[SOLVED] Problem compiling Chromium-ungoogled on a raspberry pi 4 with SARPI (SlackWare ARM-Current RPI)
Slackware - ARMThis forum is for the discussion of Slackware ARM.
Notices
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
If you have any problems with the registration process or your account login, please contact us. If you need to reset your password, click here.
Having a problem logging in? Please visit this page to clear all LQ-related cookies.
Get a virtual cloud desktop with the Linux distro that you want in less than five minutes with Shells! With over 10 pre-installed distros to choose from, the worry-free installation life is here! Whether you are a digital nomad or just looking for flexibility, Shells can put your Linux machine on the device that you want to use.
Exclusive for LQ members, get up to 45% off per month. Click here for more info.
Problem compiling Chromium-ungoogled on a raspberry pi 4 with SARPI (SlackWare ARM-Current RPI)
Hello guys n' girls and slackers.
I am trying to compile Chromium-ungoogled on my raspberry pi 4 installed with SARPI (Slackware ARM -Current)
uname -m shows:
Quote:
armv7l
lscpu shows:
Quote:
Architecture: armv7l
Byte Order: Little Endian
CPU(s): 4
On-line CPU(s) list: 0-3
Thread(s) per core: 1
Core(s) per socket: 4
Socket(s): 1
Vendor ID: ARM
Model: 3
Model name: Cortex-A72
Stepping: r0p3
CPU max MHz: 1500.0000
CPU min MHz: 600.0000
BogoMIPS: 198.00
Flags: half thumb fastmult vfp edsp neon vfpv3 tls vfpv4 idiva idivt vfpd32 lpae evtstrm crc3
Do not know if its relevent but i am using raspberry pi´s kernel
as i had to install to get the correct driver for the graphics card and sound card.
Quote:
Linux mediarpi4.not4u.lan 5.10.52-v7l+ #1441 SMP Tue Aug 3 18:11:56 BST 2021 armv7l BCM2711 GNU/Linux
Tried using Alien BOB´s chromium-ungoogled slackbuild.
Managed to edit/change the script so it starts to compile google's clang.
But here is where i get stuck as it complains that there are no compatible triple.
Quote:
error: unable to create target: 'No available targets are compatible with triple "armv7-unknown-linux-gnueabihf"'
At this point i think i have tried everything possible.
Weird thing is that the clang Driver never seems to be able to detect the correct triple:
Starting the compile of clang it shows:
Quote:
-- Performing Test HAS_MAYBE_UNINITIALIZED
-- Performing Test HAS_MAYBE_UNINITIALIZED - Success
-- Native target ARM is not selected; lli will not JIT code
-- Threads enabled.
-- Doxygen disabled.
-- Go bindings enabled.
-- Ninja version: 1.10.2
-- Could NOT find OCaml (missing: OCAMLFIND OCAML_VERSION OCAML_STDLIB_PATH)
-- Could NOT find OCaml (missing: OCAMLFIND OCAML_VERSION OCAML_STDLIB_PATH)
-- OCaml bindings disabled.
-- Could NOT find Python module pygments
-- Could NOT find Python module pygments.lexers.c_cpp
-- Could NOT find Python module yaml
-- LLVM host triple: armv7l-unknown-linux-gnueabihf
-- LLVM default target triple: armv7l-unknown-linux-gnueabihf
-- Performing Test SUPPORTS_FVISIBILITY_INLINES_HIDDEN_FLAG
-- Performing Test SUPPORTS_FVISIBILITY_INLINES_HIDDEN_FLAG - Success
I am trying to compile Chromium-ungoogled on my raspberry pi 4 installed with SARPI (Slackware ARM -Current)
Please i have run out of idears and need some help and hoping some in here has been dealing with this and maybe have a solution to this issue?
Are you intending to run this software on the Raspberry Pi 4 you're building it on? Try building it without specifying any architecture cflags. Just allow the compiler to pick them up from the system (without forcing the wrong ones) and see if it works for you.
Are you intending to run this software on the Raspberry Pi 4 you're building it on? Try building it without specifying any architecture cflags. Just allow the compiler to pick them up from the system (without forcing the wrong ones) and see if it works for you.
