Hi Nate,<div><br><div class="gmail_quote">On Thu, Aug 4, 2011 at 12:55 AM, Nathan Ridge <span dir="ltr">&lt;<a href="mailto:zeratul976@hotmail.com">zeratul976@hotmail.com</a>&gt;</span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<div><div></div><div class="h5">Why do the tokenizer and the iterator need to be in the base class at all?</div></div>
 Why not just construct them on the stack in the derived constructor?<br></blockquote><div><br></div><div>Basically, the derived class instances have very short life - all objects are created on stack. And, all the derived class need to parse tokens and have to create a Tokenizer, character separator, etc., anyway. Putting Boost.Tokenizer and character separator in Base class free me (and others) from having to create Tokenizer in the derived classes (there are too many) and I can just repeatedly use the nextToken() that I have defined in the base class to get all the tokens. Anyway, my problem is solved now - iterators become invalid after execution of the derived class constructors. Basically, I don&#39;t want anyone to write monolithic functions that do everything in the world. So, I don&#39;t want them to mix business domain logic with parsing code and vice versa and, in handling callbacks, I don&#39;t want them to parse. Just a small measure for maintainability. Since Boost.Tokenizer and character_separator will be created anyway and since derived classes have very short life and all derived classes&#39; objects get created on stack, I think there is no harm in putting the Tokenizer &amp; character separator in the base class - I am not separately storing the string-to-be-passed in the base class though as that apparently gets stored by the Tokenizer itself.</div>
<div><br></div><div>Thanks &amp; regards, Asif</div><div><br></div></div></div>