Could need some advice :-) (long)

Could need some advice :-) (long)

[Ruediger Berlich]

Hi all,

I'm currently changing from C/C++. I worked through the tutorials about
migrating from C/C++ to Ada on http://lglwww.epfl.ch/Ada/ and through
the Lovelace-Tutorial (at least most of it) and now think about starting
with a non-trivial example (since I made the experience, that this is the
best way to learn a language).

So now I could need some advice to get me started, since a lot of things
are still unclear to me or weren't covered in these tutorials.

I would like to write a procedure (or probably more than one :-),
which reads and parses configuration-files or data-files.
Therefore I first need a procedure, which reads a file of
arbitrary lenght and fills each line into a string. Since I don't know
about the lenght of each line and the number of lines in the file, I
need unbounded strings and unbounded arrays to hold the strings
(i.e. no size-limit has to be specified in advance). I read, that this
exists for strings, but didn't find a hint, that it is possible for arrays
of arbitrary type (e.g. arrays of strings). Since a string is just an array
of characters, I suspect that it's possible, though. Under C++ I would
just create a class with an overloaded []-Operator and would copy internally
the array holding the data into a larger array, if the index exceeds the
border of the array. How do I do that in Ada95 ? (I.e. do I have to do the
same as in C++ or is there a predefined package ?)
The next step is then, to take each string, and parse it for the required
information. I would like to cover two cases :
1. The file contains data in the form
1.23 4.56 7.89 ...
3.65 3.44 4.23 ...

2. it looks like
Variable1 = 1.234
Variable2 = 5.678
NKnots = 7
StringVar = "SaveResults.dat u: rwx"
...

So, when I finally have the strings holding a line, I need some functions
to parse the string.
Thereby I would like to tell the procedure at runtime, what is considered
to be a WhiteSpace-character (e.g. ' ' or 'x', if you like that) and what is
a comment (e.g. #, $ or !) .
The procedure shouldn't care, whether I say
Variable1 = 1.234   or   Variable1 1.234 (Could be done by defining '=' as
a WhiteSpace).

In case 1 it want to divide the string into substrings
holding the numbers (as characters) and then convert these strings into
numbers, thereby taking care of typos and raising exceptions,
if that happens (e.g. 1.X34 or 1.2341.456). These numbers are than filled
into an array which is of size NLinesWithNumber x NNumbersPerLine .
The array should be the result of the function.

In case 2 the function takes an array of strings as argument, holding the
names of allowed variables (e.g. 'Variable1' or 'StringVar') together with
some character specifying the type of the Variable (e.g. 'StringVar Str').
It should then convert the values (strings) belonging to the variable into
the respective type and fill it into an array.
The array for each type could be sorted in alphabetical order, so you know,
in which position you have to fill that value. I know how to do all that in
C++ (e.g. some magic programming with 'void *'), but I have no idea how to do
this type-conversion in Ada.


Any hint on what of the above is impossible in Ada and any general comment
are welcome. What I'm looking for, are a couple of keywords of Ada-features,
that you think might be relevant for the above. Also I would be glad about
hints, which of these features are already implemented in some predefined
and standardised Ada-package.

Bye and have a nice day,
Ruediger Berlich
[berlich@pc66.mppmu.mpg.de]