Re: iterator resolution



On Sat, 4 Mar 2000, Karl Nelson wrote:

> 
> This is largely for future referenece after the bulk of
> the code is working and you want to think about optimizations.
> 
> It looks like you have lots of areas which you can shave 
> places of the size.  For example, in FrooTkxtSegment you
> most likly will not need char_counts and byte counts more
> that 64k.  If it is larger just split the segment in the 
> middle.  This will save at least 10% on text with very small 
> segements while costing less than 0.01% when you need to do a split.
> 
> Similar goes for the iterator type.  Also I would recommend just
> switching to  "gpointer dumb[size]" for the iterator definition
> rather than try to match it up.  All that matter is the size of
> the declared iterator is large than the real iterator.  
> (Also easily verified with a #if statement)

why bother at all?
what's wrong about

struct _*Iter
{
  /*< private >*/
  guint cache_count;
  gpointer buffer;
  [...]
};

in the public header file?

---
ciaoTJ



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]