[BUG] FASMW huge leakage + Erroneous time measurement

Index > Compiler Internals > [BUG] FASMW huge leakage + Erroneous time measurement

Author

Thread

LocoDelAssembly
Your code has a bug

Joined: 06 May 2005
Posts: 4624
Location: Argentina

LocoDelAssembly 16 Sep 2008, 18:28

(I have edited this post)

Try compiling the attached code several times, there will be a moment in which you will receive "Error: Out of memory" permanently and your computer will have no memory left.

I don't know what is the minimal code required, sorry... But happens even on successful compilation also.

Additionally the compilation time is seriously wrong, it reports thousands of seconds when actually it takes less than 2 seconds on my computer.

Description:
Filesize:	6.46 KB
Viewed:	5826 Time(s)

Description:		Download
Filename:	esf.asm
Filesize:	40.09 KB
Downloaded:	451 Time(s)

Last edited by LocoDelAssembly on 16 Sep 2008, 20:34; edited 1 time in total

16 Sep 2008, 18:28

revolution
When all else fails, read the source

Joined: 24 Aug 2004
Posts: 20333
Location: In your JS exploiting you and your system

revolution 16 Sep 2008, 18:34

I get this with 100M fixed memory size:

Code:

flat assembler  version 1.67.28  (100000 kilobytes memory)
LocoDelAssemblyTest.asm [813]:
        LOCAL   hPen            :DWORD,\
error: illegal instruction.

16 Sep 2008, 18:34

revolution
When all else fails, read the source

Joined: 24 Aug 2004
Posts: 20333
Location: In your JS exploiting you and your system

revolution 16 Sep 2008, 18:36

How many times should I run it? I've put it in a batch file to loop and so far after a few hundreds of runs I can see no memory leak in task manager.

16 Sep 2008, 18:36

revolution
When all else fails, read the source

Joined: 24 Aug 2004
Posts: 20333
Location: In your JS exploiting you and your system

revolution 16 Sep 2008, 18:47

Oh shit, I very stupid, you said FASMW. I was using FASM. Sorry Embarassed

16 Sep 2008, 18:47

LocoDelAssembly
Your code has a bug

Joined: 06 May 2005
Posts: 4624
Location: Argentina

LocoDelAssembly 16 Sep 2008, 19:01

Yet I think it is worth to mention my settings, I'm using 65536 KB of memory and on each (failed) compilation (Ctrl+F9) I lose 64 MB of memory. I have even received the yellow balloon warning me that there is no more memory and that Windows will try to expand it.

Once I have it fully ported to FASM (I went to lunch after creating this thread), I'll check if a successful compilation also leaks memory.

16 Sep 2008, 19:01

Tomasz Grysztar

Joined: 16 Jun 2003
Posts: 8353
Location: Kraków, Poland

Tomasz Grysztar 16 Sep 2008, 20:25

It seems that it is losing the whole fasm's core thread's memory block (it is allocated only one once per compilation). Thus perhaps the assembly thread goes into some wrong place and doesn't clean up or something...

16 Sep 2008, 20:25

LocoDelAssembly
Your code has a bug

Joined: 06 May 2005
Posts: 4624
Location: Argentina

LocoDelAssembly 16 Sep 2008, 20:37

Tomasz, I have edited my post, now the code is compilable and executable and I found an extra bug (I edited the thread's title to reflect that).

16 Sep 2008, 20:37

LocoDelAssembly
Your code has a bug

Joined: 06 May 2005
Posts: 4624
Location: Argentina

LocoDelAssembly 16 Sep 2008, 22:41

OK, I've found the minimal code:

Code:

abc fix  ; the symbol must be 3 chars wide at least to trigger the bug

And this one to trigger the time bug at the same time:

Code:

qweaaaa fix

(Around 9,000.00 seconds)

So the bug may be related to http://board.flatassembler.net/topic.php?t=9131

16 Sep 2008, 22:41

< Last Thread | Next Thread >

Forum Rules:

You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You cannot attach files in this forum
You can download files in this forum