Re: confusing string error

[agent@drrob1.com]

On Sat, 15 Feb 2014 09:33:25 +0000, Simon Wright <simon@pushface.org>
wrote:

>agent@drrob1.com writes:
>
>> I was trying not to bomb the newsgroup.  But it seems you want it.  So
>> here we go.  Assume that a valid spec file exists.  I included the
>> relevent types I define in it in my comments, as it is in my actual
>> code.
>
>You are the one asking the questions, we are the ones who are trying to
>help. You do not make it easy; it just took me about half an hour to get
>the code you supply to the state where it compiles; and I still don't
>know what CAP is in
>
>   Char.CH := CAP(Element(TKNBUF,CURPOSN));
>
>so I just removed it (guessing it's supposed to convert to upper case).
>
>There is no problem if I enter "12345" or "abcde" or "1+2". However, if
>I enter "1+" or "1+2+" I get the same index error as Adam, on
>
>      CH2 := Element(Token.UStr,2);
>
>Postponing this statement to the point where we know the length is
>greater than 1, I get no exceptions.
>
>Perhaps, for the purposes of asking the question here, you should change
>Testtokenizea so that the problematic input is a constant string rather
>than being read from the input.
>
>See for example http://stackoverflow.com/help/mcve - the 'minimal' isn't
>quite so important as the 'complete', 'tested', and 'readable'.


I guess this is a case of the error code does not match the error.  I
kept getting an error message about the character appending line.

I tried GDB and stepped through, and seemed to get the same error.

I don't know why, but I agree that is the error.  I understand now
that I knew that in Modula-2 when the token was 1 character, then CH2
would just get the null termination character for all strings.

I still don't understand why stepping line by line did not show me the
correct failure point.

Thanks guys for your help

--rob

PS: CAP is a holdover from Modula-2.  I am used to it, and it is
defined in my spec file to renames Ada.Character.Handling.To_Upper