LTO now used, faster than GCC6.4+PGO
https://www.phoronix.com/scan.php?page=news_item=Firefox-Clang-LTO-All-Platforms "Firefox nightly builds are now built with the LLVM Clang compiler on all major platforms and the Linux build in particular is also now utilizing PGO optimizations too. Faster Firefox is coming thanks to this compiler work. All of Mozilla's tier-one platforms are now building the newest Firefox browser code under the Clang compiler and having LTO (Link Time Optimizations <https://www.phoronix.com/vr.php?view=17785>) enabled. That includes Linux, Mac, Android, Windows across ARM / AArch64 / x86 relying upon this open-source compiler. For now only the Linux builds also have PGO (Profile Guided Optimizations <https://www.phoronix.com/vr.php?view=26589>) enabled. With the compiler re-tooling on Linux they are seeing about 5% better performance and upwards of 18% in select tests compared to their previous GCC (6.4-based) compiler with PGO. The current Clang stack they are using is based on LLVM6 while when they transition to the upcoming LLVM/Clang 7.0 they expect 2~5% better performance on top of that. Mozilla developers did try GCC LTO as well as switching to the newest GCC8 release, but ran into problems. These new release binaries should go into effect for the Firefox 64 release."
Re: [gentoo-user] Emerge trouble with firefox and thunderbird ...
On Sat, Mar 09, 2024 at 09:16:37PM +, Peter Humphrey wrote > On Saturday, 9 March 2024 19:37:40 GMT Walter Dnes wrote: > > > On Sat, Mar 09, 2024 at 02:45:02PM +, Peter Humphr > > The real question is what else, besides clang and its libraries, > are you building that requires clang? > > Firefox. Upstream in this same thread... On Sat, Mar 09, 2024 at 08:04:06AM +, Wols Lists wrote > On 03/03/2024 23:13, Carsten Hauck wrote: > > > > Some time ago on one of my machines Thunderbird and Firefox stopped to > > compile with USE="clang". As they can be build with gcc I never digged > > too deep into that problem but maybe it's worth a shot. > > For anyone else who hits this sort of problem, I did an > > USE=-clang emerge --update @world > > (firefox and thunderbird were the only programs I thought this would > touch), and it worked. > > There were a couple of other programs that I guess got pulled in by the > changed use, but they've upgraded which is the main thing. > > Thank you very much So there are at least 2 people who've found out that Firefox can and *MUST* be built with USE="-clang". -- Roses are red Roses are blue Depending on their velocity Relative to you
[gentoo-user] Libclc and OpenCL
Hello list, Bug 596022 reports failure to build dev-libs/libclc with the latest version of sys-dev/clang. I've been trying to trace a problem running MilkyWay@home, specifically with its use of the GPU, so I'm keen to make sure that I have a properly working OpenCL system. Is my existing libclc, compiled against the previous version of clang, still good? I'm hoping this is just a build-phase problem, with nothing changed in the object code. I can't go back in clang versions without also reverting sys-devel/llvm, and maybe other things too. The bug has been open for three weeks, asking for a new libclc package to be released. It'd be good if that were to happen soon. -- Rgds Peter PS. Has anyone seen anything of Alan McK recently? He seems to have vanished.
Re: [gentoo-user] Re: chromium-65.0.3325.146 compilation warnings
On Monday, 12 March 2018 17:11:16 GMT Nikos Chantziaras wrote: > No. What these warnings mean is that the chromium build system is > passing these options to the compiler: > >-Wno-enum-compare-switch -Wno-tautological-unsigned-zero-compare >-Wno-null-pointer-arithmetic -Wno-tautological-constant-compare > > but the compiler version installed on your system doesn't support them. > > Your own CFLAGS got nothing to do with this. Furthermore, these warnings > are of no interest to you whatsoever, unless you're a chromium developer. I don't believe I possess the abilities to be developer. :-) > GCC 7.3 and Clang 6.0 do support these options, but if you're using a > version that doesn't, it's of no consequence. Warning options do not > affect code generation. OK, I think it's clang what done it. I'm on gcc-7.3.0, but clang-5.0.1. -- Regards, Mick signature.asc Description: This is a digitally signed message part.
Re: [gentoo-user] Thunderbird build failure ..
Thanks bud Dave On Sun, Jan 1, 2023, 4:22 PM cal wrote: > On 1/1/23 13:05, Wol wrote: > > On 01/01/2023 20:08, cal wrote: > >> FWIW, Thunderbird builds fine with GCC on my machine -- I'm unsure of > >> your reasons for setting your Portage compiler to clang, but you may > >> wish to use a package.env override to build Thunderbird with GCC as a > >> workaround until the problem can be fixed upstream. > > > > I don't know anything about clang ... it must be the default ... > > > > I thought part of Firefox/Thunderbird was written in Rust, so I assumed > > it was built with llvm as a matter of course. > > > > I'll just wait for it to sort itself out. > > > > Cheers, > > Wol > You're right, it looks like the Thunderbird ebuild has a clang USE > turned on by default; I didn't realize that in my earlier reply and > assumed you had overridden this yourself. Regardless, this version > built correctly on my machine, so it's worthy of investigation with > upstream which combinations of parameters may be triggering the crash. > > cal > >
Re: [gentoo-user] Thunderbird build failure ..
On 1/1/23 11:14, Wols Lists wrote: > On 01/01/2023 18:33, cal wrote: >> On 1/1/23 03:07, Wols Lists wrote: >>> I got the following build failure in my weekly emerge yesterday ... >>> >>> * Messages for package mail-client/thunderbird-102.6.0: >>> >>> * ERROR: mail-client/thunderbird-102.6.0::gentoo failed (compile >>> phase): >>> * (no error message) >>> * >>> * Call stack: >>> * ebuild.sh, line 136: Called src_compile >>> * environment, line 4782: Called die >>> * The specific snippet of code: >>> * ${virtx_cmd} ./mach build --verbose || die >>> * >>> * If you need support, post the output of `emerge --info >>> '=mail-client/thunderbird-102.6.0::gentoo'`, >>> * the complete build log and the output of `emerge -pqv >>> '=mail-client/thunderbird-102.6.0::gentoo'`. >>> * The complete build log is located at >>> '/var/tmp/portage/mail-client/thunderbird-102.6.0/temp/build.log'. >>> * The ebuild environment file is located at >>> '/var/tmp/portage/mail-client/thunderbird-102.6.0/temp/environment'. >>> * Working directory: >>> '/var/tmp/portage/mail-client/thunderbird-102.6.0/work/thunderbird-102.6.0' >>> * S: >>> '/var/tmp/portage/mail-client/thunderbird-102.6.0/work/thunderbird-102.6.0' > >> Can you post the outputs referenced above/the full build log and your >> portage configuration? > > Files attached - not sure it's everything you want, but I'm sure you'll > let me know if I've messed up ... :-) If you scroll up a bit in build.log, you can see that clang has crashed: [...] 836:17.68 clang-15: error: clang frontend command failed with exit code 139 (use -v to see invocation) 836:17.68 clang version 15.0.6 836:17.68 Target: x86_64-pc-linux-gnu 836:17.69 Thread model: posix 836:17.69 InstalledDir: /usr/lib/llvm/15/bin 836:17.69 Configuration file: /etc/clang/clang++.cfg 836:18.07 clang-15: note: diagnostic msg: 836:18.07 836:18.07 PLEASE ATTACH THE FOLLOWING FILES TO THE BUG REPORT: 836:18.07 Preprocessed source(s) and associated run script(s) are located at: 836:18.07 clang-15: note: diagnostic msg: /var/tmp/portage/mail-client/thunderbird-102.6.0/temp/Unified_cpp_dom_media_webaudio1-f85814.cpp 836:18.07 clang-15: note: diagnostic msg: /var/tmp/portage/mail-client/thunderbird-102.6.0/temp/Unified_cpp_dom_media_webaudio1-f85814.sh 836:18.07 clang-15: note: diagnostic msg: 836:18.07 If you search for "clang frontend command failed with exit code 139" it turns up a few GitHub issues which may or may not be related; you may wish to follow up with upstream on that. FWIW, Thunderbird builds fine with GCC on my machine -- I'm unsure of your reasons for setting your Portage compiler to clang, but you may wish to use a package.env override to build Thunderbird with GCC as a workaround until the problem can be fixed upstream. >>> >>> I'm wondering whether this will simply clear itself next week, seeing as >>> last week I got a very similar failure for both thunderbird and firefox. >>> >>> Could it simply be a bit of the fallout from app-alternatives? Of >>> course, it's blocking my depclean ... >> Having followed this list for quite a while, my first guess whenever >> someone fails to compile a web browser or similar heavy piece of >> software (I'm counting Thunderbird here) is that they ran out of memory. >> I would double check your MAKEOPTS and RAM size and try building with a >> smaller -j. But as noted above, perhaps if you attach the full build >> log for Thunderbird, a more obvious cause will appear in the output. >>> > > thewolery /usr/local # df > Filesystem 1K-blocks Used Available Use% > Mounted on > none 16404256 3328 16400928 1% /run > udev 10240 0 10240 0% /dev > tmpfs 16404256 0 16404256 0% > /dev/shm > /dev/dm-1 131001348 101620580 22653500 82% / > tmpfs 16404260 4 16404256 1% /tmp > /dev/mapper/vg--home-lv--data 2064042928 1472050236 487118712 76% /home > tmpfs 3280848 56 3280792 1% > /run/user/1000 > /dev/mapper/vg--home-lv--Videos 772966856 276285468 457343404 38% > /home/Videos > /dev/mapper/vg--home-lv--ISO 153707984 88546712 57280568 61% > /home/ISO > thewolery /usr/local # > > There's 28GB free disk space, surely that's enough? And iirc this > machine has 32GB ram. Or do I need to do a bit of a clean-out - I appear > to have used some 100GB for root which seems rather a lot ... > > Cheers, > Wol
Re: [gentoo-user] Thunderbird build failure ..