Yes, i want to run chromium as a web browser on my Pi 4 and figured it would be easiest to compile the software on the same machine/system
So if i understand you correctly i should just leave the compile flags empty?.
I only use -O2 flag for most of the compiling I do on Slackware ARM. Unless the source code requries specific flags to compile successfully I never specify any more than that. It just creates problems more often than not.
If anything, avoid -march and -mtune flags altogether. The -march flag specifies the target architecture. The -mtune flag specifies the target microarchitecture. You don't need these when compiling natively on ARM architecture. If you were cross-compiling on x86 then it would be a different matter.
[EDIT] Thinking about it, all you should ever need in general when compiling natively on ARM is...
Code:
SLKCFLAGS="-O2 -mcpu=native"
So try that and get rid of all the other cflags, because the -mcpu flag on ARM specifies the target architecture the same as ‑march does on x86 machines. I never use the -mcpu flag.
Quote:
-mcpu=native causes the compiler to auto-detect the CPU of the build computer. At present, this feature is only supported on GNU/Linux, and not all architectures are recognized. If the auto-detect is unsuccessful the option has no effect.
I only use -O2 flag for most of the compiling I do on Slackware ARM. Unless the source code requries specific flags to compile successfully I never specify any more than that. It just creates problems more often than not.
If anything, avoid -march and -mtune flags altogether. The -march flag specifies the target architecture. The -mtune flag specifies the target microarchitecture. You don't need these when compiling natively on ARM architecture. If you were cross-compiling on x86 then it would be a different matter.
[EDIT] Thinking about it, all you should ever need in general when compiling natively on ARM is...
Code:
SLKCFLAGS="-O2 -mcpu=native"
So try that and get rid of all the other cflags, because the -mcpu flag on ARM specifies the target architecture the same as ‑march does on x86 machines. I never use the -mcpu flag.
Tried removing the cflags and and only keeping "-O2", trying out your recommendations but still no go.
However as i looked more closely to the compile logs and realized in despair that the script never passed any cflags or cxxflags at all.
Wasted hours doing this
No clue if it makes any diffrence, but i added this code to the script:
Code:
if [[ "$ARCH" =~ arm.* ]]; then
# When building on ARM there is no need to build all the back-ends
sed -i tools/clang/scripts/build.py -e "s/targets = \x27AArch64;ARM;Mips;PowerPC;SystemZ;WebAssembly;X86\x27/targets = \x27ARM;X86;AArch64\x27/"
# And fix the C_FLAGS AND CXX_FLAGS while we are at it:
sed -i tools/clang/scripts/build.py -e "s/cflags = \[\]/cflags = \['$SLKCFLAGS'\]/"
sed -i tools/clang/scripts/build.py -e "s/cxxflags = \[\]/cxxflags = \['$SLKCFLAGS'\]/"
fi
Also found some info in this guide How To Build On ARM.
Dont know how relevent or old this is, but i figgure its worth a try?
So i´ll be puting it back to compile again for about the 10000´s time now, and will see if thsi changes the result.
Thanks for all your tips and i´ll keep this updated on how-it all goes.
Quote:
Originally Posted by Desiderius
@Minime_2003 you can download the Slackbuild of Chromium 90 of BrunoLafleur
It is a repackaging of the Debian archive, and it works very well on RPI4.
So i´ll be puting it back to compile again for about the 10000´s time now, and will see if thsi changes the result.
Thanks for all your tips and i´ll keep this updated on how-it all goes.
I have to admit, I've not looked into what you're trying to build, although it has to be said that a lot of alienBOB's work is superlative.
In order to be successful, first you need to meet failure many, many times. In my view, those who succeed first time don't learn much. If you're anything like me you'll welcome failure as a means of learning what's right or wrong in all things. However, there can come a time when constant failure can drive someone away from their vision and goals.
BrunoLafleur's GitHub repo that Desiderius gave the link to is simply repackaging the software - which means you don't have to waste time and effort building it yourself. If all you want to achieve is for it to work then that may be your best and easiest option.
I have to admit, I've not looked into what you're trying to build, although it has to be said that a lot of alienBOB's work is superlative.
