Re: uninitialized memory

[Joe Neeman]

On Monday 21 May 2007 05:41, Han-Wen Nienhuys wrote:
> Hi Joe,
>
> in
> Page_breaking::system_count_bounds (vector<Break_position> const &chunks,
>                                   bool min)
>
> a vector is resized and then indexed.  Can you document why this will never
> result in unitialized memory being passed back to the caller? I'd this
> might happen if all[sys].pscore_ is null for some systems.

In the resize call, all the elements are initialised to 1. So if 
all_[sys].pscore_ is null, it will remain as 1 (as it should, since if 
pscore_ is null then that system_spec is a piece of markup). I'll add a 
comment to this effect.