[Buildroot] Buildroot-2.18.11 fails?

[scsijon]

On 13/02/19 06:19, Peter Seiderer wrote:
> Hello scsijon,
> 
> On Mon, 11 Feb 2019 12:56:32 +1100, scsijon <scsijon at lamiaworks.com.au> wrote:
> 
>> no, same result and point of failure, i'm going to move it across to one
>> of my test machines with a later Puppy called Easy and see what happens,
>> I can containerize it on that.
> 
> Thanks for feedback (and please keep the buildroot mailing list in CC)...
> 

Hi, thought I  had, sorry must have hit the wrong button.

No, although Easy, both in the container and outside it build, it's got 
gcc6.3.0, i'm now failing at host-e2fsprogs-1.44.4 with some parts 
getting built ok and a number of internal makefile not being correctly 
modified?built in the first place with wrong data at the MKDIR_P line in 
only some, causing it to stop.

I shall leave it for now as i'm out of free time, but will come back to 
it in a week or two. I will admit, i'm more interested in why only some 
of the Makefiles are right/wrong than anything else though. I shall open 
a raw copy of the package and see if it needs a patch.

regards to all

>>
>> I'll let you know what happens, however the errors do make me thing
>> something is wrong with the source package.
> 
> ...but builds on many different machined, see e.g. the
> autobuilder results...., at least resort try to downgrade the
> selected compiler...
> 
> Regards,
> Peter
> 
>>
>>
>> -------- Forwarded Message --------
>> Subject: Re: [Buildroot] Buildroot-2.18.11 fails?
>> Date: Mon, 11 Feb 2019 09:20:51 +1100
>> From: scsijon <scsijon at lamiaworks.com.au>
>> To: Peter Seiderer <ps.report at gmx.net>
>>
>> Hi, mine is a Puppy Linux variety called Quirky Xerus64, a cross of
>> Puppy linux built with ubuntu-xenial(, the normal way we build things in
>> the puppy world), and yes it uses gcc 5.4.0, libc-2.23 and libc6-2.23.
>>
>> But what's this about "make savedefconfig", it's not in the listed build
>> steps in the docs or quickbuild steps in the readme? Maybe that's my
>> problem.
>>
>> On 11/02/19 02:09, Peter Seiderer wrote:
>>> Hello scsijon,
>>>
>>> On Sun, 10 Feb 2019 19:37:07 +1100, scsijon <scsijon at lamiaworks.com.au> wrote:
>>>    
>>>> On 10/02/19 18:14, Peter Seiderer wrote:
>>>>> Hello scsijon,
>>>>>
>>>>> On Sun, 10 Feb 2019 15:20:45 +1100, scsijon <scsijon at lamiaworks.com.au> wrote:
>>>>>       
>>>>>> I thought i'd try it after so many years.
>>>>>> I thought i'd try the last complete version before trying a snapshot.
>>>>>>
>>>>>> So, what I did was to download and unpacked the 2018.11 package from
>>>>>> your site.
>>>>>
>>>>> 2018.11.2, 2018.11.1 or 2018.11?
>>>>>       
>>>> buildroot-2018.11.tar.gz, didn't see the others on the site I was
>>>> directed too from your main page at the time. However found 2018.11.2 on
>>>> the git site so will have a go with that tonight.
>>>>>>
>>>>>> Went into make menuconfig and changed (to match what I am use to working
>>>>>> with)>
>>>>>> target Architecture to x86_64;
>>>>>> toolchain to glibc;
>>>>>> gcc version to 8.x;
>>>>>> linux kernel to 4.18 and added a defcon name;
>>>>>> exited.
>>>>>>
>>>>>> make source (to download the packages required so I could build offline);
>>>>>>
>>>>>> make.
>>>>>>
>>>>>> All went well until host-gcc-final-8.2.0 where it errored, failed and
>>>>>> exited.
>>>>>
>>>>> Should not happen ;-), mind to share your .config/defconfig file?
>>>>>       
>>>> Hope this is what you want, attached, couldn't send a . file so renamed
>>>> it, ignore the undotted.
>>>
>>> Thanks for the .config file, after 'make savedefconfig' resulted in the follwing (short)
>>> defconfig file:
>>>
>>> BR2_x86_64=y
>>> BR2_JLEVEL=1
>>> BR2_TOOLCHAIN_BUILDROOT_GLIBC=y
>>> BR2_PACKAGE_HOST_LINUX_HEADERS_CUSTOM_4_18=y
>>> BR2_GCC_VERSION_8_X=y
>>> BR2_LINUX_KERNEL=y
>>> BR2_LINUX_KERNEL_DEFCONFIG="dfgh"
>>>
>>> Could not reproduce your problem here locally, on which host system do you
>>> run your build (mine is openSUSE Tumbleweed with gcc-8.2.1 and libc-2.27,
>>> yours seem to be a gcc-5.x)?
>>>
>>> Regards,
>>> Peter
>>>    
>>>>   
>>>>> Regards,
>>>>> Peter
>>>>>       
>>>>>>
>>>>>> Before I consider proceeding further, could I have some help and advice
>>>>>> with this please. I would have thought this should have been ok.
>>>>>>
>>>>>> I have attached the relevant section of the log file.
>>>>>>
>>>>>> thanks
>>>>>> scsijon
>>>>>>
>>>>>>      
>>>>>
>>>>>       
>>>    
> 
>