POV-Ray : Newsgroups : povray.unix : Unix debugging expertise needed Server Time: 14 Dec 2018 10:59:27 GMT
  Unix debugging expertise needed (Message 1 to 7 of 7)  
From: clipka
Subject: Unix debugging expertise needed
Date: 18 Nov 2016 13:44:24
Message: <582f05b8$1@news.povray.org>
Hi folks,

can someone with access to Ubuntu _12.04_ please have a look at this
POV-Ray branch (a pending pull request), see if the binaries crap out on
their machine as well (complaining about "pure virtual method called"
and some stuff about pthread, boost and mutex), and if so, figure out why:

https://github.com/eblanca/povray

Your help would be much appreciated.


Post a reply to this message

From: Le Forgeron
Subject: Re: Unix debugging expertise needed
Date: 19 Nov 2016 11:11:29
Message: <58303361$1@news.povray.org>
Le 18/11/2016 à 14:44, clipka a écrit :
> Hi folks,
> 
> can someone with access to Ubuntu _12.04_ please have a look at this
> POV-Ray branch (a pending pull request), see if the binaries crap out on
> their machine as well (complaining about "pure virtual method called"
> and some stuff about pthread, boost and mutex), and if so, figure out why:
> 
> https://github.com/eblanca/povray
> 
> Your help would be much appreciated.
> 

Can do, have to wait till Monday (guess where it is)


Post a reply to this message

From: Le Forgeron
Subject: Re: Unix debugging expertise needed
Date: 21 Nov 2016 09:02:05
Message: <5832b80d$1@news.povray.org>
Le 19/11/2016 à 12:11, Le_Forgeron a écrit :
> Le 18/11/2016 à 14:44, clipka a écrit :
>> Hi folks,
>>
>> can someone with access to Ubuntu _12.04_ please have a look at this
>> POV-Ray branch (a pending pull request), see if the binaries crap out on
>> their machine as well (complaining about "pure virtual method called"
>> and some stuff about pthread, boost and mutex), and if so, figure out why:
>>
>> https://github.com/eblanca/povray
>>
>> Your help would be much appreciated.
>>
>
> Can do, have to wait till Monday (guess where it is)
>

Sorry to disappoint you, but it's working, at least with -D option:
* ubuntu 12.04.5 LTS
* g++ 4.6(.3)
* 64 bits system (x86_64)



-D is due to my local setting (X11 remote display is not setup to allow 
this machine to display on my screen)

With +D (or default), I got a core dumped. (but it seems unrelated to 
that branch and might be a problem for all branches)

With DISPLAY not defined, it's strange: there is an ascii preview, then 
statistic and the telnet/shell is frozen. (yet, probably still the same 
for all branches)

unix/povray --benchmark : ok
unix/povray with small scene (3.7, 1 sphere) : ok


Post a reply to this message

From: jhu
Subject: Re: Unix debugging expertise needed
Date: 13 Dec 2016 06:35:00
Message: <web.584f9650b6f3f0b2194284cc0@news.povray.org>
clipka <ano### [at] anonymousorg> wrote:
> Hi folks,
>
> can someone with access to Ubuntu _12.04_ please have a look at this
> POV-Ray branch (a pending pull request), see if the binaries crap out on
> their machine as well (complaining about "pure virtual method called"
> and some stuff about pthread, boost and mutex), and if so, figure out why:
>
> https://github.com/eblanca/povray
>
> Your help would be much appreciated.

No access to Ubuntu 12.04, but it does run just fine on Debian/testing and
FreeBSD.


Post a reply to this message

From: clipka
Subject: Re: Unix debugging expertise needed
Date: 13 Dec 2016 06:58:08
Message: <584f9c00@news.povray.org>
Am 13.12.2016 um 07:33 schrieb jhu:
> clipka <ano### [at] anonymousorg> wrote:
>> Hi folks,
>>
>> can someone with access to Ubuntu _12.04_ please have a look at this
>> POV-Ray branch (a pending pull request), see if the binaries crap out on
>> their machine as well (complaining about "pure virtual method called"
>> and some stuff about pthread, boost and mutex), and if so, figure out why:
>>
>> https://github.com/eblanca/povray
>>
>> Your help would be much appreciated.
> 
> No access to Ubuntu 12.04, but it does run just fine on Debian/testing and
> FreeBSD.

That's of little help to me. I already know that this thing runs fine on
/some/ systems.

I need someone to get this baby to cry in a controlled environment, to
figure out why it does so in our Travis CI automated build test
environemnt, and ideally find a way to soothe it there.

It might just be afraid of the dark (read: having no terminal
associated, or some such), but at the moment that's just a wild guess.


Post a reply to this message

From: jhu
Subject: Re: Unix debugging expertise needed
Date: 14 Dec 2016 01:00:06
Message: <web.5850998db6f3f0b2615a0e20@news.povray.org>
clipka <ano### [at] anonymousorg> wrote:

>
> I need someone to get this baby to cry in a controlled environment, to
> figure out why it does so in our Travis CI automated build test
> environemnt, and ideally find a way to soothe it there.
>
> It might just be afraid of the dark (read: having no terminal
> associated, or some such), but at the moment that's just a wild guess.

Someone else said it works fine on their Ubuntu 12.04 system. Perhaps it's
specific your particular hardware configuration?


Post a reply to this message

From: Le Forgeron
Subject: Re: Unix debugging expertise needed
Date: 14 Dec 2016 08:12:36
Message: <5850fef4$1@news.povray.org>
Le 14/12/2016 à 01:59, jhu a écrit :
> clipka <ano### [at] anonymousorg> wrote:
>
>>
>> I need someone to get this baby to cry in a controlled environment, to
>> figure out why it does so in our Travis CI automated build test
>> environemnt, and ideally find a way to soothe it there.
>>
>> It might just be afraid of the dark (read: having no terminal
>> associated, or some such), but at the moment that's just a wild guess.
>
> Someone else said it works fine on their Ubuntu 12.04 system. Perhaps it's
> specific your particular hardware configuration?
>
>
pure virtual called is usually a problem of code vs compiler vs linker.
(child class declared the method, but did not implement it)

It can also be a problem of a fool pointer being interpreted as a 
pointer to a class that is not the expected one anymore. (such as to a 
deleted instance) and then, it can depend a lot on the actual system 
usage about what is found in the memory for that pointer. Memory 
problems can be due to race or serialization between cores.

I tried it on 64 bits ubuntu 12.04, with g++ 4.6.3, and 4 cores.

It could also be tied to the subversion of g++ (I cannot update that 
system, and maybe 12.04 did not stay at 4.6.3), or 32 vs 64.


Post a reply to this message

Copyright 2003-2008 Persistence of Vision Raytracer Pty. Ltd.