On 1/1/23 13:05, Wol wrote: > On 01/01/2023 20:08, cal wrote: >> FWIW, Thunderbird builds fine with GCC on my machine -- I'm unsure of >> your reasons for setting your Portage compiler to clang, but you may >> wish to use a package.env override to build Thunderbird with GCC as a >> workaround until the problem can be fixed upstream. > > I don't know anything about clang ... it must be the default ... > > I thought part of Firefox/Thunderbird was written in Rust, so I assumed > it was built with llvm as a matter of course. > > I'll just wait for it to sort itself out. > > Cheers, > Wol You're right, it looks like the Thunderbird ebuild has a clang USE turned on by default; I didn't realize that in my earlier reply and assumed you had overridden this yourself. Regardless, this version built correctly on my machine, so it's worthy of investigation with upstream which combinations of parameters may be triggering the crash. cal
Re: [gentoo-user] Re: Emerge trouble with firefox and thunderbird ...
On 10/03/24 at 01:50, mp666 wrote: On Sat, 9 Mar 2024 08:04:06 +, Wols Lists wrote: For anyone else who hits this sort of problem, I did an USE=-clang emerge --update @world (firefox and thunderbird were the only programs I thought this would touch), and it worked. There were a couple of other programs that I guess got pulled in by the changed use, but they've upgraded which is the main thing. Thank you very much Cheers, Wol This is a known problem. It generally shows itself with older architectures like AMD Phenom II, Bulldozer, Intel Core 2 etc. "-march=native" in the make.conf file was the culprit, IIRC, and replacing it with core2 or amd's equivalent option solved it. I, instead, put "-clang" in my package.use file, for firefox. No problem with Firefox builds since then. The CPU of the machine in question is in deed an old AMD. It's good to know the reason for that build-failures, thanks a lot. I certainly will stick to "-clang" in my package.use. Greetings, Carsten
Re: [gentoo-user] llvm compile error
On Fri, 24 Jan 2014 19:44:38 +0100 Silvio Siefke siefke_lis...@web.de wrote: Hello, On Fri, 24 Jan 2014 09:00:26 +0200 Alan McKinnon alan.mckin...@gmail.com wrote: llvm-3.4 is not hardmasked. It is marked ~x86 which is something very different. Correct, i mean Clang in version 3.4 is hardmasked. I want install clang alone llvm i not need. I think mesa use llvm too but i has not set the use flag. https://packages.gentoo.org/package/sys-devel/clang I don't know why I'm replying, as I'm no expert. So, I could be way off, but: From a google search: Clang /ˈklæŋ/ is a compiler front end for the C, C++, Objective-C and Objective-C++ programming languages. It uses LLVM as its back end ... The mistake you made is using --autounmask-write That feature writes local package.unmask entries to satisfy dependencies. It's a very blunt tool, it blindly keywords whatever it thinks it needs to and when it goes wrong, it goes very wrong quickly. The tool write package.accept_keywords, a package.unmask i have not on system. Yes its a shit tool, normal i make echo package ~x86 package.a.. What should do? That's what you should do. The output will tell you what files need updating. In some cases, I just let it write, like for a big perl package with lots of modules from the overlay. Otherwise it's faster to just copy and paste the output. Should i compile without portage? No. Should i not use? When i want emerge clang is masked, ok why? https://bugs.gentoo.org/show_bug.cgi?id=408963 maybe? FreeBSD use it and say is stable. FreeBSD maybe not the reference on earth but the BSD's make a good job. When i saw all versions of Clang is masked. Isn't FBSD more similar than, say, Linux, to what OSX is, sort of? [...] So i think for some packages is accept between risk and compromiss. I find better x86 flag better as install software in /usr without portage. Yep. Sometimes you have to unmask stuff, but portage can handle it then. Anyway, I hope any of this helps. Good luck! Cheers, -- - - - Michael Higgins - michael_higg...@iinet.com 503-473-5882
[gentoo-user] Re: new computer : any advice ?
Gevisz gmail.com> writes: > on-board video card. Just to avoid extra heating and aircraft noise > produced by R4770. Fanless video cards are wonderful. I have had many over the years but this one is still my (silent) favorite:: 01:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Cape Verde PRO [Radeon HD 7750 / R7 250E] I did notice in a gentoo blog that openmp is a testing option for Clang-3.7 now? [1] Try not to loose faith, we all have had bad experiences, but clustering, distributed and systems aggregation codes are rapidly coalescing into something wonderful, so . keep the faith.. bro. ;-) wwr, James [1] http://blog.cafarelli.fr/2015/09/testing-clang-3-7-0-openmp-support/
[gentoo-user] Re: llvm clang ABI confusion
James tampabay.rr.com> writes: > > 1. Relax your stance and accept that some software out there that you > > might want is 32 bits > > 2. Refuse to have 32 bits, so give up on llvm and clang. Find something > > else and move on. > > You must pick one of those two. There is no magic hidden solution that > > magically lets you have something and also not have it. > OK, (1) can be moved to another system or in a container for those > experiments. I'll choose door number (2). oops transposed the response ==>> choose option 1. Option 2 is for minimalist cluster systems testing. James
Re: [gentoo-user] kdevelop broken (llvm slot issue)
Am Sonntag, 19. August 2018, 17:57:55 CEST schrieb Andrew Udvare: > > I am not having issues with KDevelop with Clang support and I have > everything on the latest version: > > LLVM 6.0.1-r1 libffi ncurses > Clang 6.0.1 +static-analyzer LLVM_TARGETS="AMDGPU BPF NVPTX X86" > KDevelop 5.2.3 gdbui hex plasma qmake welcomepage > kDevelop-php 5.2.3 > kdevelop-python 5.2.3 > Mesa 18.* classic dri3 egl gallium gbm gles2 llvm wayland > Thanks for the quick response. Upgrading to mesa 18 solves the problem. Regards Alex
[gentoo-user] kdevelop broken (llvm slot issue)
Hi there, After recent upgrades of mesa, llvm, clang etc kdevelop does not work anymore. It crashes immediately after start with errors : CommandLine Error: Option 'help-list' registered more than once! LLVM ERROR: inconsistency in registered CommandLine options This issue is covered by bug https://bugs.gentoo.org/651658, which is open since March 2018 and no progress since also March 2018. It seems as if multiple slots of llvm cause the problems. mesa pulls in llvm: 5, while other programs pull in llvm:6 (via clang:6) Does anyone have an idea how to get a working kdevelop again? Regards Alex
[gentoo-user] Re: kdevelop broken (llvm slot issue)
On 19/08/18 18:21, Alexander Puchmayr wrote: After recent upgrades of mesa, llvm, clang etc kdevelop does not work anymore. It crashes immediately after start with errors [...] It seems as if multiple slots of llvm cause the problems. mesa pulls in llvm: 5, while other programs pull in llvm:6 (via clang:6) Does anyone have an idea how to get a working kdevelop again? Have you tried disabling the llvm USE flag of mesa? AFAIK it's only used for the software rendering backend, which you probably don't need. Try disabling it, and then do: emerge -auDN --changed-deps --with-bdeps=y @world followed by: emerge -a --depclean which should unmerge llvm:5.
Re: [gentoo-user] Thunderbird build failure ..
On 01/01/2023 20:08, cal wrote: FWIW, Thunderbird builds fine with GCC on my machine -- I'm unsure of your reasons for setting your Portage compiler to clang, but you may wish to use a package.env override to build Thunderbird with GCC as a workaround until the problem can be fixed upstream. I don't know anything about clang ... it must be the default ... I thought part of Firefox/Thunderbird was written in Rust, so I assumed it was built with llvm as a matter of course. I'll just wait for it to sort itself out. Cheers, Wol
[gentoo-user] Re: Emerge trouble with firefox and thunderbird ...
On Sat, 9 Mar 2024 08:04:06 +, Wols Lists wrote: > For anyone else who hits this sort of problem, I did an > > USE=-clang emerge --update @world > > (firefox and thunderbird were the only programs I thought this would > touch), and it worked. > > There were a couple of other programs that I guess got pulled in by the > changed use, but they've upgraded which is the main thing. > > Thank you very much > > Cheers, > Wol This is a known problem. It generally shows itself with older architectures like AMD Phenom II, Bulldozer, Intel Core 2 etc. "-march=native" in the make.conf file was the culprit, IIRC, and replacing it with core2 or amd's equivalent option solved it. I, instead, put "-clang" in my package.use file, for firefox. No problem with Firefox builds since then.
