Re: [cp-patches] FYI: Replaced use of JvAssert by assert and removed java-assert.h

[Dalibor Topic]

Tom Tromey wrote:
>>>>>>"Dalibor" == Dalibor Topic <address@hidden> writes:
> 
> 
>>>`(1 << k) < 128' failed.
> 
> 
> mprec overflow...
> 
> 
>>>#0  0x00002aaaab028df0 in raise () from /lib/libc.so.6
> 
> 
> ... 64 bit machine ...
> 
> Dalibor> JvAssert only kicked in if DEBUG was defined, which assert always
> Dalibor> kicks in unless NDEBUG is defined. Other than that, they do the same
> Dalibor> checks, so it may be a real bug in mprec.c. Could you investigate on
> Dalibor> your arch/os?
> 
> The buffer size workaround I applied to mprec isn't completely safe --
> it just works in more cases.  This is why we need the newer
> dynamically-allocating mprec.
> 

Could you elaborate some more? I am not very familar with mprec's bugs,
yet. :)

cheers,
dalibor topic