[PC-BSD Dev] [Fwd: Re: virtualbox com object bandaid & qt4 problem?]

Joe Maloney jpm820 at gmail.com
Thu Dec 26 12:04:59 PST 2013


It was not fixed.  While it does work in stock FreeBSD 10 if you do not use KDE it does not work whatsoever in PCBSD.  So I’ve been continuing the discussion over at the thread below.  

http://lists.freebsd.org/pipermail/freebsd-emulation/2013-December/011171.html

Here it is pointed out to me that there is a experimental port which uses 95 % clang.  It does indeed fix the segfault issue.  

https://redports.org/browser/virtualbox/emulators

However in the thread he references the port maintainer explains that this idea should not be applied to his port.  He believes that the issues with GCC should be resolved instead for him to be able to continue to commit patches upstream with oracle.  That kind of makes sense I think.  Though I’m on the fence as I’d rather have everything compile with CLANG.

So I’m not sure if this will get fixed in time for 10-RELEASE.  I’m not sure how critical the virtualbox gui is to PCBSD as a feature but there are a few other options if you don’t want to wait that long.  

1.) change the port in your fork to use mostly clang
2.) use the experimental 4.3.4 port from Jung-uk Kim at redports instead of the maintained version by decke

I can assist with either of those items if needed.  I’d probably recommend number 1 over number 2 although number 2 would require less effort.  Otherwise I’m personally fine with just waiting it out if everyone else is as I can just build the port myself to fix.

Joe Maloney


On Dec 17, 2013, at 3:32 PM, Joe Maloney <jpm820 at gmail.com> wrote:

> FYI the segfault issue should be fixed as well now.
> 
> http://lists.freebsd.org/pipermail/freebsd-emulation/2013-December/011156.html
> 
> Joe Maloney
> 
> -------- Forwarded Message --------
> From: Joe Maloney <jpm820 at gmail.com>
> To: dev at lists.pcbsd.org
> Subject: Re: virtualbox com object bandaid & qt4 problem?
> Date: Sat, 14 Dec 2013 12:52:34 -0600
> 
> The good news is that the virtualbox com errors have been fixed in the new port.  I’m still running into segfault issues when running the qt4 app even with a new world and kernel.  I’ve submitted some backtrace info to the freebsd-emulation mailing list starting here.
> 
> http://lists.freebsd.org/pipermail/freebsd-emulation/2013-December/011126.html
> 
> I’ve also been informed by a FreeBSD user in the freebsd-vbox channel running rc1 that they did not have this problem.  If it still doesn’t work in the next ISO I will do some testing with stock freebsd to see if I can get it to work that way and figure out what the differences are if it does.  Not really a critical issue for me as I don’t really use vbox anymore.  Just wanted to try to help get all the little things fixed.
> 
> Joe Maloney
> 
> 
> On Dec 8, 2013, at 1:54 PM, Joe Maloney <jpm820 at gmail.com> wrote:
> 
>> So I've been looking in the virtualbox cannot create com object error.
>> Creating a VirtualBox.xml under /root/./VirtuaBox will fix this.  It
>> seems that the executable is no longer creating this file as it should.
>> There was a PR created for this one that looks like it hasn't been
>> resolved yet.
>> 
>> http://www.freebsd.org/cgi/query-pr.cgi?pr=183085
>> 
>> Creating the file is enough to get VBoxManage to work from command line
>> and phpvirtualbox to function.  The next problem seems to be the QT
>> application segfaulting.  I've noticed some others mentioned having to
>> unload env variables such as LANG, some QT4 variables with the
>> experimental port.  However I've tried all this in a plain xsession and
>> verified that no variables were loading using the env command and no
>> dice.  
>> 
>> I know normally I could debug gtk applications with gdb.  Not sure if
>> there may be something similar for qt4.  Really feels like maybe
>> something isn't right for launching the qt4 app though as it still
>> segfaults there.
>> 
>> I'm going to give the experimental ports a try next though I don't have
>> much hope that will fix.  Although some have reported the experimental
>> port works after unloading variables.  In which case that will be
>> annoying to have to unload things like LANG and not sure what other
>> effects that might have.  Just thought I would pass on what I've found
>> so far in case it might help solve.
>> 
>> Joe Maloney
>> 
>> 
>> 
>> 
> 
> 
> 



More information about the Dev mailing list