Re: [gentoo-user] llvm / clang compile error
On 08/16/2016 04:43 PM, Silvio Siefke wrote: > Hello, > > I'm trying to install llvm for days. But all versions fail with the > error message "internal compiler error: Killed (program cc1plus)". > > The last lines in the build log. > > [1820/2304] /usr/lib64/ccache/bin/x86_64-pc-linux-gnu-g++ > -DCLANG_ENABLE_ARCMT -DCLANG_ENABLE_OBJC_REWRITER > -DCLANG_ENABLE_STATIC_ANALYZER -D_GNU_SOURCE -D__STDC_CONSTANT_MACROS > -D__STDC_FORMAT_MACROS -D__STDC_LIMIT_MACROS -DNDEBUG -O2 -pipe -fPIC > -fvisibility-inlines-hidden -Wall -W -Wno-unused-parameter -Wwrite-strings > -Wcast-qual -Wno-missing-field-initializers -pedantic -Wno-long-long > -Wno-maybe-uninitialized -Wnon-virtual-dtor -Wno-comment -Werror=date-time > -std=c++11 -ffunction-sections -fdata-sections -fno-common > -Woverloaded-virtual -fno-strict-aliasing -fPIC -Itools/clang/lib/Sema > -I/var/tmp/portage/sys-devel/llvm-3.7.1-r3/work/llvm-3.7.1.src/tools/clang/lib/Sema > > -I/var/tmp/portage/sys-devel/llvm-3.7.1-r3/work/llvm-3.7.1.src/tools/clang/include > -Itools/clang/include -Iinclude > -I/var/tmp/portage/sys-devel/llvm-3.7.1-r3/work/llvm-3.7.1.src/include -MMD > -MT tools/clang/lib/Sema/CMakeFiles/clangSema.dir/SemaExprObjC.cpp.o -MF > tools/clang/lib/Sema/CMakeFiles/clangSema.dir/SemaExprObjC.cpp.o.d -o > tools/clang/lib/Sema/CMakeFiles/clangSema.dir/SemaExprObjC.cpp.o -c > /var/tmp/portage/sys-devel/llvm-3.7.1-r3/work/llvm-3.7.1.src/tools/clang/lib/Sema/SemaExprObjC.cpp > FAILED: /usr/lib64/ccache/bin/x86_64-pc-linux-gnu-g++ -DCLANG_ENABLE_ARCMT > -DCLANG_ENABLE_OBJC_REWRITER -DCLANG_ENABLE_STATIC_ANALYZER -D_GNU_SOURCE > -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -D__STDC_LIMIT_MACROS > -DNDEBUG -O2 -pipe -fPIC -fvisibility-inlines-hidden -Wall -W > -Wno-unused-parameter -Wwrite-strings -Wcast-qual > -Wno-missing-field-initializers -pedantic -Wno-long-long > -Wno-maybe-uninitialized -Wnon-virtual-dtor -Wno-comment -Werror=date-time > -std=c++11 -ffunction-sections -fdata-sections -fno-common > -Woverloaded-virtual -fno-strict-aliasing -fPIC > -Itools/clang/lib/ASTMatchers/Dynamic > -I/var/tmp/portage/sys-devel/llvm-3.7.1-r3/work/llvm-3.7.1.src/tools/clang/lib/ASTMatchers/Dynamic > > -I/var/tmp/portage/sys-devel/llvm-3.7.1-r3/work/llvm-3.7.1.src/tools/clang/include > -Itools/clang/include -Iinclude > -I/var/tmp/portage/sys-devel/llvm-3.7.1-r3/work/llvm-3.7.1.src/include -MMD > -MT > tools/clang/lib/ASTMatchers/Dynamic/CMakeFiles/clangDynamicASTMatchers.dir/Registry.cpp.o > -MF > tools/clang/lib/ASTMatchers/Dynamic/CMakeFiles/clangDynamicASTMatchers.dir/Registry.cpp.o.d > -o > tools/clang/lib/ASTMatchers/Dynamic/CMakeFiles/clangDynamicASTMatchers.dir/Registry.cpp.o > -c > /var/tmp/portage/sys-devel/llvm-3.7.1-r3/work/llvm-3.7.1.src/tools/clang/lib/ASTMatchers/Dynamic/Registry.cpp > {standard input}: Assembler messages: > {standard input}:265459: Warning: end of file in string; '"' inserted > x86_64-pc-linux-gnu-g++: internal compiler error: Killed (program cc1plus) > Please submit a full bug report, > > This happen with version 3.5, 3.8, 3.7. The crazy thing without set flag > clang, the installation work without problems. In Bug Report i found > equal bug, but there seem march item the problem. This i have not set. > > Have someone a idea why llvm hate clang? > > Thank's > > Silvio > > > # emerge --info '=sys-devel/llvm-3.7.1-r3::gentoo' > Portage 2.2.28 (python 2.7.10-final-0, default/linux/amd64/13.0/no-multilib, > gcc-5.4.0, glibc-2.22-r4, 3.14.32--std-ipv6-64-hz1000 x86_64) > = > System Settings > = > System uname: > Linux-3.14.32--std-ipv6-64-hz1000-x86_64-Intel-R-_Atom-TM-_CPU_N2800_@_1.86GHz-with-gentoo-2.2 > KiB Mem: 2024472 total, 1766584 free > KiB Swap:2046972 total, 1229132 free > Timestamp of repository gentoo: Tue, 16 Aug 2016 11:45:01 + > sh bash 4.3_p42-r1 > ld GNU ld (Gentoo 2.25.1 p1.1) 2.25.1 > ccache version 3.2.4 [enabled] > app-shells/bash: 4.3_p42-r1::gentoo > dev-java/java-config: 2.2.0-r3::gentoo > dev-lang/perl:5.20.2::gentoo > dev-lang/python: 2.7.10-r1::gentoo, 3.4.3-r1::gentoo > dev-util/ccache: 3.2.4::gentoo > dev-util/cmake: 3.3.1-r1::gentoo > dev-util/pkgconfig: 0.28-r2::gentoo > sys-apps/baselayout: 2.2::gentoo > sys-apps/openrc: 0.21.3::gentoo > sys-apps/sandbox: 2.10-r1::gentoo > sys-devel/autoconf: 2.13::gentoo, 2.69::gentoo > sys-devel/automake: 1.11.6-r1::gentoo, 1.14.1::gentoo, 1.15::gentoo &
[gentoo-user] llvm / clang compile error
Hello, I'm trying to install llvm for days. But all versions fail with the error message "internal compiler error: Killed (program cc1plus)". The last lines in the build log. [1820/2304] /usr/lib64/ccache/bin/x86_64-pc-linux-gnu-g++ -DCLANG_ENABLE_ARCMT -DCLANG_ENABLE_OBJC_REWRITER -DCLANG_ENABLE_STATIC_ANALYZER -D_GNU_SOURCE -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -D__STDC_LIMIT_MACROS -DNDEBUG -O2 -pipe -fPIC -fvisibility-inlines-hidden -Wall -W -Wno-unused-parameter -Wwrite-strings -Wcast-qual -Wno-missing-field-initializers -pedantic -Wno-long-long -Wno-maybe-uninitialized -Wnon-virtual-dtor -Wno-comment -Werror=date-time -std=c++11 -ffunction-sections -fdata-sections -fno-common -Woverloaded-virtual -fno-strict-aliasing -fPIC -Itools/clang/lib/Sema -I/var/tmp/portage/sys-devel/llvm-3.7.1-r3/work/llvm-3.7.1.src/tools/clang/lib/Sema -I/var/tmp/portage/sys-devel/llvm-3.7.1-r3/work/llvm-3.7.1.src/tools/clang/include -Itools/clang/include -Iinclude -I/var/tmp/portage/sys-devel/llvm-3.7.1-r3/work/llvm-3.7.1.src/include -MMD -MT tools/clang/lib/Sema/CMakeFiles/clangSema.dir/SemaExprObjC.cpp.o -MF tools/clang/lib/Sema/CMakeFiles/clangSema.dir/SemaExprObjC.cpp.o.d -o tools/clang/lib/Sema/CMakeFiles/clangSema.dir/SemaExprObjC.cpp.o -c /var/tmp/portage/sys-devel/llvm-3.7.1-r3/work/llvm-3.7.1.src/tools/clang/lib/Sema/SemaExprObjC.cpp FAILED: /usr/lib64/ccache/bin/x86_64-pc-linux-gnu-g++ -DCLANG_ENABLE_ARCMT -DCLANG_ENABLE_OBJC_REWRITER -DCLANG_ENABLE_STATIC_ANALYZER -D_GNU_SOURCE -D__STDC_CONSTANT_MACROS -D__STDC_FORMAT_MACROS -D__STDC_LIMIT_MACROS -DNDEBUG -O2 -pipe -fPIC -fvisibility-inlines-hidden -Wall -W -Wno-unused-parameter -Wwrite-strings -Wcast-qual -Wno-missing-field-initializers -pedantic -Wno-long-long -Wno-maybe-uninitialized -Wnon-virtual-dtor -Wno-comment -Werror=date-time -std=c++11 -ffunction-sections -fdata-sections -fno-common -Woverloaded-virtual -fno-strict-aliasing -fPIC -Itools/clang/lib/ASTMatchers/Dynamic -I/var/tmp/portage/sys-devel/llvm-3.7.1-r3/work/llvm-3.7.1.src/tools/clang/lib/ASTMatchers/Dynamic -I/var/tmp/portage/sys-devel/llvm-3.7.1-r3/work/llvm-3.7.1.src/tools/clang/include -Itools/clang/include -Iinclude -I/var/tmp/portage/sys-devel/llvm-3.7.1-r3/work/llvm-3.7.1.src/include -MMD -MT tools/clang/lib/ASTMatchers/Dynamic/CMakeFiles/clangDynamicASTMatchers.dir/Registry.cpp.o -MF tools/clang/lib/ASTMatchers/Dynamic/CMakeFiles/clangDynamicASTMatchers.dir/Registry.cpp.o.d -o tools/clang/lib/ASTMatchers/Dynamic/CMakeFiles/clangDynamicASTMatchers.dir/Registry.cpp.o -c /var/tmp/portage/sys-devel/llvm-3.7.1-r3/work/llvm-3.7.1.src/tools/clang/lib/ASTMatchers/Dynamic/Registry.cpp {standard input}: Assembler messages: {standard input}:265459: Warning: end of file in string; '"' inserted x86_64-pc-linux-gnu-g++: internal compiler error: Killed (program cc1plus) Please submit a full bug report, This happen with version 3.5, 3.8, 3.7. The crazy thing without set flag clang, the installation work without problems. In Bug Report i found equal bug, but there seem march item the problem. This i have not set. Have someone a idea why llvm hate clang? Thank's Silvio # emerge --info '=sys-devel/llvm-3.7.1-r3::gentoo' Portage 2.2.28 (python 2.7.10-final-0, default/linux/amd64/13.0/no-multilib, gcc-5.4.0, glibc-2.22-r4, 3.14.32--std-ipv6-64-hz1000 x86_64) = System Settings = System uname: Linux-3.14.32--std-ipv6-64-hz1000-x86_64-Intel-R-_Atom-TM-_CPU_N2800_@_1.86GHz-with-gentoo-2.2 KiB Mem: 2024472 total, 1766584 free KiB Swap:2046972 total, 1229132 free Timestamp of repository gentoo: Tue, 16 Aug 2016 11:45:01 + sh bash 4.3_p42-r1 ld GNU ld (Gentoo 2.25.1 p1.1) 2.25.1 ccache version 3.2.4 [enabled] app-shells/bash: 4.3_p42-r1::gentoo dev-java/java-config: 2.2.0-r3::gentoo dev-lang/perl:5.20.2::gentoo dev-lang/python: 2.7.10-r1::gentoo, 3.4.3-r1::gentoo dev-util/ccache: 3.2.4::gentoo dev-util/cmake: 3.3.1-r1::gentoo dev-util/pkgconfig: 0.28-r2::gentoo sys-apps/baselayout: 2.2::gentoo sys-apps/openrc: 0.21.3::gentoo sys-apps/sandbox: 2.10-r1::gentoo sys-devel/autoconf: 2.13::gentoo, 2.69::gentoo sys-devel/automake: 1.11.6-r1::gentoo, 1.14.1::gentoo, 1.15::gentoo sys-devel/binutils: 2.25.1-r1::gentoo sys-devel/gcc:5.4.0::gentoo sys-devel/gcc-config: 1.7.3::gentoo sys-devel/libtool:2.4.6::gentoo sys-devel/make: 4.1-r1::gentoo sys-kernel/linux-headers: 4.3::gentoo (virtual/os-headers) sys-libs/glibc: 2.22-r4::gentoo Repositories: gentoo location: /usr/portage sync-type: rsync sync-uri: rsync://rsync2.fr.gentoo.org/gentoo-portage/ priority: -1000 pentoo location: /var/lib/layman/pen
