BTW: Frank and Peter - the "Readln" bug still exists under DJGPP.

This is my workaround; Readln ( MyText, s ); i := length ( s ); if s [ i ] = #13 then Delete ( s, i, 1);

I am not sure whether the fact that "s" contains the #13 character at the end of it is a GPC problem, or a DJGPP problem.

Strange -- I can't see this bug! I got the binary distribution of gpc-980511 (*) for DJGPP, and tested the following program.

(*) The files are dated 11 May -- though GPC gives 2.8.1 as its version number (under Linux as well). Is this correct, Peter?

program Chief23; var t:text; s:string(42); begin rewrite(t,'test.dat'); writeln(t,'OK'); reset(t); readln(t,s); if length(s)=2 then writeln(s) else writeln('Failed: ',length(s),' ',s); close(t) end.

It writes "OK", though test.dat clearly contains a #13 before the #10.

So, if this bug still shows on your system, it must depend on the presence of DosEmu (but why should it?), the version of DJGPP or the phase of the moon. Can anyone else reproduce it?

Frank