I'm working on modernizing Rosetta Code's infrastructure. Starting with communications. Please accept this time-limited open invite to RC's Slack.. --Michael Mol (talk) 20:59, 30 May 2020 (UTC)

Halt and catch fire

From Rosetta Code
Task
Halt and catch fire
You are encouraged to solve this task according to the task description, using any language you may know.
Task

Create a program that crashes as soon as possible, with as few lines of code as possible. Be smart and don't damage your computer, ok?

The code should be syntactically valid. It should be possible to insert [a subset of] your submission into another program, presumably to help debug it, or perhaps for use when an internal corruption has been detected and it would be dangerous and irresponsible to continue.

References


Related Tasks



11l[edit]

Translation of: Nim
assert(0B)

6502 Assembly[edit]

Upon executing this byte as code, the processor will halt. No interrupts can occur either. This does not occur on 65c02-based hardware such as the Apple II or Atari Lynx.

 db $02

This version works on all 6502 models:

forever:
jmp forever

This code is often written as JMP $ which means the same thing. (In addition to the hexadecimal token, $ can refer to the value of the program counter at that instruction's address.

8080 Assembly[edit]

Translation of: Z80 Assembly
di
hlt

8086 Assembly[edit]

Translation of: Z80 Assembly

Disabling interrupts prior to a HLT command will cause the CPU to wait forever.

cli
hlt

68000 Assembly[edit]

The 68000 can only read words or longs at even addresses. Attempting to do so at an odd address will crash the CPU.

CrashBandicoot equ $100001
LEA CrashBandicoot,A0
MOVE.W (A0),D0

Ada[edit]

procedure Halt_And_Catch_Fire is
begin
raise Program_Error with "Halt and catch fire";
end Halt_And_Catch_Fire;
Output:
raised PROGRAM_ERROR : Halt and catch fire

ALGOL 68[edit]

This program will crash immediately on startup.

( print( ( 1 OVER 0 ) ) )

ALGOL W[edit]

This won't halt the CPU but the program will crash immediately on startup.

assert false.

Arturo[edit]

0/0
Output:
>> Runtime | File: halt and catch fire.art
     error | Line: 1
           | 
           | uncaught system exception:
           | division by zero

AWK[edit]

 
# syntax: GAWK -f HALT_AND_CATCH_FIRE.AWK
#
# This won't halt the CPU but the program will crash immediately on startup
# with "error: division by zero attempted".
BEGIN { 1/0 }
#
# This will heat up the CPU, don't think it will catch on fire.
BEGIN { while(1) {} }
#
# Under TAWK 5.0 using AWKW will immediately abort.
BEGIN { abort(1) }
 
Output:
gawk: C:\AWK\HALT_AND_CATCH_FIRE.AWK:5: error: division by zero attempted

C[edit]

int main(){int a=0, b=0, c=a/b;}

C++[edit]

Use an unhandled exception to crash the program.

#include <stdexcept>
int main()
{
throw std::runtime_error("boom");
}
Output:
terminate called after throwing an instance of 'std::runtime_error'
  what():  boom

The output depends on the compiler and platform but should be similar.

C#[edit]

Works with: C sharp version 9

This throws a DivideByZeroException at runtime.

int a=0,b=1/a;

This will throw a compile-time exception, so technically not a valid solution.

int a=1/0;

This one-liner also works

throw new System.Exception();

Computer/zero Assembly[edit]

STP

Crystal[edit]

raise "fire"

F#[edit]

 
0/0
 
Output:
[ERROR] FATAL UNHANDLED EXCEPTION: System.DivideByZeroException: Attempted to divide by zero.
exit status 1

Factor[edit]

REPL[edit]

Causing a stack underflow is trivial; just call any word that expects arguments with an empty data stack.

+
Output:
Data stack underflow

script[edit]

This crashes because Factor expects the data stack to be empty at the end of a program. However, it is not here.

1
Output:
Quotation's stack effect does not match call site
quot      [ 1 ]
call-site ( -- )
(U) Quotation: [ c-to-factor => ]
    Word: c-to-factor
(U) Quotation: [ [ (get-catchstack) push ] dip call => (get-catchstack) pop* ]
(O) Word: command-line-startup
(O) Word: run
(O) Word: load-vocab
(O) Method: M\ vocab (require)
(O) Word: load-source
(O) Word: wrong-values
(O) Method: M\ object throw
(U) Quotation: [
        OBJ-CURRENT-THREAD special-object error-thread set-global
        current-continuation => error-continuation set-global
        [ original-error set-global ] [ rethrow ] bi

deployed[edit]

When deploying as a standalone executable, a main word and vocabulary must be declared. The stack effect checker must be satisfied, so we can't rely on either of the tricks used before. Therefore die is called instead.