[gentoo-user] You need at least GCC 4.7.x or Clang = 3.0 for C++11-specific compiler flags
Hello mates, I am facing the following issue, not sure why I have been trying to fix it for 3 days now. I've also installed Clang, but the same results. Additional information: ~AMD64 emerge --info here: http://tny.cz/c2210f24 Build log: ^[[32;01m * ^[[39;49;00mPackage:net-libs/webkit-gtk-2.0.4 ^[[32;01m * ^[[39;49;00mRepository: gentoo ^[[32;01m * ^[[39;49;00mMaintainer: gn...@gentoo.org ^[[32;01m * ^[[39;49;00mUSE:amd64 elibc_glibc geoloc gstreamer introspection jit kernel_linux userland_GNU webgl ^[[32;01m * ^[[39;49;00mFEATURES: preserve-libs sandbox userpriv usersandbox ^[[31;01m*^[[0m ERROR: net-libs/webkit-gtk-2.0.4::gentoo failed (pretend phase): ^[[31;01m*^[[0m You need at least GCC 4.7.x or Clang = 3.0 for C++11-specific compiler flags ^[[31;01m*^[[0m ^[[31;01m*^[[0m Call stack: ^[[31;01m*^[[0m ebuild.sh, line 93: Called pkg_pretend ^[[31;01m*^[[0m webkit-gtk-2.0.4.ebuild, line 93: Called die ^[[31;01m*^[[0m The specific snippet of code: ^[[31;01m*^[[0mdie You need at least GCC 4.7.x or Clang = 3.0 for C++11-specific compiler flags ^[[31;01m*^[[0m ^[[31;01m*^[[0m If you need support, post the output of `emerge --info '=net-libs/webkit-gtk-2.0.4::gentoo'`, ^[[31;01m*^[[0m the complete build log and the output of `emerge -pqv '=net-libs/webkit-gtk-2.0.4::gentoo'`. ^[[31;01m*^[[0m The complete build log is located at '/var/tmp/portage/net-libs/webkit-gtk-2.0.4/temp/build.log'. ^[[31;01m*^[[0m The ebuild environment file is located at '/var/tmp/portage/net-libs/webkit-gtk-2.0.4/temp/die.env'. ^[[31;01m*^[[0m Working directory: '/usr/lib64/portage/pym' ^[[31;01m*^[[0m S: '/var/tmp/portage/net-libs/webkit-gtk-2.0.4/work/webkitgtk-2.0.4' Any help? -- Carlos Sura.- www.carlossura.com www.carlossura.com/blog
Re: [gentoo-user] llvm compile error
FreeBSD use it and say is stable. FreeBSD maybe not the reference on earth but the BSD's make a good job. When i saw all versions of Clang is masked. Isn't FBSD more similar than, say, Linux, to what OSX is, sort of? I could be way off. But I think of OSX as being FBSD built for people that don't know how to use a computer.
Re: [gentoo-user] llvm / clang compile error
On Wed, 17 Aug 2016 08:36:12 -0400 jameswrote: > On 08/16/2016 04:43 PM, Silvio Siefke wrote: > > Please submit a full bug report https://bugs.gentoo.org/show_bug.cgi?id=591514 Silvio
Re: [gentoo-user] llvm / clang compile error
On Wed, 17 Aug 2016 08:36:12 -0400 jameswrote: Okay i will open Bug Report. But can it be that the memory / swap is to small? cat /proc/meminfo MemTotal:2024472 kB cat /proc/swaps FilenameTypeSizeUsedPriority /dev/sda2 partition 2046972 1741732 -1 Silvio
Re: [gentoo-user] llvm / clang compile error
On 08/17/2016 10:24 AM, siefke_lis...@web.de wrote: On Wed, 17 Aug 2016 08:36:12 -0400 james <gar...@verizon.net> wrote: On 08/16/2016 04:43 PM, Silvio Siefke wrote: Please submit a full bug report https://bugs.gentoo.org/show_bug.cgi?id=591514 Silvio Reading the bug, here is a request for specific info:: " Alex Xu (Hello71) 2016-08-17 15:26:17 UTC please attach example build.log" If you work with the requests on BGO, then they are much more likely to work on your bug, or at least process. If there are other similar bugs in BGO, then stay on top of them too. I've only mildly used clang, at this point, so it'll take somebody else to spot what's wrong, or who knows clang more deeply. But, now you can keep your posts short and precise in gentoo-user and reference your but like this BGO#591514. Also, if you are not copied on the postings to your bug, go ahead and add yourself directly to the "cc list". You can also use gentoo forums or an appropriate gentoo irc channel to look for CLANG/llvm types of folks, referring them to your posted bug too, so your posts there can be short and easy to read. [1] [1] https://wiki.gentoo.org/wiki/Google_Summer_of_Code/2016/Ideas/Clang_native_support hth, James
Re: [gentoo-user] llvm / clang compile error
On Tue, 16 Aug 2016 21:40:40 -0400 Fernando Rodriguezwrote: Hi, > If you havent, try disabling ccache to rule that out. i have do now. Same mistake. Same error message. Silvio pgp0yrcGvf49z.pgp Description: PGP signature
Re: [gentoo-user] Is there a reason why LLVM/Clang ebuilds don't support "mutislot"?
On Sun, Aug 28, 2016 at 11:58:03PM -0400, P Levine wrote: > Other distros like Ubuntu support the installation of multiple versions of > LLVM/Clang side by side. One of the things Clang is really good at is > support for the most recently approved upcoming features of the C++17 > standard. The best support for testing such features is with the latest > sys-devel/llvm-. However if I want to compile Mesa against a stable > version LLVM/Clang as well, I don't get that option. I'm not sure such a configuration is fully supported upstream, but the way ubuntu (and debian) does this is (was?) painfully broken. If I recall correctly they first build it in one location and then move it around and try to partially solve things through symlinks. This totally destroys llvm-config and LLVM's cmake find_package module meaning those things have hard-coded paths that have nothing to do with where the things are installed on the system. Simple example, the following CMakeLists.txt cmake_minimum_required(VERSION 3.4) project(llvm-test) find_package(LLVM) errors out on ubuntu with: CMake Error at /usr/share/llvm-3.8/cmake/LLVMConfig.cmake:178 (include): include could not find load file: /usr/share/llvm/cmake/LLVMExports.cmake Call Stack (most recent call first): CMakeLists.txt:4 (find_package) CMake Error at /usr/share/llvm-3.8/cmake/LLVMConfig.cmake:181 (include): include could not find load file: /usr/share/llvm/cmake/LLVM-Config.cmake Call Stack (most recent call first): CMakeLists.txt:4 (find_package) -- Configuring incomplete, errors occurred! Fixing this one path leads to more suffering further down the road. The way gentoo maintainers package LLVM is much saner and developer friendly.
[gentoo-user] Re: Clang has gone walkabout
On 04/10/2017 03:58 PM, Simon Thelen wrote: Try running `env-update && source /etc/profile'. Your path should be extended by /etc/profile.env which is generated from /etc/env.d/10llvm-9995. Just logout/login. "source" will help in the current shell.
[gentoo-user] Re: sys-devel/llvm and LLVM_TARGETS
On 12/03/2022 10:43, Dale wrote: https://bugs.gentoo.org/767700 Is that the one? It mentions the target but I don't quite understand the why. The biggest thing, will this break something if I let it do it? No. Unlike GCC, LLVM/Clang is always a cross-compiler. This just enables some extra targets. It won't actually affect anything other than perhaps the binaries becoming a bit larger.
Re: [gentoo-user] Emerge trouble with firefox and thunderbird ...
On Sunday, 10 March 2024 07:17:27 GMT Walter Dnes wrote: > So there are at least 2 people who've found out that Firefox can and > *MUST* be built with USE="-clang". Ah. I'll change my USE flag straight away. Thanks Walter. -- Regards, Peter.
Re: [gentoo-user] Firefox fails to compile. crc32 error??
