Test suite results for test file tbs/tb0364.pp

Test run data :

Run ID:
Operating system: linux
Processor: i386
Version: 3.3.1
Fails/OK/Total: 30/9632/9662
Version: 3.3.1
Full version: 3.3.1-15681-g54a73ae3eb-unpushed
Comment: -Fl/usr/lib/gcc/i686-linux-gnu/12 -O4 -gwl -Fd
Machine: gcc26
Category: 1
SVN revisions: 5e00358f13:d6cf3c98d0:e657d6a07d:54a73ae3eb
Submitter: pierre
Date: 2024/05/09 00:01:00
Previous run: 947626
Next run: 948944

Hide skipped tests

Hide successful tests

Test file "tbs/tb0364.pp" information:

t_id 582
t_adddate 2003/10/03
t_result 0
t_knownrunerror 0

Detailed test run results:

tr_idruntr_oktr_skiptr_result
526064237948292TrueFalseSuccessfully run

Record count: 1

No log of 948292.

Source:

uses
   sysutils;

{ comment by submitter:
  The following statement (which works in Delphi)
    result:=Format('%10.n', [ival*1.0]);
  generated an unhandled exception (and said: Missing argument in format "").
  Checking the Delphi documentation, it agrees with the fpc documentation
  (units.pdf), that a dot should be followed by a  (but Delphi does
  not appear to explicitly state that prec should be an integer).
  It appears that Delphi is treating this like %10.0n, although it is
  potentially undefined behaviour.  The fpc documentation indicates I
  should get an EConversionError exception if there are problems.
  (Actually the documentation may be inconsistent, since it also says
  I may get an EConvertError exception.)

  If I change the format string to %10.0n, the program runs OK using
  fpc, however, my thousand separators do not appear.
}

var
   s : string;
   ival : integer;

begin
   ThousandSeparator:='.';
   DecimalSeparator:=',';
   ival:=1234;
   s:=Format('%10.n', [ival*1.0]);
   writeln('s: "',s,'"');
   if s<>'     1.234' then
     begin
        writeln('Problem with Format');
        halt(1);
     end;
end.

Link to SVN view of tbs/tb0364.pp source.