Pointer-to-pointer realloc problem

[Herhor]

Hello marvinla!

See how these old drags got carried away on my advise. It is quite
clear they are _too backward persons_ to be able to think progressively.
They suggested you to use almost forty years old malloc/free shit if
there is _much simpler and secure_ new/delete alternative. It isn't done
without reason because otherwise these buzzards would have to *close*
their antiquated newsgroup and move themselves into some *software
museum* like this at: http://dn.codegear.com/museum.

So blow off them and do as I advocated you!

 

[Army1987]

*(p+1)+1 is the same as *p+2 which is the same as *(p+2), so
*(*(p+1)+1) is the same as **(p+2). I suspect you wanted *((*(p+1))+1)

Huh?
*(p + 1) + 1 is p[1] + 1, *p + 2 is p[0] + 2, and *(p + 2) is p[2].
*(*(p+1)+1) is p[1][1], and **(p + 2) is p[2][0].

 

[Richard Heathfield]

Herhor said:

Hello marvinla!

See how these old drags got carried away on my advise.

Nope. They just pointed out that you're wrong, which you are.

It is quite
clear they are _too backward persons_ to be able to think
progressively.

If "think progressively" == "think wrongly", then you might have a
point, but it doesn't, and you don't.

You are, quite simply, incorrect. Learn how to learn from corrections.

 

[mark_bluemel]

[Something that probably translates as "I is 133t!"]

I'm currently classifying this guy as a troll unless/until I see
evidence to the contrary.

 

[marvinla]

Hello marvinla!

See how these old drags got carried away on my advise. It is quite
clear they are _too backward persons_ to be able to think progressively.
They suggested you to use almost forty years old malloc/free shit if
there is _much simpler and secure_ new/delete alternative. It isn't done
without reason because otherwise these buzzards would have to *close*
their antiquated newsgroup and move themselves into some *software
museum* like this at:http://dn.codegear.com/museum.

So blow off them and do as I advocated you!

Herhor,
I'm writing my code in C, not in C++. I know about the new and delete
in C++, and
IF I was using C++, of course I would use them.

See you, and thank you guys for the great explanations!

 

[CBFalconer]

.... snip ...

In your suggested pattern, the name 'attemptcorrection' is rather
misleading because, in those cases when actual correction is
possible (rather than, say, an emergency data save), you can't
reasonably do it after malloc. You need a malloc replacement:[1]

No you don't. For example, the program may simply announce that
memory is full, therefore it is not accepting more input, and
proceed to process what it has.

 

[Default User]

Hello marvinla!

See how these old drags got carried away on my advise.

Plonk the troll



Brian

 

[Ben Bacarisse]

... snip ...

In your suggested pattern, the name 'attemptcorrection' is rather
misleading because, in those cases when actual correction is
possible (rather than, say, an emergency data save), you can't
reasonably do it after malloc. You need a malloc replacement:[1]

No you don't. For example, the program may simply announce that
memory is full, therefore it is not accepting more input, and
proceed to process what it has.

That is not what I'd call 'attemptcorrection'. But I agree, if you
*don't* want to attempt any significant correction, then your outline
suits the situation perfectly.

 

[CBFalconer]

... snip ...

In your suggested pattern, the name 'attemptcorrection' is rather
misleading because, in those cases when actual correction is
possible (rather than, say, an emergency data save), you can't
reasonably do it after malloc. You need a malloc replacement:[1]

No you don't. For example, the program may simply announce that
memory is full, therefore it is not accepting more input, and
proceed to process what it has.

That is not what I'd call 'attemptcorrection'. But I agree, if
you *don't* want to attempt any significant correction, then your
outline suits the situation perfectly.

That IS a significant correction. It changes the objective of the
program according to resources available. I use it (or the
equivalent) in many situations. You can see one in my hashlib
package.

 

[David Thompson]

Herhor said:

Fine, given appropriate definitions of new and delete, and provided
delete isn't 0, of course. Otherwise, the behaviour is undefined.

Also if new==INT_MIN /* or LONG etc */ && delete==-1
on most (2sC) systems. <G>
- formerly david.thompson1 || achar(64) || worldnet.att.net