Re: KI_FILE count precondition

On Thu, 18 Jan 2007 18:45:59 +0100, Eric Bezault <ericb-D6Qt/9opevxWk0Htik3J/[email protected]>  
wrote:

> If this code is in Gobo then it needs to be fixed. Indeed there
> is no guarantee that `read_string' will read `filesize' characters.
> It could well be implemented to read only one character even if
> the file is much bigger. So we need a loop until the end of file
> is reached. Therefore the value of `filesize' that you provide is
> arbitrary. Even more under Windows with the %R%N end-of-line.

I understand and will fix the location ASAP, definitely before the  
release. The original code was not written by me, but obviously I didn't  
see the problem. Perhaps the documentation string of 'read_string' could  
be updated to say clearly that read_string might read less than specified.  
Obviously, I has to read at least one character.

>> It would be best if there was a 'read_string(-1)' that would just read  
>> in
>> everything available.
>
> This -1 is very ugly. This might be common practice in
> other languages, but in Eiffel I would prefer to have a new
> feature `read_all' or `read_to_end' or something like that.
> Then I wonder whether this routine should write to `last_string'
> or to a string given as argument (like `read_to_string').

True, I think such a functionality would be very helpful, as many people  
have probably the same problem and a common solution with a reasonable  
buffer size would prevent common mistakes.

Bernd