Re: Fedora issues

[Andrew Bernard]

Dear Phil,

Thanks. I’m aware that without passing you the score this will possibly never 
get fixed. My issue occurs on Fedora 21, 22 and Mint 17, which is Ubuntu based. 
At this point, I am going to try a binary chop debugging approach. I had hoped 
the previous posting of the failed assertion message may be of some help, but 
it is possibly too uncertain as to how the code gets there.

lilypond: 
/home/gub/NewGub/gub/target/linux-64/src/lilypond-git.sv.gnu.org--lilypond.git-release-unstable/flower/include/interval.hh:226:
 T Interval_t<T>::center() const [with T = double]: Assertion `!is_empty ()' 
failed.

Fedora seems to have a bunch of issues at the moment. Is there any way that the 
development team could try a build on Fedora?

I understand my issue and the gs issue are separate. Sadly, this combination 
kills lilypond totally for me, or at least restricts me to never being able to 
go further than 2.19.20. I tried using gdb on the program but the output 
showing the crash stack frame is overwhelming and impossible to understand 
without being an expert on the internals.

Is lilypond restricted to 32 bit?

Andrew



On 31 May 2015 at 01:59:30, Phil Holmes (address@hidden) wrote:

Please remember the "reply all"
 
I think the problem you have with your complex score and the simple score 
failure are not the same?  I believe the simple issue is to do with GhostScript 
and how it interacts with the installed version in Fedora, although I am not an 
expert in this area.  I also believe Masamichi Hosoda is looking at what could 
be causing it.
 
Your more complex score might fail on any or no other OSes: without a sample 
that causes the crash, it is never likely to be able to be fixed.
 
The binaries for LilyPond are built on my 32 bit Ubuntu 14.04 virtual machine.  
They are tested by any of the community that uses LilyPond on whatever 
operating system they have installed.