Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.java.help > #697

Re: Why would one use += 1 at the place of ++?

Show all headers | View raw

David Lamb wrote:
> Merciadri Luca wrote:
>> I've heard that it is generally better to use += 1 at the place of ++. For me,
>> they do exactly the same thing, maybe except that the former is
>> actually splitted into different atomic operations, when the latter
>> might be done in an atomic operation?

> I think ++ exists only because the hardware on which C was developed had an
> addressing mode that did it atomically in the very special case of referring to
> something = array[index++];
> I have no idea what goes on with the Java byte code specifically (ie whether
> ++ is defined to be atomic) but any decent optimizer ought to treat x++ and x
> +=1 each in whatever way is optimal.

Nope.

++ is not atomic.  This has nothing to do with being "optimal" or not.  Also, 
its semantics are not the same as += 1.  By definition, 'x += 1;' is identical 
in operation to 'x = (type) (x + 1);' except for the number of times the 
variable is evaluated.

"A compound assignment expression of the form E1 op= E2 is equivalent to E1 = 
(T)((E1) op (E2)), where T is the type of E1, except that E1 is evaluated only 
once."
- JLS, §15.26.2

The op= construction works on all types that support the 'op'.  ++ only works 
on numeric types.

The post- and pre-increment operators and compound assignment operators have 
different precedence.

There's no cast, widening or narrowing conversions or value set conversion in 
a pre- or post-increment operator expression.

And of course, the post-increment operators yield a different expression value 
from the pre-increment or compound assignment operators.

-- 
Lew
Honi soit qui mal y pense.
http://upload.wikimedia.org/wikipedia/commons/c/cf/Friz.jpg

Back to comp.lang.java.help | Previous | Next — Previous in thread | Next in thread | Find similar

Thread

Why would one use += 1 at the place of ++? Merciadri Luca <Luca.Merciadri@student.ulg.ac.be> - 2011-05-19 09:19 +0200
  Re: Why would one use += 1 at the place of ++? Bent C Dalager <bcd@pvv.ntnu.no> - 2011-05-19 10:22 +0000
  Re: Why would one use += 1 at the place of ++? David Lamb <dalamb@cs.queensu.ca> - 2011-05-19 06:40 -0400
    Re: Why would one use += 1 at the place of ++? Jeff Higgins <jeff@invalid.invalid> - 2011-05-19 07:58 -0400
    [OT] Re: Why would one use += 1 at the place of ++? Eric Sosman <esosman@ieee-dot-org.invalid> - 2011-05-19 08:03 -0400
      Re: [OT] Re: Why would one use += 1 at the place of ++? Patricia Shanahan <pats@acm.org> - 2011-05-19 07:35 -0700
        Re: [OT] Re: Why would one use += 1 at the place of ++? Eric Sosman <esosman@ieee-dot-org.invalid> - 2011-05-19 22:57 -0400
          Re: [OT] Re: Why would one use += 1 at the place of ++? rossum <rossum48@coldmail.com> - 2011-05-20 11:00 +0100
          Re: [OT] Re: Why would one use += 1 at the place of ++? Patricia Shanahan <pats@acm.org> - 2011-05-20 12:55 -0700
      Re: [OT] Re: Why would one use += 1 at the place of ++? David Lamb <dalamb@cs.queensu.ca> - 2011-05-19 17:58 -0400
        Re: [OT] Re: Why would one use += 1 at the place of ++? Eric Sosman <esosman@ieee-dot-org.invalid> - 2011-05-19 23:13 -0400
    Re: Why would one use += 1 at the place of ++? Lew <noone@lewscanon.com> - 2011-05-19 08:43 -0400
  Re: Why would one use += 1 at the place of ++? Eric Sosman <esosman@ieee-dot-org.invalid> - 2011-05-19 07:51 -0400
  Re: Why would one use += 1 at the place of ++? Jeff Higgins <jeff@invalid.invalid> - 2011-05-19 08:03 -0400
    Re: Why would one use += 1 at the place of ++? Lew <noone@lewscanon.com> - 2011-05-19 08:08 -0400
  Re: Why would one use += 1 at the place of ++? Roedy Green <see_website@mindprod.com.invalid> - 2011-05-19 10:23 -0700
    Re: Why would one use += 1 at the place of ++? Merciadri Luca <Luca.Merciadri@student.ulg.ac.be> - 2011-05-19 21:15 +0200
      Re: Why would one use += 1 at the place of ++? Lew <noone@lewscanon.com> - 2011-05-19 15:29 -0400

csiph-web