I cannot see any segfaults or errors in there after taking a quick glimpse (only that the make somehow fails). Are you sure you logged everything? I would try using another gcc version.
[quote=“fp1_wo_sw_updates, post:378, topic:11600”]
I cannot see any segfaults or errors in there after taking a quick glimpse (only that the make somehow fails). Are you sure you logged everything?[/quote]
I probably didn’t log stderr, but only stdout. :-/ I’ll give it another try.
I used the default gcc of Ubuntu 14.04. So I believe other have successfully compiled using exactly that version, thus I was doubting that the gcc version is the fault. I can nevertheless try to downgrade the version, I would like to avoid compiling gcc myself…
Somehow I messed up output/error redirect and the log file is empty But here the segfault copied from the bash output (I tried make -j1
):
target thumb C++: third_party_WebKit_Source_core_webcore_remaining_gyp <= external/chromium_org/third_party/WebKit/Source/core/css/parser/MediaQueryTokenizer.cpp
target thumb C++: third_party_WebKit_Source_core_webcore_remaining_gyp <= external/chromium_org/third_party/WebKit/Source/core/css/parser/SizesAttributeParser.cpp
target thumb C++: third_party_WebKit_Source_core_webcore_remaining_gyp <= external/chromium_org/third_party/WebKit/Source/core/css/resolver/AnimatedStyleBuilder.cpp
In file included from external/chromium_org/third_party/WebKit/Source/core/rendering/style/StyleInheritedData.h:29:0,
from external/chromium_org/third_party/WebKit/Source/core/rendering/style/RenderStyle.h:50,
from external/chromium_org/third_party/WebKit/Source/core/rendering/style/CachedUAStyle.h:25,
from external/chromium_org/third_party/WebKit/Source/core/css/resolver/StyleResolverState.h:35,
from external/chromium_org/third_party/WebKit/Source/core/css/resolver/AnimatedStyleBuilder.cpp:58:
external/chromium_org/third_party/WebKit/Source/platform/fonts/Font.h:163:103: internal compiler error: Segmentation fault
FloatRect selectionRectForComplexText(const TextRun&, const FloatPoint&, int h, int from, int to) const;
^
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://source.android.com/source/report-bugs.html> for instructions.
make: *** [out/target/product/FP2/obj/STATIC_LIBRARIES/third_party_WebKit_Source_core_webcore_remaining_gyp_intermediates/third_party/WebKit/Source/core/css/resolver/AnimatedStyleBuilder.o] Fehler 1
#### make failed to build some targets (03:39:23 (hh:mm:ss)) ####
I still used gcc 4.8.4
, but can try with another version next weekend.
The last lines of dmesg
within the vm don’t show anything interesting as far as I can tell:
[ 3.255941] init: failsafe main process (578) killed by TERM signal
[ 3.305337] sound hdaudioC0D0: autoconfig: line_outs=1 (0x3/0x0/0x0/0x0/0x0) type:line
[ 3.305340] sound hdaudioC0D0: speaker_outs=0 (0x0/0x0/0x0/0x0/0x0)
[ 3.305342] sound hdaudioC0D0: hp_outs=0 (0x0/0x0/0x0/0x0/0x0)
[ 3.305343] sound hdaudioC0D0: mono: mono_out=0x0
[ 3.305344] sound hdaudioC0D0: inputs:
[ 3.305346] sound hdaudioC0D0: Line=0x5
[ 4.288898] init: plymouth-upstart-bridge main process ended, respawning
[ 4.300936] init: plymouth-upstart-bridge main process ended, respawning
[ 11.406803] input: spice vdagent tablet as /devices/virtual/input/input4
[ 367.851814] systemd-hostnamed[2100]: Warning: nss-myhostname is not installed. Changing the local hostname might make it unresolveable. Please install nss-myhostname!
And on the host something is jamming dmesg
, I only see plenty of following lines:
[15687.562449] ACPI: \_SB_.PCI0.LPCB.EC0_.ECRD: 1 arguments were passed to a non-method ACPI object (RegionField) (20150619/nsarguments-230)
[15687.562468] ACPI: \_SB_.PCI0.LPCB.EC0_.ECRD: 1 arguments were passed to a non-method ACPI object (RegionField) (20150619/nsarguments-230)
[15689.567889] ACPI: \_SB_.PCI0.LPCB.EC0_.ECRD: 1 arguments were passed to a non-method ACPI object (RegionField) (20150619/nsarguments-230)
[15689.567912] ACPI: \_SB_.PCI0.LPCB.EC0_.ECRD: 1 arguments were passed to a non-method ACPI object (RegionField) (20150619/nsarguments-230)
As far as I followed sensors
the temperature has always not been higher than 88°C, with 105°C being the limit for high and critical.
