Mail #10753

Back to main page of archive

Previous mail   Next mail   Unformatted/full headers
Overview  10 days   Subject   Date   Thread   Author  

From: Waldek Hebisch
Subject: boolean functions as conditions
Date: 13 Sep 2004, 14:37:07

Artur Kornilowicz wrote:
> Hello,
> 
> I think gpc does not work properly when one gives a boolean function as 
> a condition in repeat until loop. Please find an example below.
> I tested it on 2 different machines working under Linux and Darwin.
> 
> When you change
> 
> repeat until F1(a,b);
> 
> by
> 
> repeat x := F1(a,b); until x;
> 
> in procedure F2, then it works.
> 
> 
> Greetings
> Artur
> 
> program tester;
> 
> function F1(a,b : string):boolean;
> begin
>     writeln('F1(',a,',',b,')',length(a),' ',length(b));
>     F1:=true;
> end;
> 
> procedure F2(a,b : string);
> begin
> //   writeln('F2(',a,',',b,')',length(a),' ',length(b)); //not necessary
> 	//when you add above, below works
>     repeat until F1(a,b); //incorrect result
>     repeat until F1('3','3'); //correct result //not necessary
> end;
> 
> var c: string; //not necessary
> 
> begin
>    c:='1'; //not necessary
>    repeat until F1(c,c); //correct result //not necessary
>    F2('2','2');
> end.
> 
> 
> Remarks:
> 
> a) to have smaller example you can delete all lines with comment
> "not necessary" at the end of the line.
> 
> b) when F1 and F2 have only one parameter they seem to work.
> 
> Outputs:
> 
> Mac, Darwin
> Version: gpc 20030830, based on gcc-3.3.2
> F1(2,1)1 1
> F1(2,)1 1 // why the second printing is empty?
> F1(2,1)1 1
> 
> Intel, Linux
> gpc 20040516, based on gcc-3.3.1
> artur@alioth:~> ./a.out
> F1(1,1)1 1
> Segmentation fault
> artur@alioth:~>
> 
> 

Later Ernst-Ludwig Bohnen wrote:

> It seems to me that the problem is using the suspicious type 'string'.
> Declare 'type s=string (100)' and change all types 'string' by 's'.
> This works on my Linux with gpc version 20040516, based on gcc-3.2.3.
> F1(1,1)1 1
> F1(2,2)1 1
> F1(3,3)1 1

Only line:

var c: string;

is suspicious. In fact the compiler treats it as:

var c: string(255);

Arthur's version of `F1' can handle strings of arbitrary length
(well, up to stack limit). Changing that to fixed length works around
the problem. In this case better workaround is probably passing
strings as `const' parameters (avoids useless copy):

function F1(const a,b : string):boolean;

However, the original is supposed to work. The patch below below
should fix the problem:

--- gpc-20040516/p/parse.y~	2004-09-11 03:03:51.000000000 +0200
+++ gpc-20040516/p/parse.y	2004-09-13 14:12:38.000000000 +0200
@@ -1818,10 +1818,11 @@
       { emit_nop (); expand_start_loop_continue_elsewhere (1); }
     pushlevel statement_sequence p_until
       { LOCATION_NOTE (@5); expand_loop_continue_here (); }
-    boolean_expression poplevel
+    boolean_expression
       {
         LOCATION_NOTE (@7);
         expand_exit_loop_if_false (0, build_pascal_unary_op (TRUTH_NOT_EXPR, $7));
+        poplevel_expand (0, 1);
         expand_end_loop ();
       }
   ;

 

Previous mail   Next mail   Unformatted/full headers
Overview  10 days   Subject   Date   Thread   Author  

Replies

In reply to

Back to main page of archive

Note: This page contains information that does not originate from the owner of this web site, but from the authors of the mails archived. The owner of this web site is not responsible for the content of such information. Any use of that infomation requires the consent of the respective author.

Where WWW addresses (URLs) in the mails archived are marked as hyperlinks, this is only for the comfort of the reader. The content of the web pages linked to like this does not necessarily reflect the opinion of the owner of this web site or of the authors of the mails archived. The owner of this web site is not responsible for the content of such web pages. Those pages are explicitly not to be considered as part of the content of this page, but merely as references.

This page was created by Crystal 0.999 (Linux 2.4.27/i686).