Daniel Pielmeier wrote: > Looks like it is related to -march=native. > > See bug https://bugs.gentoo.org/838373 > Nailed it. After the post by Michael, I edited the kernel several times. I'm sure I ruled that thing out. I had some crc stuff not there but same error with or without. I admit, I wouldn't have thought of a kernel driver issue but the error did sort of lead that way. Given the CPU FLAGS is about the CPU and its instruction set, it could be a option. I would never have thought of the CFLAGS in make.conf. I'm not sure how you figured that out either. o_O Before trying new settings for that, I tried newer versions of clang, llvm and their friends to see if there was a fix that wasn't applied to older versions. When you unmask one thing, it snowballs a bit. Anyway, after getting a more recent version of clang and llvm, it still failed. At that point, I suspected that I had ruled out those packages. I did some digging to find what my CFLAG settings should be if done manually. It took some digging tho. Once I set that to the manual way, it compiled successfully on the first try. So, my backup rig now has a web browser. That's good. ;-) I wonder why that bug report wasn't in the search results when I was digging for the error??? If I see something Gentoo related, I always look. While at it. I don't think there is a way but I may have missed it. As a example, when I wanted to unmask/keyword specific versions of llvm and clang, is there a tool to find out what all else has to also be unmasked/keyworded? When I tried to do that for the packages I wanted, I ended up running emerge to find more that had to be added to the list. This is the list I ended up with. sys-devel/clang sys-devel/llvm sys-devel/clang-common sys-devel/clang-runtime sys-devel/clang-toolchain-symlinks sys-libs/compiler-rt sys-libs/compiler-rt-sanitizers sys-libs/libomp sys-devel/llvm-toolchain-symlinks sys-devel/llvmgold It would be nice if there is a way to just run a command once and it spits out a friendly list that can be added to the proper file. Save annoying some electrons and all. ;-) Thanks much to you both. I hope if someone else runs into this, this thread pops up in the search results. Save someone some head scratching. Dale :-) :-) P. S. Switched computer case door to left side hinge the other day. Put in the drive trays and figuring out if there is a way to put drives in other places and still have lots of fans. :-D
Re: [gentoo-user] Epic list of total FAIL.
Seeing segfaults in a compile like that makes me question your hardware rather than the gentoo tools. Are you sure your hardware is in a functional state? I update my world fairly often and am running ~amd64 and have not experienced the issue you are experiencing. :( On Thu, Aug 20, 2015 at 4:26 PM, Alan Grimes alonz...@verizon.net wrote: My five year old CPU has been working it's ass off the last few days doing a full --emptytree world to try to purge the system of the ncurses clusterfuck. Here is the list of FAIL. A few of these might be stale listings because I didn't purge the directory before running this. Also a few of these might work if I tried it again because a document generator was in ncurses-fail mode when it attempted to build the relevant package. One of these, however, was OMFGROTFLMFAO funny... You'll see what I mean below. What does it say about the state of linux when the compiler is too broken to compile a bug reporter module? =P I usually tolerate a moderate failure %-age but these packages are far too important to the things I need to do to be acceptable. =| tortoise portage # pwd /var/tmp/portage tortoise portage # tree -L 2 . ├── app-arch │ └── rpm-4.12.0.1 ├── app-doc │ └── doxygen-1.8.10-r1 ├── app-office │ ├── libreoffice-4.4.5.2 │ └── texmacs-1.99.2-r1 ├── dev-db │ └── mysql-workbench-6.3.4 ├── dev-dotnet │ └── nuget-2.8.3 ├── dev-java │ └── antlr-3.1.3-r3 ├── dev-libs │ ├── libcdio-0.93 │ ├── libcdio-paranoia-0.93_p1 │ └── libevdev-1.4.3 ├── dev-util │ ├── kdevplatform-1.7.1 │ └── monodevelop-5.9.5.9 ├── kde-apps │ ├── kdesdk-kioslaves-4.14.3 │ └── libkdcraw-4.14.3 ├── media-gfx │ └── digikam-4.12.0 ├── media-libs │ ├── libkface-4.12.0 │ └── mesa-10.6.3 ├── media-sound │ └── playmidi-2.5-r2 ├── media-video │ └── vcdimager-0.7.24 ├── sci-libs │ └── gdal-2.0.0 └── sys-devel └── llvm-3.6.2 36 directories, 0 files tortoise portage # ### In file included from /usr/lib/gcc/x86_64-pc-linux-gnu/4.9.3/include/g++-v4/functional:55:0, from /var/tmp/portage/sys-devel/llvm-3.6.2/work/llvm-3.6.2.src/tools/clang/lib/StaticAnalyzer/Core/../../../include/clang/Basic/SourceLocation.h:22, from /var/tmp/portage/sys-devel/llvm-3.6.2/work/llvm-3.6.2.src/tools/clang/lib/StaticAnalyzer/Core/../../../include/clang/StaticAnalyzer/Core/BugReporter/BugReporter.h:18, from /var/tmp/portage/sys-devel/llvm-3.6.2/work/llvm-3.6.2.src/tools/clang/lib/StaticAnalyzer/Core/BugReporter.cpp:15: /usr/lib/gcc/x86_64-pc-linux-gnu/4.9.3/include/g++-v4/tuple: In constructor ‘constexpr std::tuple_T1, _T2::tuple(_U1, _U2) [with _U1 = clang::ento::LikelyFalsePositiveSuppressionBRVisitor*; _U2 = std::default_deleteclang::ento::LikelyFalsePositiveSuppressionBRVisitor; template-parameter-2-3 = void; _T1 = clang::ento::BugReporterVisitor*; _T2 = std::default_deleteclang::ento::BugReporterVisitor]’: /usr/lib/gcc/x86_64-pc-linux-gnu/4.9.3/include/g++-v4/tuple:539:19: internal compiler error: Segmentation fault constexpr tuple(_U1 __a1, _U2 __a2) ^ Please submit a full bug report, with preprocessed source if appropriate. See https://bugs.gentoo.org/ for instructions. /bin/rm: cannot remove ‘/var/tmp/portage/sys-devel/llvm-3.6.2/work/llvm-3.6.2.src-abi_x86_64.amd64/tools/clang/lib/StaticAnalyzer/Core/Release/BugReporter.d.tmp’: No such file or directory /var/tmp/portage/sys-devel/llvm-3.6.2/work/llvm-3.6.2.src/Makefile.rules:1514: recipe for target '/var/tmp/portage/sys-devel/llvm-3.6.2/work/llvm-3.6.2.src-abi_x86_64.amd64/tools/clang/lib/StaticAnalyzer/Core/Release/BugReporter.o' failed make[5]: *** [/var/tmp/portage/sys-devel/llvm-3.6.2/work/llvm-3.6.2.src-abi_x86_64.amd64/tools/clang/lib/StaticAnalyzer/Core/Release/BugReporter.o] Error 1 make[5]: *** Waiting for unfinished jobs -- IQ is a measure of how stupid you feel. Powers are not rights. -- The irony is that Bill Gates claims to be making a stable operating system and Linus Torvalds claims to be trying to take over the world. -- seen on the net
Re: [gentoo-user] Epic list of total FAIL.
On Thursday, August 20, 2015 4:26:08 PM Alan Grimes wrote: My five year old CPU has been working it's ass off the last few days doing a full --emptytree world to try to purge the system of the ncurses clusterfuck. Here is the list of FAIL. A few of these might be stale listings because I didn't purge the directory before running this. Also a few of these might work if I tried it again because a document generator was in ncurses-fail mode when it attempted to build the relevant package. One of these, however, was OMFGROTFLMFAO funny... You'll see what I mean below. What does it say about the state of linux when the compiler is too broken to compile a bug reporter module? =P I usually tolerate a moderate failure %-age but these packages are far too important to the things I need to do to be acceptable. =| tortoise portage # pwd /var/tmp/portage tortoise portage # tree -L 2 . ├── app-arch │ └── rpm-4.12.0.1 ├── app-doc │ └── doxygen-1.8.10-r1 ├── app-office │ ├── libreoffice-4.4.5.2 │ └── texmacs-1.99.2-r1 ├── dev-db │ └── mysql-workbench-6.3.4 ├── dev-dotnet │ └── nuget-2.8.3 ├── dev-java │ └── antlr-3.1.3-r3 ├── dev-libs │ ├── libcdio-0.93 │ ├── libcdio-paranoia-0.93_p1 │ └── libevdev-1.4.3 ├── dev-util │ ├── kdevplatform-1.7.1 │ └── monodevelop-5.9.5.9 ├── kde-apps │ ├── kdesdk-kioslaves-4.14.3 │ └── libkdcraw-4.14.3 ├── media-gfx │ └── digikam-4.12.0 ├── media-libs │ ├── libkface-4.12.0 │ └── mesa-10.6.3 ├── media-sound │ └── playmidi-2.5-r2 ├── media-video │ └── vcdimager-0.7.24 ├── sci-libs │ └── gdal-2.0.0 └── sys-devel └── llvm-3.6.2 36 directories, 0 files tortoise portage # ### In file included from /usr/lib/gcc/x86_64-pc-linux-gnu/4.9.3/include/g++-v4/functional:55:0, from /var/tmp/portage/sys- devel/llvm-3.6.2/work/llvm-3.6.2.src/tools/clang/lib/StaticAnalyzer/Core/../../../include/clang/Basic/SourceLocation.h:22, from /var/tmp/portage/sys- devel/llvm-3.6.2/work/llvm-3.6.2.src/tools/clang/lib/StaticAnalyzer/Core/../../../include/clang/StaticAnalyzer/Core/BugReporter/BugReporter.h:18, from /var/tmp/portage/sys- devel/llvm-3.6.2/work/llvm-3.6.2.src/tools/clang/lib/StaticAnalyzer/Core/BugReporter.cpp:15: /usr/lib/gcc/x86_64-pc-linux-gnu/4.9.3/include/g++-v4/tuple: In constructor ‘constexpr std::tuple_T1, _T2::tuple(_U1, _U2) [with _U1 = clang::ento::LikelyFalsePositiveSuppressionBRVisitor*; _U2 = std::default_deleteclang::ento::LikelyFalsePositiveSuppressionBRVisitor; template-parameter-2-3 = void; _T1 = clang::ento::BugReporterVisitor*; _T2 = std::default_deleteclang::ento::BugReporterVisitor]’: /usr/lib/gcc/x86_64-pc-linux-gnu/4.9.3/include/g++-v4/tuple:539:19: internal compiler error: Segmentation fault constexpr tuple(_U1 __a1, _U2 __a2) ^ Please submit a full bug report, with preprocessed source if appropriate. See https://bugs.gentoo.org/ for instructions. /bin/rm: cannot remove ‘/var/tmp/portage/sys-devel/llvm-3.6.2/work/llvm-3.6.2.src- abi_x86_64.amd64/tools/clang/lib/StaticAnalyzer/Core/Release/BugReporter.d.tmp’: No such file or directory /var/tmp/portage/sys- devel/llvm-3.6.2/work/llvm-3.6.2.src/Makefile.rules:1514: recipe for target '/var/tmp/portage/sys-devel/llvm-3.6.2/work/llvm-3.6.2.src- abi_x86_64.amd64/tools/clang/lib/StaticAnalyzer/Core/Release/BugReporter.o' failed make[5]: *** [/var/tmp/portage/sys-devel/llvm-3.6.2/work/llvm-3.6.2.src- abi_x86_64.amd64/tools/clang/lib/StaticAnalyzer/Core/Release/BugReporter.o] Error 1 make[5]: *** Waiting for unfinished jobs The ncurses ebuild is indeed broken, I ran into the same problem before. But you received good advice on your last thread (build libtinfo on another system and copy it or just try symlinking it to ncurses), if you'd followed it you would a got your system back up in a few minutes. Good luck, -- Fernando Rodriguez
[gentoo-user] Epic list of total FAIL.