I couldn't agree more, that guy has done massive and amazing work. No doubt about it. A brilliant guy
Quote:
Originally Posted by Exaga
In order to be successful, first you need to meet failure many, many times. In my view, those who succeed first time don't learn much. If you're anything like me you'll welcome failure as a means of learning what's right or wrong in all things. However, there can come a time when constant failure can drive someone away from their vision and goals.
Thats true, Just get sligtly anoyed when wating 5-6 hours and the results in failure...
And to try something new one has to wait yet another 5-6 hours...really time consuming thats all =)
But i don't mind learning from what i am doing and trying to achieve.
Quote:
Originally Posted by Exaga
BrunoLafleur's GitHub repo that Desiderius gave the link to is simply repackaging the software - which means you don't have to waste time and effort building it yourself. If all you want to achieve is for it to work then that may be your best and easiest option.
Yes, it might be the easiest way resolve it in the end if its not possible to build on the system it self.
However i don't like adding loads of third party (Non slackware libs and dependencies) and i´d like to see if i can get chromium to be able to use Widevine on ARM which i suspect is not even possible if using a debian build for ahole lot of reasons.
Last edited by Minime_2003; 08-19-2021 at 01:10 PM.
Thats true, Just get sligtly anoyed when wating 5-6 hours and the results in failure...
And to try something new one has to wait yet another 5-6 hours...really time consuming thats all =)
Have you specified multi jobs with the 'make -j$(nproc)' option? You can also use the -k option to keep-going which continues as much as possible after an error. This dramatically speeds up any compilation process in general. If you've already included this then I'm sorry to waste your time.
Have you specified multi jobs with the 'make -j$(nproc)' option? You can also use the -k option to keep-going which continues as much as possible after an error. This dramatically speeds up any compilation process in general. If you've already included this then I'm sorry to waste your time.
I usally do this in other build scripts, but as chromium uses loads of phyton stuff for the builds, i dont know
where everything is located.
however i suspect the script is using some kind of "-j" command, as i can see 4-5 threds of the compile prosseses when looking in "top" during the build.
Cant say that its worth trying to get the compiler to keep going when the compilre reports googles clang compiler as broken (Needed later to compile chromium).
So its ok for it to stop there as the error comes at the end when the system is to test the compiler for errors.
I dont get eny other other errors during the compile, so i dont know why or how to get around nore solve this issue?
But as it looks at the moment, its not possible to build googles clang on Slackware Arm...
So i´v been thinking, would it be possible to ise the systems clang version?
How-to solve the problem with googles build using lld as this does not exist in the system?
No worries, all tips and idears are more than welcome and I really appreciate it.
No worries, all tips and idears are more than welcome and I really appreciate it.
I looked into the chromium.SlackBuild by alienBOB and it uses clang to build by default, not gcc. So all the advice I've posted about gcc options is irrelevant.
I looked into the chromium.SlackBuild by alienBOB and it uses clang to build by default, not gcc. So all the advice I've posted about gcc options is irrelevant.
Its ok, well it depends on what system you are running.
Ex. if one uses x86_64 the script downloads googles prebuilt llvm/clang binaries to compile chromium.
If you use i686 the script downloads googles llvm/clang source code and compiles it with gcc i think, then after it has compiled llvm/clang it
switches to googles llvm/clang to compile chromium.
Gave up trying to compile googles llvm/clang on arm, as i looks it does not support building this on arm systems.
So the new goal is to get the script to compile by using the systems own clang instead.
/usr/bin/chromium
cat: /etc/debian_version: No such file or directory
/usr/lib/chromium/chromium: error while loading shared libraries: libjsoncpp.so.24: cannot open shared object file: No such file or directory
I imagine I could fake out /etc/debian_version but will jsoncpp (1.9.2) take care of libjsoncpp.so.24?
@Desiderius which libjsoncpp.so you have installed?
Last edited by glorsplitz; 08-20-2021 at 07:38 PM.
Gave up trying to compile googles llvm/clang on arm, as i looks it does not support building this on arm systems.
So the new goal is to get the script to compile by using the systems own clang instead.
Just to satisfy my curiosity, why Chromium? Is it specifically required for something you're doing/planning or simply your preferred browser?
LinuxQuestions.org is looking for people interested in writing
Editorials, Articles, Reviews, and more. If you'd like to contribute
content, let us know.