USE: kernel IN: a : b ( -- ) die ; MAIN: b
Output:
You have triggered a bug in Factor. Please report.
critical_error: The die word was called by the library.: 0
Starting low level debugger...
Basic commands:
  q ^Z             -- quit Factor
  c                -- continue executing Factor - NOT SAFE
  t                -- throw exception in Factor - NOT SAFE
  .s .r .c         -- print data, retain, call stacks
  help             -- full help, including advanced commands

>

FALSE[edit]

Any function with the exception of ^ (read from stdin) or ß (flush stdin) will cause a stack underflow.

.

Alternatively, the FALSE interpreter expects the stack to be empty at the end of the program's execution, and so leaving a value on the stack is also a valid strategy for crashing the program.

 0

Fermat[edit]

Defines, then calls, a function with no parameters that calls itself. A segfault occurs.

Func S=S. S;

This alternative is five bytes longer but crashes more thoroughly; after a warning about end of line inside a string literal it locks my computer up for a good 2-3 minutes before exiting to the command prompt.

while 1 do !' od;

FreeBASIC[edit]

Instant segfault.

poke 0,0

This alternative crashes the compiler.

#define A() B()
#define B() A()
A()

Go[edit]

This wouldn't survive go fmt which would stretch it out to 5 lines. However, that's not compulsory and the task says do it in as few lines as possible.

package main; import "fmt"; func main(){a, b := 0, 0; fmt.Println(a/b)}


An alternative shorter line would be:

package main; func main(){panic(0)}

GW-BASIC[edit]

0 gosub 0

Haskell[edit]

An alternative to the following is to use undefined.

main = error "Instant crash"

Julia[edit]

To crash the running program:

@assert false "Halt and catch fire."
Output:
ERROR: AssertionError: Halt and catch fire.

To crash the LLVM virtual machine running Julia with Exception: EXCEPTION_ILLEGAL_INSTRUCTION:

unsafe_load(convert(Ptr{Int}, C_NULL))

Liberty BASIC[edit]

This is just one possibility.

Let

Lua[edit]

Tricks could be used to shorten this, particularly from interactive REPL, where -_ would be enough (i.e., attempt arithmetic on a nil global), or from a file _() would be enough (i.e., attempt to call a nil global). This instead focuses on the "be useful elsewhere" aspect of the task, because both seem short-enough as-is:

error(1)
Output:
1
stack traceback:
        [C]: in function 'error'
        stdin:1: in main chunk
        [C]: in ?

Or:

assert(false)
Output:
stdin:1: assertion failed!
stack traceback:
        [C]: in function 'assert'
        stdin:1: in main chunk
        [C]: in ?

Nim[edit]

One possibility:

assert false

Another solution with the same number of characters (we could also use mod instead of div):

echo 1 div 0

Pascal[edit]

Works with: Free Pascal
Do an illegal memory access at $0
begin pByte($0)^ := 0 end.
Output:
Runtime error 216 at $0000000000401098

Perl[edit]

This is not a syntax error, it is a fatal run time error. See "perldoc perldiag".

&a
Output:
Undefined subroutine &main::a called at line 1.

PL/M[edit]

This will terminate the program by restarting CP/M.

100H: GOTO 0; EOF

Phix[edit]

I normally and quite often just use this:

?9/0

The ? means print and/but obviously the 9/0 triggers a fatal error before it gets that far.

Output:
C:\Program Files (x86)\Phix\test.exw:1
attempt to divide by 0

Global & Local Variables

--> see C:\Program Files (x86)\Phix\ex.err
Press Enter...

Alternatives include crash("some message") which produces similar output, and abort(n) which is somewhat quieter with abort(0) meaning (immediately) terminate normally without an error. All of those can be caught by try/catch: should you want to get properly brutal and defeat any active exception handler you can/must resort to inline assembly:

try
    #ilASM{
       [PE32]
           push 1 -- uExitCode
           call "kernel32","ExitProcess"
       [PE64]
           sub rsp,8*5
           mov rcx,1 -- uExitCode
           call "kernel32","ExitProcess"
       [ELF32]
           xor ebx, ebx 
           mov eax, 1  -- SYSCALL_EXIT 
           int 0x80 
       [ELF64]
           mov rax,231 -- sys_exit_group(rdi=int error_code) 
           xor rdi,rdi
           syscall
   }
catch e
    ?e
end try

No output, the try/catch is just for show. ExitProcess/sys_exit are the only non-catchable things I know of, apart from a few other deliberates such as quitting the debugger, and aside from being technically difficult to catch it seems reasonable to classify them as direct actions rather than errors, and that way excuse the non-catchableness.
(I suppose [ok, actually know that] you could also write inline assembly that fubars the call stack to [effectively or quite deliberately] disable any active exception handler[s])

Python[edit]

0/0
Output:
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ZeroDivisionError: division by zero

Quackery[edit]

Ripping lumps out of core definitions will do the trick.

 > quackery

Welcome to Quackery.

Enter "leave" to leave the shell.

Building extensions.

/O> ' tuck take
... 