My five year old CPU has been working it's ass off the last few days doing a full --emptytree world to try to purge the system of the ncurses clusterfuck. Here is the list of FAIL. A few of these might be stale listings because I didn't purge the directory before running this. Also a few of these might work if I tried it again because a document generator was in ncurses-fail mode when it attempted to build the relevant package. One of these, however, was OMFGROTFLMFAO funny... You'll see what I mean below. What does it say about the state of linux when the compiler is too broken to compile a bug reporter module? =P I usually tolerate a moderate failure %-age but these packages are far too important to the things I need to do to be acceptable. =| tortoise portage # pwd /var/tmp/portage tortoise portage # tree -L 2 . ├── app-arch │ └── rpm-4.12.0.1 ├── app-doc │ └── doxygen-1.8.10-r1 ├── app-office │ ├── libreoffice-4.4.5.2 │ └── texmacs-1.99.2-r1 ├── dev-db │ └── mysql-workbench-6.3.4 ├── dev-dotnet │ └── nuget-2.8.3 ├── dev-java │ └── antlr-3.1.3-r3 ├── dev-libs │ ├── libcdio-0.93 │ ├── libcdio-paranoia-0.93_p1 │ └── libevdev-1.4.3 ├── dev-util │ ├── kdevplatform-1.7.1 │ └── monodevelop-5.9.5.9 ├── kde-apps │ ├── kdesdk-kioslaves-4.14.3 │ └── libkdcraw-4.14.3 ├── media-gfx │ └── digikam-4.12.0 ├── media-libs │ ├── libkface-4.12.0 │ └── mesa-10.6.3 ├── media-sound │ └── playmidi-2.5-r2 ├── media-video │ └── vcdimager-0.7.24 ├── sci-libs │ └── gdal-2.0.0 └── sys-devel └── llvm-3.6.2 36 directories, 0 files tortoise portage # ### In file included from /usr/lib/gcc/x86_64-pc-linux-gnu/4.9.3/include/g++-v4/functional:55:0, from /var/tmp/portage/sys-devel/llvm-3.6.2/work/llvm-3.6.2.src/tools/clang/lib/StaticAnalyzer/Core/../../../include/clang/Basic/SourceLocation.h:22, from /var/tmp/portage/sys-devel/llvm-3.6.2/work/llvm-3.6.2.src/tools/clang/lib/StaticAnalyzer/Core/../../../include/clang/StaticAnalyzer/Core/BugReporter/BugReporter.h:18, from /var/tmp/portage/sys-devel/llvm-3.6.2/work/llvm-3.6.2.src/tools/clang/lib/StaticAnalyzer/Core/BugReporter.cpp:15: /usr/lib/gcc/x86_64-pc-linux-gnu/4.9.3/include/g++-v4/tuple: In constructor ‘constexpr std::tuple_T1, _T2::tuple(_U1, _U2) [with _U1 = clang::ento::LikelyFalsePositiveSuppressionBRVisitor*; _U2 = std::default_deleteclang::ento::LikelyFalsePositiveSuppressionBRVisitor; template-parameter-2-3 = void; _T1 = clang::ento::BugReporterVisitor*; _T2 = std::default_deleteclang::ento::BugReporterVisitor]’: /usr/lib/gcc/x86_64-pc-linux-gnu/4.9.3/include/g++-v4/tuple:539:19: internal compiler error: Segmentation fault constexpr tuple(_U1 __a1, _U2 __a2) ^ Please submit a full bug report, with preprocessed source if appropriate. See https://bugs.gentoo.org/ for instructions. /bin/rm: cannot remove ‘/var/tmp/portage/sys-devel/llvm-3.6.2/work/llvm-3.6.2.src-abi_x86_64.amd64/tools/clang/lib/StaticAnalyzer/Core/Release/BugReporter.d.tmp’: No such file or directory /var/tmp/portage/sys-devel/llvm-3.6.2/work/llvm-3.6.2.src/Makefile.rules:1514: recipe for target '/var/tmp/portage/sys-devel/llvm-3.6.2/work/llvm-3.6.2.src-abi_x86_64.amd64/tools/clang/lib/StaticAnalyzer/Core/Release/BugReporter.o' failed make[5]: *** [/var/tmp/portage/sys-devel/llvm-3.6.2/work/llvm-3.6.2.src-abi_x86_64.amd64/tools/clang/lib/StaticAnalyzer/Core/Release/BugReporter.o] Error 1 make[5]: *** Waiting for unfinished jobs -- IQ is a measure of how stupid you feel. Powers are not rights.
Re: [gentoo-user] Epic list of total FAIL.
Let me describe what I see. This can't be a clusterfuck, as it is affecting only you. No-one else to my knowledge is reporting problems caused by ncurses. So, it is then highly likely that you have a setup that the devs did not consider, and it is rare (if not unique). So, what exactly did you do to fuck your system up this badly? Don;t say I ran emerge world as lots of other people do that without issue. Before that, perhaps long ago, what did YOU do that caused this current issue? Ranting on the list might make you feel better, but is not likely to fix your problem. Just saying. On 20/08/2015 22:26, Alan Grimes wrote: My five year old CPU has been working it's ass off the last few days doing a full --emptytree world to try to purge the system of the ncurses clusterfuck. Here is the list of FAIL. A few of these might be stale listings because I didn't purge the directory before running this. Also a few of these might work if I tried it again because a document generator was in ncurses-fail mode when it attempted to build the relevant package. One of these, however, was OMFGROTFLMFAO funny... You'll see what I mean below. What does it say about the state of linux when the compiler is too broken to compile a bug reporter module? =P I usually tolerate a moderate failure %-age but these packages are far too important to the things I need to do to be acceptable. =| tortoise portage # pwd /var/tmp/portage tortoise portage # tree -L 2 . ├── app-arch │ └── rpm-4.12.0.1 ├── app-doc │ └── doxygen-1.8.10-r1 ├── app-office │ ├── libreoffice-4.4.5.2 │ └── texmacs-1.99.2-r1 ├── dev-db │ └── mysql-workbench-6.3.4 ├── dev-dotnet │ └── nuget-2.8.3 ├── dev-java │ └── antlr-3.1.3-r3 ├── dev-libs │ ├── libcdio-0.93 │ ├── libcdio-paranoia-0.93_p1 │ └── libevdev-1.4.3 ├── dev-util │ ├── kdevplatform-1.7.1 │ └── monodevelop-5.9.5.9 ├── kde-apps │ ├── kdesdk-kioslaves-4.14.3 │ └── libkdcraw-4.14.3 ├── media-gfx │ └── digikam-4.12.0 ├── media-libs │ ├── libkface-4.12.0 │ └── mesa-10.6.3 ├── media-sound │ └── playmidi-2.5-r2 ├── media-video │ └── vcdimager-0.7.24 ├── sci-libs │ └── gdal-2.0.0 └── sys-devel └── llvm-3.6.2 36 directories, 0 files tortoise portage # ### In file included from /usr/lib/gcc/x86_64-pc-linux-gnu/4.9.3/include/g++-v4/functional:55:0, from /var/tmp/portage/sys-devel/llvm-3.6.2/work/llvm-3.6.2.src/tools/clang/lib/StaticAnalyzer/Core/../../../include/clang/Basic/SourceLocation.h:22, from /var/tmp/portage/sys-devel/llvm-3.6.2/work/llvm-3.6.2.src/tools/clang/lib/StaticAnalyzer/Core/../../../include/clang/StaticAnalyzer/Core/BugReporter/BugReporter.h:18, from /var/tmp/portage/sys-devel/llvm-3.6.2/work/llvm-3.6.2.src/tools/clang/lib/StaticAnalyzer/Core/BugReporter.cpp:15: /usr/lib/gcc/x86_64-pc-linux-gnu/4.9.3/include/g++-v4/tuple: In constructor ‘constexpr std::tuple_T1, _T2::tuple(_U1, _U2) [with _U1 = clang::ento::LikelyFalsePositiveSuppressionBRVisitor*; _U2 = std::default_deleteclang::ento::LikelyFalsePositiveSuppressionBRVisitor; template-parameter-2-3 = void; _T1 = clang::ento::BugReporterVisitor*; _T2 = std::default_deleteclang::ento::BugReporterVisitor]’: /usr/lib/gcc/x86_64-pc-linux-gnu/4.9.3/include/g++-v4/tuple:539:19: internal compiler error: Segmentation fault constexpr tuple(_U1 __a1, _U2 __a2) ^ Please submit a full bug report, with preprocessed source if appropriate. See https://bugs.gentoo.org/ for instructions. /bin/rm: cannot remove ‘/var/tmp/portage/sys-devel/llvm-3.6.2/work/llvm-3.6.2.src-abi_x86_64.amd64/tools/clang/lib/StaticAnalyzer/Core/Release/BugReporter.d.tmp’: No such file or directory /var/tmp/portage/sys-devel/llvm-3.6.2/work/llvm-3.6.2.src/Makefile.rules:1514: recipe for target '/var/tmp/portage/sys-devel/llvm-3.6.2/work/llvm-3.6.2.src-abi_x86_64.amd64/tools/clang/lib/StaticAnalyzer/Core/Release/BugReporter.o' failed make[5]: *** [/var/tmp/portage/sys-devel/llvm-3.6.2/work/llvm-3.6.2.src-abi_x86_64.amd64/tools/clang/lib/StaticAnalyzer/Core/Release/BugReporter.o] Error 1 make[5]: *** Waiting for unfinished jobs -- Alan McKinnon alan.mckin...@gmail.com
Re: [gentoo-user] eix and bad colors.
