Re: [Boost-users] [1.34] [iostreams] zlib_compressor works in 1_33, dies in 1_34_0

From: KSpam (keesling_spam_at_[hidden])
Date: 2007-05-08 12:47:11

Jeff,

I modified your test code to move the data internal to the source code.  I 
also added a configuration option to use the bzip2 compressor instead of the 
zlib compressor.  I have attached my version of the test code, along with a 
Makefile.

Here are some specs for my test machine:

Platform: Gentoo Linux x86_64
Kernel: 2.6.19-gentoo-r5
Compiler: gcc version 4.1.1 (Gentoo 4.1.1-r3)
RAM: 2 GB

When I run test_zip_1_33_1, test_bzip2_1_33_1, and test_bzip2_1_34_0, all 
tests run successfully without any noticeable delay or memory overhead (40000 
iterations).

When I run test_zip_1_34_0, I see a significant memory spike that eats all of 
my available RAM.  As the iterations increase, the delay between iterations 
increases.  I ran this test for 20000 iterations because 40000 iterations 
slows my machine to a grinding halt.  I did not receive a bad_alloc 
exception, but I would have preferred that to the near system crash that 
occurred!

This is a showstopper to using boost_1_34_0 on my projects.  Thanks for 
pointing it out!

Justin

On Thursday 03 May 2007 17:46, Jeff wrote:
> Hello,
>
> The code below works fine in 1_33, but throws a bad_alloc exception
> after some 10-20K iterations in 1_34_0.
>
> Am I doing something wrong? Did the API or usage requirements change?
>
> Thanks,
> Jeff