[mythtv-users] backend crash - fatal flex scanner internal error

[Michael T. Dean]

On 10/16/06 06:48, Chris Rouch wrote:

>On 10/13/06, Michael T. Dean <mtdean at thirdcontact.com> wrote:
>  
>
>>On 10/13/06 03:22, Chris Rouch wrote:
>>
>>    
>>
>>>Mythbackend crashed again yesterday. the last message in the logs is
>>>
>>>fatal flex scanner internal error--end of buffer missed
>>>
>>>
>>>      
>>>
>>Broken toolchain--not Myth's fault.  (And something that would make me
>>question the validity of every app on the system.)
>>
>>    
>>
>>>This happened while no-one was using it, and about 30 minutes before
>>>anything was scheduled to record. This is on FC4 with mythtv from
>>>atrpms (0.20-139, based on SVN r11293).
>>>
>>>Does anyone know if this has been fixed recently either in atrpms
>>>(latest is based on SVN r11499) or more recently in SVN, or should I
>>>raise a bug?
>>>
>>>      
>>>
>>That might be hard to figure out.  After all, there's nothing in that
>>error message that would be useful for a search of the mailing list archives
>>
>>http://www.gossamer-threads.com/lists/engine?list=mythtv&do=search_results&search_forum=all&search_string=flex+scanner&search_type=AND
>>
>>So, you'd never find that thread on -dev.  ;)
>>    
>>
>Oops. My bad.
>
>But having now looked at the list archives, I'm confused by the OP's
>solution: "I resolved the fatal flex scan error by: - upgrade gcc,
>glibc and all the binutils". I'm running a fully updated FC4 system,
>so I expect this solutions is not possible.
>

It doesn't mean you have a bad /version/ of some program in your 
toolchain.  It means you have a broken toolchain (i.e. the /files/ in 
your toolchain are broken/your install is broken/your system is 
corrupted/whatever).  This is why I wouldn't trust the system.

Mike