Hi GPC'ers!
Before I forget: Reading specs from /usr/local/stow/gpc-CVS/lib/gcc-lib/i686-pc-linux-gnu/2.95.2/specs gpc version 20001116, based on 2.95.2 19991024 (release)
The following test program crashes GPC. Maybe my statement is not legal, but anyway the compiler should just complain, not just lie down and die.
: program Test; : : type : PT1 = ^T1; : PT2 = ^T2; : : type : T1 = record : a : Integer; : b : Real; : end; : T2 = record : a : Integer; : b : Real; : end; : : begin : end.
Splicing together the type definitions, the program compiles fine. So there's no real problem for me, except that I think the compiler should just complain. OTOH, I don't know if the statement was correct, either.
: program Test; : : type : PT1 = ^T1; : PT2 = ^T2; : T1 = record : a : Integer; : b : Real; : end; : T2 = record : a : Integer; : b : Real; : end; : : begin : end.
Greetings,
Nicola Girardi wrote:
Before I forget: Reading specs from /usr/local/stow/gpc-CVS/lib/gcc-lib/i686-pc-linux-gnu/2.95.2/specs gpc version 20001116, based on 2.95.2 19991024 (release)
The following test program crashes GPC. Maybe my statement is not legal, but anyway the compiler should just complain, not just lie down and die.
Yes, it's not legal (the use and declaration must be in the *same* `type' block). Yes, GPC should not crash (nicola1.pas).
Frank