On 12/08/2012 12:26:35 AM, Dale wrote: If I figure out something or Helmut's config works, I'll post back. Here is my ebuild app-portage/eix-.ebuild in my local overlay. Hopefully it helps, Helmut. # Copyright 1999-2012 Gentoo Foundation # Distributed under the terms of the GNU General Public License v2 # $Header: $ EAPI=5 EGIT_REPO_URI=git://git.berlios.de/${PN} EGIT_PROJECT=${PN}.git [ -n ${EVCS_OFFLINE} ] || EGIT_REPACK=true WANT_LIBTOOL=none PLOCALES=de ru inherit autotools bash-completion-r1 eutils git-2 l10n multilib DESCRIPTION=Search and query ebuilds, portage incl. local settings, ext. overlays, version changes, and more HOMEPAGE=http://eix.berlios.de; SRC_URI= PROPERTIES=live LICENSE=GPL-2 SLOT=0 KEYWORDS= IUSE=clang debug +dep doc nls optimization security strong-optimization sqlite tools zsh-completion RDEPEND=app-shells/push sqlite? ( =dev-db/sqlite-3 ) nls? ( virtual/libintl ) DEPEND=${RDEPEND} clang? ( sys-devel/clang ) sys-devel/gettext pkg_setup() { if has_version ${CATEGORY}/${PN}-0.25.3; then local eixcache=${EROOT}/var/cache/${PN} ! test -f ${eixcache} || rm -f -- ${eixcache} fi } src_prepare() { epatch_user eautopoint eautoreconf } src_configure() { econf $(use_with sqlite) $(use_with doc extra-doc) \ $(use_with zsh-completion) \ $(use_enable nls) $(use_enable tools separate-tools) \ $(use_enable security) $(use_enable optimization) \ $(use_enable strong-optimization) $(use_enable debug debugging) \ $(use_with prefix always-accept-keywords) \ $(use_with dep dep-default) \ $(use_with clang nongnu-cxx clang++) \ --with-ebuild-sh-default=/usr/$(get_libdir)/portage/bin/ebuild.sh \ --with-portage-rootpath=${ROOTPATH} \ --with-eprefix-default=${EPREFIX} \ --docdir=${EPREFIX}/usr/share/doc/${PF} \ --htmldir=${EPREFIX}/usr/share/doc/${PF}/html } src_install() { default dobashcomp bash/eix keepdir /var/cache/${PN} fowners portage:portage /var/cache/${PN} fperms 775 /var/cache/${PN} } pkg_postinst() { # fowners in src_install doesn't work for owner/group portage: # merging changes this owner/group back to root. use prefix || chown portage:portage ${EROOT}var/cache/${PN} local obs=${EROOT}var/cache/eix.previous ! test -f ${obs} || ewarn Found obsolete ${obs}, please remove it }
Re: [gentoo-user] Sweet Sweet Portage
On 08/15/2016 06:02 PM, james wrote: > Well, > > I brought this up before. No need for argument, just test it out > for yourself. > > Multiple times (over the last few weeks) I have run 'emerge -uDNvp @world' > and there are issues to deal with manually. > > For example 'One or more updates/rebuilds have been skipped due to a > dependency conflict', type of fudd and other types of fudd is the result, not > all the time, but maybe 50% of the time. > > > Now, routinely, all I do is immediately issue this command > 'emerge -uDt @world' and go have a coffee. An AMD 8 core, 32G workstation > does it's thang, leaving me a with just a smile after the work is complete. > No other actions, nadda, ziltch. Immediately, I then run 'emerge -uDNvp > world' (again, and routinely I get:: > > "These are the packages that would be merged, in order: > > Calculating dependencies... done! > > Total: 0 packages, Size of downloads: 0 KiB" > > > Just try it for yourself. It's like clockwork now. Smooth. I have over 1500 > packages installed on a mostly stable but hacked out /usr/local/portage/ and > maybe 10% of the packages, that are much newer, but portage is sweet, sweet, > sweet now. There is inherent magic now, but, > I do not have time to ferret it out. Sure I can dive in, manually, > and I have done this to fix things, but, 'emerge -uDt @world' fixes things, > automagically; dozens of times as I update 3 or 4 times a week. > > I don't know exactly what's going on but I think something is wrong so it's not so sweet. I think you got a conflict that's not being resolved and not being pulled by the second command. What happends if you add --with-bdeps=y --backtrack=30? Also try without the -p (I think it runs more code like autounmask etc so it may cause the extra output). Is clang/llvm stuff still popping up on the list of skipped packages? I remember a similar conflict around the time of your first post and it turned out that the latest stable clang blocks the latest stable libclc. So the tree is (still) broken. For most users it's not a problem because portage pulls the right version of clang but if you have clang on your world file it updates it to the latest and you get those conflicts. I fixed it by masking all versions of clang >3.6 -- Fernando Rodriguez
Re: [gentoo-user] Sweet Sweet Portage
On 08/16/2016 02:49 PM, Fernando Rodriguez wrote: On 08/15/2016 06:02 PM, james wrote: Well, I brought this up before. No need for argument, just test it out for yourself. Multiple times (over the last few weeks) I have run 'emerge -uDNvp @world' and there are issues to deal with manually. For example 'One or more updates/rebuilds have been skipped due to a dependency conflict', type of fudd and other types of fudd is the result, not all the time, but maybe 50% of the time. Now, routinely, all I do is immediately issue this command 'emerge -uDt @world' and go have a coffee. An AMD 8 core, 32G workstation does it's thang, leaving me a with just a smile after the work is complete. No other actions, nadda, ziltch. Immediately, I then run 'emerge -uDNvp world' (again, and routinely I get:: "These are the packages that would be merged, in order: Calculating dependencies... done! Total: 0 packages, Size of downloads: 0 KiB" Just try it for yourself. It's like clockwork now. Smooth. I have over 1500 packages installed on a mostly stable but hacked out /usr/local/portage/ and maybe 10% of the packages, that are much newer, but portage is sweet, sweet, sweet now. There is inherent magic now, but, I do not have time to ferret it out. Sure I can dive in, manually, and I have done this to fix things, but, 'emerge -uDt @world' fixes things, automagically; dozens of times as I update 3 or 4 times a week. I don't know exactly what's going on but I think something is wrong so it's not so sweet. I think you got a conflict that's not being resolved and not being pulled by the second command. What happends if you add --with-bdeps=y --backtrack=30? Also try without the -p (I think it runs more code like autounmask etc so it may cause the extra output). Is clang/llvm stuff still popping up on the list of skipped packages? I remember a similar conflict around the time of your first post and it turned out that the latest stable clang blocks the latest stable libclc. So the tree is (still) broken. For most users it's not a problem because portage pulls the right version of clang but if you have clang on your world file it updates it to the latest and you get those conflicts. I fixed it by masking all versions of clang >3.6 Nope, but on gentoo-dev there is a big announce about LLVM(clang). No issues with the system, I run the latest portage and when issues popup, granted that are easy to fix, manually 'emerge -uDt world' precludes the need to fix them. Afterwards, running 'emerge -uDNvp world' just comes back completely clean. NO idea what's going on, but *I* have verified this now dozens and of times over the recent weeks. Instead of running 'emerge -uDNv world' I simple run 'emerge -uDt world' and the manual (trivial) items just dissappear, it runs to completion and all is just spanky fine. I do not have the will, nore inclination to dig deeply, but the benefit is most wonderful. ymmv. But follow the evidence emerge -uDNvp world emerge -Dut world emerge -uDNvp world It's just that simple:: 3 commands, no others. fabulous! You don't believe me, then just ignore the thread. ymmv. --nobody cares, as I'm done with this thread. James
[gentoo-user] Re: [OT] LiveCDs for Uni students
systemrescuecd? Too complicated. What I mean by this is that the user upon booting has options!!! Do you kick into a graphical environment, do you copy everything to memory, do you. you get the idea. The problem is that these are first year students in a common first year, they have not yet decided upon their stream, could be Civil, Chem, Mech etc and don't see any benefit in doing programming but have to pass the subject. My requirement is that it boots into a GUI, preferably straight into the environment, no username/password, has dhcp, a decent editor, a browser and gcc or clang. Andrew System Rescue CD fulfils the above requirements. You can go straight to XFCE, be root with no user/password, by default automatically configures network with DHCP. Editor is vim (I believe), browser is Midori, and gcc is present; I didn't think of looking for clang. Tom
Re: [gentoo-user] Re: [OT] LiveCDs for Uni students
On 03/08/2014 09:24 PM, Thomas Mueller wrote: systemrescuecd? Too complicated. What I mean by this is that the user upon booting has options!!! Do you kick into a graphical environment, do you copy everything to memory, do you. you get the idea. The problem is that these are first year students in a common first year, they have not yet decided upon their stream, could be Civil, Chem, Mech etc and don't see any benefit in doing programming but have to pass the subject. My requirement is that it boots into a GUI, preferably straight into the environment, no username/password, has dhcp, a decent editor, a browser and gcc or clang. Andrew System Rescue CD fulfils the above requirements. You can go straight to XFCE, be root with no user/password, by default automatically configures network with DHCP. Editor is vim (I believe), browser is Midori, and gcc is present; I didn't think of looking for clang. Tom Thanks everyone for your comments. In my research, I stumbled across sax, www.sax.org, and it does all that I need. I can either burn it to a cd or a stick, it has the apps I need and is quite small, just under 300MB. Andrew
Re: [gentoo-user] Paraview 4.2 not building
