Help! "for (;;)"?

[Perro Flaco]

Hi everybody!

I've got a very simple, but very difficult to me, question: can anybody
explain me this?

for (;;)
{
(...)
}
I just don't get it. Inside the loop, I've got a couple of "break".

Thank you for your help!

[mlimber]

Perro Flaco wrote:

Hi everybody!

I've got a very simple, but very difficult to me, question: can anybody
explain me this?

for (;;)
{
(...)
}
I just don't get it. Inside the loop, I've got a couple of "break".

Thank you for your help!

It's an infinite loop (same as "while(true) { /*...*/ }"). The "break"s
are the only way to get out of it.

Cheers! --M

[Sumit Rajan]

"Perro Flaco" <fa****************@gmail.com> wrote in message
news:11*********************@g14g2000cwa.googlegro ups.com...

Hi everybody!

I've got a very simple, but very difficult to me, question: can anybody
explain me this?

for (;;)
The above line starts the loop. The is no exit condition so the loop will go
on forever -- unless you "break" out of it.
{
(...)
}

Regards,
Sumit.
--
Sumit Rajan <su****@msdc.hcltech.com>

[Perro Flaco]

In only five minutes I've got two answers to my question. This is
great! :)

Thank you very much!
¡Muchas gracias!

[Kwan Lim]

This is called a mid-test loop or multi-exit loop. It can be used to
avoid a common software engineering problem: duplicate code.
For example,

read(input, d)
while (!eof(input)) do
do something with d;
read(input, d);
end

- here, read(input, d) is duplicated.

vs.

while(true) do
read(input, d);
if(eof(input)) break;
do something with d;
end

- here there's no code duplication

[Perro Flaco]

Your example is very welcome.

Again, thanks!

[Ferdi Smit]

mlimber wrote:

It's an infinite loop (same as "while(true) { /*...*/ }"). The "break"s
are the only way to get out of it.

Well, not the "only" way. You could issue a 'return' (heck even a
'goto'...yikes) or more importantly an exception might be thrown
implicitly by other code in the loop; perhaps in a completely different
source module. Always consider exceptions and what they might do.
--
Regards,

Ferdi Smit (M.Sc.)
Email: Fe********@cwi.nl
Room: C0.07 Phone: 4229
INS3 Visualization and 3D Interfaces
CWI Amsterdam, The Netherlands

[deane_gavin]

Kwan Lim wrote:

This is called a mid-test loop or multi-exit loop. It can be used to
avoid a common software engineering problem: duplicate code.
For example,

read(input, d)
while (!eof(input)) do
do something with d;
read(input, d);
end

- here, read(input, d) is duplicated.

vs.

while(true) do
read(input, d);
if(eof(input)) break;
do something with d;
end

- here there's no code duplication

I am changing the subject here so not detracting from your point at all
....

I think it's worth mentioning that with C++ streams that pseudo-code
should look like this (note: still no code duplication)

while (read(input, d))
do something with d
end
if (!eof(input)) ... something went wrong reading from the stream

I want to make that point because there is a common misunderstanding
about how to use eof with streams.

Gavin Deane

[deane_gavin]

Ferdi Smit wrote:

mlimber wrote:

It's an infinite loop (same as "while(true) { /*...*/ }"). The "break"s
are the only way to get out of it.

Well, not the "only" way. You could issue a 'return' (heck even a
'goto'...yikes) or more importantly an exception might be thrown
implicitly by other code in the loop; perhaps in a completely different
source module. Always consider exceptions and what they might do.

Always write exception safe code so you never have to worry about what
exceptions might do :-)

Gavin Deane

[Julián Albo]

Kwan Lim wrote:

This is called a mid-test loop or multi-exit loop. It can be used to
avoid a common software engineering problem: duplicate code.
For example,

read(input, d)
while (!eof(input)) do
do something with d;
read(input, d);
end

- here, read(input, d) is duplicated.

vs.

while(true) do
read(input, d);
if(eof(input)) break;
do something with d;
end

- here there's no code duplication

I think a better solution in this case and many similar others is something
like that (pseudocode):

bool read_eof (input, d)
{
read (input, d);
return ! eof (input);
}

while (read_eof (input, d) )
do_something;

--
Salu2

[Karl Heinz Buchegger]

Julián Albo wrote:

while(true) do
read(input, d);
if(eof(input)) break;
do something with d;
end

- here there's no code duplication

I think a better solution in this case and many similar others is something
like that (pseudocode):

bool read_eof (input, d)
{
read (input, d);
return ! eof (input);
}

while (read_eof (input, d) )
do_something;

Actually this is not a better solution in C++.
The main problem is, that you still try to use eof()
for controlling the loop.

There is nothing wrong with what Deane posted
and it is the way to go:

while( read( input, d ) )
do_something;

if( !eof( input ) )
something_went_wrong;

Note that this loop terminates even if some other thing
then eof() happens. Things like: user removed device (floppy, cd,
memory stick, ...) during the read operation, the network connection
broke down, the modem had a transmission failure, phone line broke
or simply a bad block on the hard disc.

--
Karl Heinz Buchegger
kb******@gascad.at