TRY-EXCEPT doesnt work in fpc 3.0.0!!! [solved]

[Magicus]

 up vote
0
down vote
favorite
I dont understand, why the code below doesnt get executed?!

Im using: FPC 3.0.0 on Ubuntu 16.04

try
 begin
   //f_start is a record which holds an array within.   
   start_arr := f_start^.Arr;
   (start_arr + index)^ := value;
   //...Fehlerursache
 end;
except On EAccessViolation do
 begin
   //...this code SHOULD BE executed...
 end;�

And the Exception: SIGSEGV: EAccessViolation is raised by....

but the Except block doesnt get executed, why??

Pls can someone help me, im really frustrated now..

[Eugene Loza]

If I remember correctly (yeah... I've also had a headache with this )...
You should "ignore" this exception type. You can do this in project options, or (easier) just click ignore when the exception occurs.

[wp]

Please show more details. We are only guessing what you mean:

• "f_start is a record which holds an array within", but: start_arr := f_start^.Arr --> f_start is not a record, it is a pointer to a record!
• if f_start^.Arr is an array then start_arr muist be an array as well. How can you dereference (start_arr + index)^?

Show all declarations. Otherwise this is just a joke!

[molly]

Besides what was already written, if you put a nil into f_start then you most certainly gain an access violation and the except block is executed.

[Magicus]

Ok, sry:

f_start is a pointer to a record, which holds a pointer to an array, as u know, the start-adress of an array:

it looks like:

TBlockPtr = ^Tblock;

TBlock = record
  next: TBlockPtr;
  Arr: ^integer;
end;

the local variable: start_arr is equal to: f_start.Arr !

[molly]

f_start is a pointer to a record, which holds a pointer to an array, as u know, the start-adress of an array:

Still not seeing any array, sorry. At the most it could be a linked list (that wasn't generated).

the local variable: start_arr is equal to: f_start.Arr !

So a pointer to an integer... and you index that manually..... that a bit out of the ordinary unless you meant to say that is the array.

Using your structures (and fill in the rest/unknowns myself) it still generates an access violation for me.

[Magicus]

What is so hard to understand???

And i guess, you missed the point of this topic, i doesnt even matter that it throws any exception, but i WONDER, why the Try-Except Block isnt executed....

When this would work, i would handle the exception by myself, i dont need your advice for that, just dont know, when EAccessViolation is thrown, it doesnt jump to the "except" block.....

focus on the topic and dont get lost by unnecessary things..

[molly]

What is so hard to understand???

Trying to understand what part of your memory you are overwriting.

And i guess, you missed the point of this topic,

No i did not. You are by not providing the minimal information required to answer your question.

i doesnt even matter that it throws any exception, but i WONDER, why the Try-Except Block isnt executed....

Becuase you are doing something stupid like overwriting memory. But, we will never know because you leave us in the dark.

When this would work, i would handle the exception by myself, i dont need your advice for that, just dont know, when EAccessViolation is thrown, it doesnt jump to the "except" block.....

focus on the topic and dont get lost by unnecessary things..

Ok so asking for how your data is structured and what you do with it is unnecessary... thanks for completely wasting my time.

To answer your original question: "it is magic."

[wp]

focus on the topic and dont get lost by unnecessary things..

Hey come on! If you don't provide all information needed nobody can decide what is unnecessary. With that attitude nobody will be willing to help you.

[balazsszekely]

but i WONDER, why the Try-Except Block isnt executed....

Replace EAccessViolation with Exception like this:

Code: Pascal  [Select][+][-]

1.  try
2.   //f_start is a record which holds an array within.  
3.   start_arr := f_start^.Arr;
4.   (start_arr + index)^ := value;
5.   //...Fehlerursache
6.  except
7.    on e: Exception do
8.    begin
9.      ShowMessage('Crap');
10.      //...
11.    end;
12. end;
13.  

If still not works run the executable outside the IDE.

PS: Please be polite with others, they are trying to help you after all.

[bylaardt]

What is so hard to understand???

And i guess, you missed the point of this topic, i doesnt even matter that it throws any exception, but i WONDER, why the Try-Except Block isnt executed....

When this would work, i would handle the exception by myself, i dont need your advice for that, just dont know, when EAccessViolation is thrown, it doesnt jump to the "except" block.....

focus on the topic and dont get lost by unnecessary things..

Probably is not a EAccessViolation

Code: Pascal  [Select][+][-]

1.  try  
2.   //your hided code
3.   except
4.     on e:Exception do
5.      shomessage('Error type is '+e.classname);
6.   end;
7.  
8.  

[Magicus]

How can't overwriting memory not lead to try-except block, when this is true, what you say, this is the most crappiest and fatal design by fpc-developer...

In every other language you are able to catch those errors!

Ok: when you want to know my structure so badly, here is it...

I am creating a TBlock: (defined in some earlier posts)

in my class, i create such a memory-block which has:

* a "next" pointer to the same type
* a pointer to the start-adress of an array

so the TBLOCK is just an pointerholderrecord, if you want and my idea is, in the constructor i create such a block of memory with an initial-capacity, for example: 1000 array- values and 1-next pointer which is nil at the first-time

and in my: Write(index, value) function,

i want to write an element at a given position, ok:

But when i run the write(...) in a for loop of 40000 items, at one index, (mostly by round about index 33000)
SIGSEGV is raising and from there it doesnt execute..

[molly]

And to counterpart the statement "TRY-EXCEPT doesnt work in fpc 3.0.0!!!"

Code: Pascal  [Select][+][-]

1. program ForceException;
2.  
3. {$MODE OBJFPC}{$H+}
4.  
5. uses
6.   SysUtils;
7.  
8. Type
9.   PSomeData = ^TSomeData;
10.   TSomeData = Integer;
11. var
12.   SomePointer: PSomeData;
13. begin
14.   SomePointer := nil;
15.  
16.   try
17.     SomePointer^ := $12345678;
18.    except
19.     On E: EAccessViolation do
20.     begin
21.       WriteLn('Oops acces violation');
22.     end;
23.   end;
24. end.
25.  

So, exception based on acces violation _DO_ work in fpc 3.0 (and all other version of FPC for that matter). Just not for your case because you are doing something else (and that information about that something else i was trying to gather, in order to be able to anwer the question why the except block does not execute for you).

[Magicus]

Im sry for the rude tone, but im really pissed of by this problem currently...

Ok, i switched from EAccessViolation to general Exception, but it still doesnt work...

[rvk]

How can't overwriting memory not lead to try-except block, when this is true, what you say, this is the most crappiest and fatal design by fpc-developer...

Have you tried Run > Run without Debugging like suggested multiple times already here?

You'll see it works fine.
So it's not a problem of FPC.
It's only that Lazarus (or the debugger GDB) trips over the exception (like already mentioned).

AFAIK you can't Try/Except an "External: SIGSEGV".