Quackery system damage detected.
Python reported: maximum recursion depth exceeded


Raku[edit]

++8

Syntactically: Valid.

Semantically: Change the mathematical concept of 8 to 9, either in your whole computer, or maybe the whole universe.

Fails with this run-time error:

Output:
Cannot resolve caller prefix:<++>(Int:D); the following candidates
match the type but require mutable arguments:
    (Mu:D $a is rw)
    (Int:D $a is rw --> Int:D)

The following do not match for other reasons:
    (Bool $a is rw)
    (Mu:U $a is rw)
    (Num:D $a is rw)
    (Num:U $a is rw)
    (int $a is rw --> int)
    (num $a is rw --> num)
  in block <unit> at -e line 1

Alternately, and perhaps more community condoned, to end the program as soon as possible without trying to change the Laws of the Universe, you could just enter:

die
In REPL:
Died
  in block <unit> at <unknown file> line 1

Same character count, exits the program as soon as possible (though trappable if desired through the exception system,) and it looks more like an intentional act rather than a typo. Plus, you can add a message that will be added when it dies to explain why.

Here is a silly alternative : A standalone Unicode counterpart for the yada yada yada operator takes up 3 code units but visually just a single codepoint,

Output:
cat test.raku ; wc test.raku
…
1 1 4 test.raku

raku -c test.raku ; echo $?
Syntax OK
0

raku test.raku ; echo $?
Stub code executed
  in block <unit> at test.raku line 1

1

However when I tried to combine all to test against the Test module, the last one somehow lived through an EVAL,

use Test;
 
dies-ok { ++8 };
dies-ok { die };
dies-ok {};
 
eval-dies-ok '++8';
eval-dies-ok 'die';
eval-dies-ok '…' ;
Output:
ok 1 -
ok 2 -
ok 3 -
ok 4 -
ok 5 -
not ok 6 -
# Failed test at all.raku line 11

so it is indeed at one's discretion whether this one is qualified as a crasher.

REXX[edit]

Version 1[edit]

_=1;_+=
There is no output shown in the DOS window.   This REXX program (using Regina REXX) also
crashes the Microsoft DOS window (application).
ooRexx shows this:
H:\>rexx c2
     1 *-* _+=
Error 35 running H:\c2.rex line 1:  Invalid expression.
Error 35.918:  Missing expression following assignment instruction.

Version 2[edit]

one statement is enough

_+=1
H:\>regina crash
     1 +++ _+=1
Error 41 running "H:\crash.rex", line 1: Bad arithmetic conversion
Error 41.1: Non-numeric value ("_") to left of arithmetic operation "+="

H:\>rexx crash
     1 *-* _+=1
Error 41 running H:\crash.rex line 1:  Bad arithmetic conversion.
Error 41.1:  Nonnumeric value ("_") used in arithmetic operation.

Version 3[edit]

even shorter

+
H:\>rexx crash
     1 *-* +
Error 35 running H:\crash.rex line 1:  Invalid expression.
Error 35.901:  Prefix operator "+" is not followed by an expression term.

H:\>regina crash
Error 35 running "H:\crash.rex", line 1: Invalid expression
Error 35.1: Invalid expression detected at "

Ring[edit]

 
try
see 5/0
catch
see "Catch!" + nl + cCatchError
done
 
Output:
Catch!
Error (R1) : Can't divide by zero !

Ruby[edit]

 
raise
 
Output:
1:in `<main>': unhandled exception

Rust[edit]

Rust provides the panic! macro for stopping execution when encountering unrecoverable errors. This results in a crash, rather than a normal exit.

 
fn main(){panic!("");}
 
Output:
thread 'main' panicked at '', src\main.rs:1:12
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

Swift[edit]

Swift provides a built-in function whose sole purpose is to stop execution in the event of unrecoverable errors. This is different from the standard exit function, as it causes an actual trap (i.e. program crash). As such, it uses the special return value of Never, which allows it to be used in returns that normally expect another type.

fatalError("You've met with a terrible fate, haven't you?")
Output:
$ ./.build/x86_64-apple-macosx/release/Runner
Runner/main.swift:11: Fatal error: You've met with a terrible fate, haven't you?
Illegal instruction: 4

Tiny BASIC[edit]

0 gosub 0

Vlang[edit]

fn main() { panic(0) }

Wren[edit]

Fiber.abort("")

XBS[edit]

Calling the error function in the standard library should stop all running code.

error("Crashed");
Output:
{XBS}: CodeError: Crashed

XPL0[edit]

This overflows the stack. It gives a "Segmentation fault" under Raspberry Pi OS, and just hangs (in some cases such that Ctrl+Alt+Del doesn't even work) under MS-DOS.

proc Recurse; Recurse; Recurse

Z80 Assembly[edit]

The CPU will halt and will require a reset. (Earlier there was a mention that the Game Boy is different in this regard - that was an error; it is not.)

di
halt