Re: GNAT: Performance of String functions

["Tarjei T. Jensen" <tarjei@online.no>]

Robert Dewar wrote:
> 
> Tarjei said
> 
> <<VAX Pascal passes strings by descriptors. The string pointed to by
> descriptor is of type counted string. A counted string has a maximum
> size, an actual size and the string itself. I'm not sure anymore if the
> descriptor contains the maximum size of the string or if the programmer
> have to keep track of this.
> 
> This means that comparing Ada bounded strings and VAX Pascal strings
> compares two different ways of doing things. Each time you change the
> size of an Ada bounded string there can be serious overhead.
> >>
> 
> So, clearly Vax Pascal strings are to be compared with bounded strings,
> not with unbounded strings, and presumably there is no automatic
> finalization involved.
> 
> I am not sure what your "serious" overhead message is about, makes no
> sense to me, but perhaps it is just not clear.

I wrote bounded string when I meant unbounded string. Silly me.

As unbounded strings are allocated on the heap there is a potential
serious overhead in allocating and deallocating space for these strings
as they change size. The serious overhead can include system calls to
allocate or free space. I expect that it also include wholesale copying
of the contents of the strings.

I have not checked the GNAT source for how unbounded strings are
handled.

Greetings,


-- 
// Tarjei T. Jensen 
//    tarjei@online.no || voice +47 51 62 85 58
//   Support you local rescue centre: GET LOST!