On 12/30/2014 05:26 PM, Helmut Jarausch wrote: On 12/30/2014 05:44:16 AM, Andrew Lowe wrote: Hi all, Anyone out there using Paraview? A quick question before I put all of the supporting doco, build/error logs etc, together. I'm attempting to get the latest version in portage, V4.2, built but I'm running into problems at the 45% stage, Linking libvtkRenderingOpenGL. Has anyone else come across any problems in this area? Any thoughts on how to fix the problem. Hi, it's the same here. It looks like a template error. I'm using gcc 4.9.2 which is very strict w.r.t. the C++ standard. I suggest, you create a bug report. Helmut Will do. I also have Clang installed on this machine and set that to compile paraview. Unlike under gcc where it failed at trying to link libvtkRenderingOpenGL, it is now failing when trying to link libvtkRenderingVolume. The libvtkRenderingOpenGL problem appears to not exist under clang.. Andrew
Re: [gentoo-user] llvm / clang compile error
On 08/16/2016 04:43 PM, Silvio Siefke wrote: Please submit a full bug report Seems like a good recommendation. Also, once the bug is filed, you can provide all the details in the bug report and reference those details in your gentoo-user threads. That way the gentoo-user thread can focus on a specific item or two, because the full details are readily available in BGO. The bug report will enable a wider group to look at your complete details and keep the discussion (here on gentoo-user) short and focused. Just a suggestion, ymmv. Often when I run into issues, I just go to a minimal flag set and just use one processor to compile, as a routine, baseline test-compile run. Also on my system, I have this:: Installed versions: 3.7.1-r3(03:48:50 PM 07/13/2016)(clang libffi ncurses python static-analyzer ABI_X86="32 64 -x32" PYTHON_TARGETS="python2_7" hth, James
Re: [gentoo-user] Re: Spectre-NG
On 09/05/18 19:18, Martin Vaeth wrote: > As mentioned, I wonder why gcc/clang do not yet support this > horribly slow but spectre-safe option. It can't be that hard to > implement in the actual code-producing back-end. Given the response by the gcc team to security people complaining that gcc was optimising out security-sensitive code (namely, a two-fingered salute near enough), I doubt the gcc team would have any interest in optimisations that SLOWED DOWN the resultant code. I suspect that might be one of the forces driving the kernel towards CLANG - a development team that is not obsessed with performance at the expense of breaking any code that uses undefined features. Unfortunately, when dealing with hardware, one is forced to rely on undefined features. A strong point of C, until the compiler decides to go "rogue" on you ... Cheers, Wol
Re: [gentoo-user] Re: gcc 7.3 + kernel 4.15 = spectre_v2 fixed
On Mon, Jan 29, 2018 at 7:50 PM, Ian Zimmerman <i...@very.loosely.org> wrote: > On 2018-01-29 20:11, Adam Carter wrote: > >> Comparing the contents of /sys/devices/system/cpu/vulnerabilities/spectre_v2 >> >> With gcc 7.2 + kernel 4.14.15; >> Intel system shows; Vulnerable: Minimal generic ASM retpoline >> AMD system shows: Vulnerable: Minimal AMD ASM retpoline >> >> With gcc 7.3 + kernel 4.15.0; >> Intel system shows; Mitigation: Full generic retpoline >> AMD system shows' Mitigation: Full AMD retpoline > > Is there a simple way, with the upstream (kernel.org) sources, to force > a compiler different from the system default? If there is, it's not in the > README, and a simple grep over the Makefiles also doesn't enlighten. > > I am not ready to activate a keyworded gcc for general use. > > -- > Please don't Cc: me privately on mailing lists and Usenet, > if you also post the followup to the list or newsgroup. > To reply privately _only_ on Usenet, fetch the TXT record for the domain. > To compile the kernel with a different compiler, the method shown below may be used, e.g.: make CC=clang See [1], for details: Building the kernel with Clang: [1] https://lwn.net/Articles/734071/
Re: [gentoo-user] Comparatively slow clang startup vs. gcc - reasons?
On Tue, Oct 9, 2018 at 1:55 PM Holger Hoffstätte wrote: > Can we do something about this? I remember that llvm had optional static > libs, which apparently were removed completely. Is there something > that can be done with linker tricks (better relocation info?) when > building llvm/clang to speed up the .so loading? > > curious, > Holger > Has anyone tested with -DLLVM_BUILD_LLVM_DYLIB=ON and -DLLVM_LINK_LLVM_DYLIB=ON? It seems to be made to solve this kind of issue. LLVM_BUILD_LLVM_DYLIB: If enabled, the target for building the libLLVM shared library is added. This library contains all of LLVM’s components in a single shared library. Defaults to OFF. This cannot be used in conjunction with BUILD_SHARED_LIBS. Tools will only be linked to the libLLVM shared library if LLVM_LINK_LLVM_DYLIB is also ON. The components in the library can be customised by setting LLVM_DYLIB_COMPONENTS to a list of the desired components. LLVM_LINK_LLVM_DYLIB: If enabled, tools will be linked with the libLLVM shared library. Defaults to OFF. Setting LLVM_LINK_LLVM_DYLIB to ON also sets LLVM_BUILD_LLVM_DYLIB to ON.
Re: [gentoo-user] Remove rust completely
On 12/05/2022 02:41, Mansour Al Akeel wrote: And yes, the compile time is one of the factors in not wanting it on my system. The second factor is a natural reaction toward feeling that I am forced to have it. Another reason is the growing collection of compilers and development tools and their build time (gcc, bin-utils, llvm, clang ... etc.) and now rust. Tongue in cheek, but have you tried doing away with gcc, clang etc? If you're going to use a source-based system, a bunch of compilers "comes with the territory". Many people see Rust as a "better C than C", so that's why it's becoming popular. (Oh, and does Rust have its own compiler, or is Rust just part of llvm?). The way things are going you might find all you need is the llvm collection, and gcc will be obsolete ... Cheers, Wol
Re: [gentoo-user] Thunderbird build failure .. [Gone away]
On 01/01/2023 21:05, Wol wrote: On 01/01/2023 20:08, cal wrote: FWIW, Thunderbird builds fine with GCC on my machine -- I'm unsure of your reasons for setting your Portage compiler to clang, but you may wish to use a package.env override to build Thunderbird with GCC as a workaround until the problem can be fixed upstream. I don't know anything about clang ... it must be the default ... I thought part of Firefox/Thunderbird was written in Rust, so I assumed it was built with llvm as a matter of course. I'll just wait for it to sort itself out. Just to say it's finally sorted itself out without me doing anything, all the while causing random failures in the build chain ie I think Firefox and LLVM all failed along the way ... Cheers, Wol
Re: [gentoo-user] Emerge trouble with firefox and thunderbird ...
On 03/03/2024 23:13, Carsten Hauck wrote: So I don't know what's going on, but basically Mozilla won't emerge, and I don't know why ... Cheers, Wol Did the other 19 package emerge OK? Are the mozilla progs crashing when running, or when emerging? If emerging, the log is just console output, as indecipherable as we know it sometimes can be. If they crash when running, try running from command line. Some time ago on one of my machines Thunderbird and Firefox stopped to compile with USE="clang". As they can be build with gcc I never digged too deep into that problem but maybe it's worth a shot. For anyone else who hits this sort of problem, I did an USE=-clang emerge --update @world (firefox and thunderbird were the only programs I thought this would touch), and it worked. There were a couple of other programs that I guess got pulled in by the changed use, but they've upgraded which is the main thing. Thank you very much Cheers, Wol
[gentoo-user] Re: You need at least GCC 4.7.x or Clang = 3.0 for C++11-specific compiler flags
Nevermind mates, I've just upgraded the GCC to latest and it worked out. Thanks! On 27 July 2013 20:55, Carlos Sura carlos.su...@googlemail.com wrote: Hello mates, I am facing the following issue, not sure why I have been trying to fix it for 3 days now. I've also installed Clang, but the same results. Additional information: ~AMD64 emerge --info here: http://tny.cz/c2210f24 Build log: ^[[32;01m * ^[[39;49;00mPackage:net-libs/webkit-gtk-2.0.4 ^[[32;01m * ^[[39;49;00mRepository: gentoo ^[[32;01m * ^[[39;49;00mMaintainer: gn...@gentoo.org ^[[32;01m * ^[[39;49;00mUSE:amd64 elibc_glibc geoloc gstreamer introspection jit kernel_linux userland_GNU webgl ^[[32;01m * ^[[39;49;00mFEATURES: preserve-libs sandbox userpriv usersandbox ^[[31;01m*^[[0m ERROR: net-libs/webkit-gtk-2.0.4::gentoo failed (pretend phase): ^[[31;01m*^[[0m You need at least GCC 4.7.x or Clang = 3.0 for C++11-specific compiler flags ^[[31;01m*^[[0m ^[[31;01m*^[[0m Call stack: ^[[31;01m*^[[0m ebuild.sh, line 93: Called pkg_pretend ^[[31;01m*^[[0m webkit-gtk-2.0.4.ebuild, line 93: Called die ^[[31;01m*^[[0m The specific snippet of code: ^[[31;01m*^[[0mdie You need at least GCC 4.7.x or Clang = 3.0 for C++11-specific compiler flags ^[[31;01m*^[[0m ^[[31;01m*^[[0m If you need support, post the output of `emerge --info '=net-libs/webkit-gtk-2.0.4::gentoo'`, ^[[31;01m*^[[0m the complete build log and the output of `emerge -pqv '=net-libs/webkit-gtk-2.0.4::gentoo'`. ^[[31;01m*^[[0m The complete build log is located at '/var/tmp/portage/net-libs/webkit-gtk-2.0.4/temp/build.log'. ^[[31;01m*^[[0m The ebuild environment file is located at '/var/tmp/portage/net-libs/webkit-gtk-2.0.4/temp/die.env'. ^[[31;01m*^[[0m Working directory: '/usr/lib64/portage/pym' ^[[31;01m*^[[0m S: '/var/tmp/portage/net-libs/webkit-gtk-2.0.4/work/webkitgtk-2.0.4' Any help? -- Carlos Sura.- www.carlossura.com www.carlossura.com/blog -- Carlos Sura.- www.carlossura.com www.carlossura.com/blog