Your compiler is dying while trying to compile webkit I think, not sure if due to memory or code, but I would recommend trying gcc 4.8.5. A bit never … but not too new
vm: 6GB of mem, 8GB swap
host: 8GB mem, 8.5GB swap
(I hope that is enough)
I’ll try updating my gcc next weekend.
When I try to sync the repo, it only fetches 3 or 4 of the 515 projects… happens both on ssh and html connection.
Does anyone have any idea why that would be?
Nevermind, now it seems like it is downloading properly. I only had to wait much longer than I expected…
make completed successfully (03:14:09 (hh:mm:ss))
Hurray! For those of you who have problems, here is what I did:
- Took an old harddrive, wiped it and installed Ubuntu 14.04 LTS.
- Accepted automated updates that Ubuntu wanted to do
- Followed the official guide step by step.
Here is the output of gcc --version, in case it helps anyone:
gcc (Ubuntu 4.8.4-2ubuntu1~14.04) 4.8.4
I will not try flashing my phone until there is a way to install openGAPPS. I’m afraid I just cannot live without the Play Store. If anyone can help, please check out this thread:
As @teh_flame just wrote that he successfully compiled using gcc 4.8.4 I gave it another try. Unfortunately it failed again (building chromium it seems), see build log (only 1.4MB of text compared to 2.3MB of log last try).
After a hint in the IRC channel that memory might be a problem I first of all increased swap of the vm (now: 6.5GB memory + 12GB swap) and tried again. Unfortunately still failing (build log).
You can install openGAPPS, just try it
Hi @sjjh,
that’s weird. I’m compiling with Ubuntu 14.04 Server (default gcc) in a KVM-VM which works perfectly, altough I dedicated 10GB of RAM to the VM.
Have you tried starting over? Installing a new VM with Ubuntu 14.04 Server, updating the machine, installing the packages, redownload (or copy) the repository. If that doesn’t help, maybe reduce the number of parallel tasks (make -j2 instead of make -j8).
Thanks @jnsp!
My physical machine only has 8GB of RAM, thus I’ll need to compensate with swap (I hope that doesn’t make any difference – except for speed)
I always tried with j-1
only anyway.
I can set-up another vm to try on a different system. But to be honest I would prefer to understand what went wrong and fix it, if I knew how to debug. Nevertheless I’ll start over with a new vm to see how it goes.
Before switching to docker I used a VM with 4G ram and 4G swap, which seems to be the minimum
Now I’m using a docker container on my host with 8G ram
Over here it fails with 6.5GB ram and 12GB swap. So there must be a different reason…
Only things that still come to my mind: Broken memory or … wrong gcc/software setup. I know other people build with it with your version here, but I’m just trying to explain the error message I saw that you posted here. But as far as I know, you never tried another gcc Or tried to reinstall the old one.
Regarding “different reason” I was referring to the amount of memory/swap available.
But I’m currently setting up a new vm (ubuntu server 14.04 lts), so let’s see if it works there finally.
Okay, that would also be a new software setup. If all fails, try running memcheck on the system. ;(
The lastest pull (“sync”) from today still builds here, I have only seen a few old January changes from an external party that seems to fix FP issues upstream (Thunder Software Technology Co., Ltd). No problems. Not sure if there is an other -dev tree hidden somewhere. I use what is described in the HOWTO above.
Yeah, thought it couldn’t hurt to start from scratch… Hope the physical memory is okay – the machine is not too old. :-/[quote=“fp1_wo_sw_updates, post:393, topic:11600”]
The lastest pull (“sync”) from today still builds here,
[/quote]
Good to know. Currently fetching it. I’ll see how it goes…
And it failed again (at building chromium), see buildlog. Repeating specs: vm, ubuntu 14.04 lts server, gcc 4.8.4, 6GB memory, 6GB swap.
@fp1_wo_sw_updates, I’ll run memcheck
but as I never had any freezes or the like yet I believe that it’s not a hardware problem with memory on the host.
I couldn’t find it, as the thread is quite long already, what kind of virtualization do you use?
I might never have named it… :-/ I’m using kvm (via virt-manager/libvirt).