RE: [Dev-C++] 'New Source File' bug

[Ioannis V. <no...@ya...>]

> -----Original Message-----
> From: dev...@li... 
> [mailto:dev...@li...] On Behalf 
> Of Devin Carless
> Sent: Sunday, February 17, 2002 6:47 PM
> To: dev...@li...
> Subject: Re: [Dev-C++] 'New Source File' bug
> 
> 
>   Oh, I know.  They're wonderful-looking things.  I'm 
> currently only using them as fancy dynamic arrays.
>   But while you can do a lot with vectors, vectors can do a 
> lot with you.  I had Vec.size() in a conditional, and the 
> conditional consistently got it wrong.  It worked, however, 
> when I replaced
> Vec.size() with a variable, Size, which had been intialized to
> Vec.size() a line before.
>   That's only one reason I'm very cautious with vectors.


This is broken compiler. Did it happen with MINGW32?

Although i did not see any code, value wrapping may had occured.

E.g. 

  for(int i=0; i<Vec.size(); i++)
  { ...}

  When variable i reached the maximum int value it may wrapped around
(according to the standard it is undefined behavior).

The correct & safe synatx is assuming Vec to be vector<int>:

for(vector<int>::size_type i=0; i<Vec.size(); i++)
{ ...}


     

> > Anyway i think it is our lack of knowledge, not vectors. 
> When i finish 
> > the 16th chapter i will resend an email telling you how you can to 
> > assignments using vector constructors.
> 
>   There's a saying that goes 'It's a poor workman who blames 
> his tools.' And many times it's been my mistake; I'm a poor 
> workman.  But there are other times when I've caught 
> something funny going on.  (push_back() refusing to work was 
> one of them.)

This is broken implementation issue.


> > I am confident that standard containers are better than built in 
> > arrays.
> 
>   Absolutely true: the reason I'm using vectors is because 
> arrays are /worse./  The only thing I think arrays have to 
> recommend them is that they can be intialized like so:
> 
>     char* frogs[4] = { "Green", "Tree", "Bull", "Muppet" };
> 
>   (Yes, I know char* is frowned upon, but I've found some 
> interesting traps trying to initialize a string to a string 
> constant of the form "Blah" when passed as an argument.)
>   For my project, in which certain subclasses have great 
> tracts of constant data, this kind of initialization would be ideal.


I am working using standard containers all the time and never
encountered any problem (except that MING doesn't support vector::at()
).



Ioannis
 
* Ioannis Vranos
* Programming pages: http://www.noicys.d2g.com
* Alternative URL: http://run.to/noicys