guile-devel
[Top][All Lists]
Advanced

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

Re: Another alternative string representation proposal


From: Dirk Herrmann
Subject: Re: Another alternative string representation proposal
Date: Tue, 26 Sep 2000 11:20:03 +0200 (MEST)

address@hidden (Carl R. Witty) writes:

> > Is owner_p really necessary?  If a char-field is read-only, it is not
> > the owner of the character sequence.  If it is mutable or immutable,
> > we can force the char-field to own the sequence.
> 
> I can see at least one potential use for a mutable, not-owned string:
> it could be used by a guile wrapper for mmap.  Although probably it
> would be better to use a different, specialized type for this; so
> that's not much of an argument...

I'd like to give another example for a case, where mutable, not-owned strings
make sense.  Assume the following function f, which takes an already existing
string, but initializes or overwrites it with something.


void
f (SCM aString, unsigned int n)
{
  /* Fill the first n characters of aString with something. */
}


A possible use of function f might be to initialize a temporary string:


void
g ()
{
  char tmp_chars[4];
  SCM tmp_string = scm_make_mutable_not_owned_string (tmp_chars, 4);

  gh_display (f (tmp_string, 4));
}


It's obvious that this gives much better performance than having to allocate
a small memory region just to have a mutable string.

Still, I don't want to claim that we really need mutable strings, but we
should be aware of potential uses before we come to a decision about them.

Best regards
Dirk




reply via email to

[Prev in Thread] Current Thread [Next in Thread]