Exception when testing non-existent property of Uint32Array and bug with throw inside switch()

[espruino-discuss]

Posted at 2018-02-25 by mrQ

just a "picture" without words;)

process.on('uncaughtException', function(err) {
  console.log('!!!UNCAUGHTEXCEPTION!!!',err);
});

const arr= new Uint32Array([1,2,3,4]);     // or any other ArrayView

console.log('test E...');
if (!E.xyz) console.log('E ok');   // works as expected
console.log('test arr...');
if (!arr.xyz) console.log('arr ok'); // throws exception - not as expected
console.log('done all ok');

EspruinoWIFI 1v96

[espruino-discuss]

Posted at 2018-02-25 by mrQ

Some more suspicious things...

This behaves as expected:

process.on('uncaughtException',(err)=>{
  console.log('!!!UNCAUGHTEXCEPTION!!!',err);
});
const arr= new Uint32Array([1,2,3,4]);

console.log('-A1-');
if (!!arr.xyz) console.log('-A2-');
console.log('-A3-');

But this

process.on('uncaughtException',(err)=>{
  console.log('!!!UNCAUGHTEXCEPTION!!!',err);
});
const arr= new Uint32Array([1,2,3,4]);

console.log('-B1-');
try{
  if (!!arr.xyz) console.log('-B2-');
  console.log('-B3-');
}
catch(e){
  console.log('-B4-',e);
}
console.log('-B5-');

reports some strange error (see 2nd UNCAUGHTEXCEPTION message):

-B1-
!!!UNCAUGHTEXCEPTION!!! Error {
  "msg": "Field or method \"xyz\" does not already exist, and can't create it on Uint32Array",
  "type": "Error",
  "stack": " at line 2 col 12\n  if (!!arr.xyz) console.log('-B2-');\n           ^\n"
 }
!!!UNCAUGHTEXCEPTION!!! SyntaxError {
  "msg": "Got catch expected EOF",
  "type": "SyntaxError",
  "stack": " at line 1 col 1\ncatch(e){\n^\n"
 }
-B5-

Maybe another topic relates to this: i think that exceptions got lost without any notice when they occured inside a switch statement. i try to come up with some test code for this, too...

[espruino-discuss]

Posted at 2018-02-25 by mrQ

ok, as promised some sample code to reproduce the second problem when throwing inside a switch statement.

without switch(), everything behaves as expected:

process.on('uncaughtException',(err)=>{
  console.log('!!!UNCAUGHTEXCEPTION!!!',err);
});
/*
expect output === achieved output:
  -D0-
  -D1-
  CATCHED ... "simulate crash"
  -D8-
  -D9-
*/
function test( crit){
  try{
    const arr= new Uint32Array([1,2,3,4]);
    console.log('-D1-');
    if( crit) throw new Error('simulate crash');	// such as thrown by testing  Uint32Array.xyz
    console.log('-D2-');
  }
  catch(err){
    console.log('CATCHED',err);
  }
  console.log('-D8-');
}

console.log('-D0-');
test('crash');
console.log('-D9-');

adding the switch() brings up some fancy extras:

process.on('uncaughtException',(err)=>{
  console.log('!!!UNCAUGHTEXCEPTION!!!',err);
});
function test( crit){
  try{
    const arr= new Uint32Array([1,2,3,4]);
    console.log('-C1-');
    switch (crit) {
      case 'ok':  
        console.log('-C2-C3-');	
        break;

      default:  
        console.log('-C2-');
        if( crit) throw new Error('simulate crash');	// such as thrown by testing  Uint32Array.xyz
        console.log('-C3-');
        break;
    }		
    console.log('-C4-');
  }
  catch(err){
    console.log('CATCHED',err);
  }
  console.log('-C8-');
}

console.log('-C0-');
test('crash');
console.log('-C9-');

expected output (tested on chrome):

  -C0-
  -C1-
  -C2-
  CATCHED Error: simulate crash
  -C8-
  -C9-

achieved output:

  -C0-
  -C1-
  -C2-
  -C4-                                            * should not exist
  -C8-
  !!!UNCAUGHTEXCEPTION!!! ... "simulate crash"    * should be CATCHED
  -C9-

[gfwilliams]

Posted at 2018-02-26 by @gfwilliams

I'll look into the exception on Typed Arrays - it shouldn't complain about it at that point - only when you try and write it. I filed an issue for it here: espruino/Espruino#1351

For your try... catch I bet you just pasted it into the left-hand side? Espruino bracket-counts to see when it should execute a line, so if you write code with all brackets closed and hit enter then it's executed. If you upload from the right-hand side of the IDE it's got some stuff in there to counteract that.

Thanks for the report on the switch as well - I'll have a look at that. There's now an issue filed for it here espruino/Espruino#1352