Re: [Bug-wget] Test 16130 from latest wget2 fills up disk

[Tim Ruehsen]

On Wednesday, March 22, 2017 5:05:46 PM CET Dagobert Michelsen wrote:
> Hi,
> 
> the test 16130 from git head of wget2 fills up the disk during running the
> test:
> 
> address@hidden [global]:~/slave/wget2-solaris10-i386/build > cd tests/
> address@hidden [global]:~/slave/wget2-solaris10-i386/build/tests > du
> -sh .* 8.8G   .
>  8.9G   ..
>  186K   .deps
>  413K   .libs
>  8.8G   .test_16130
> address@hidden [global]:~/slave/wget2-solaris10-i386/build/tests > cd
> .test_16130/ address@hidden
> [global]:~/slave/wget2-solaris10-i386/build/tests/.test_16130 > ls -la
> total 18407066
> drwx------   2 buildbot csw            4 Mar 22 16:30 .
> drwx------   7 buildbot csw          175 Mar 22 16:30 ..
> -rw-------   1 buildbot csw            0 Mar 22 16:30 1.txt
> -rw-------   1 buildbot csw      9412149248 Mar 22 16:47 2.txt
> address@hidden
> [global]:~/slave/wget2-solaris10-i386/build/tests/.test_16130 > tail 2.txt
> Content-Length: 17
> Content-Type: text/plain
> Transfer-Encoding: chunked
> 
> FFFFFFF4
> Garbagexhtml+xml,application/xml;q=0.9,*/*;q=0.8
> User-Agent: wget2/1.0.0
> Connection: keep-alive
> 
> loghostlocalhos��p�    ��
> address@hidden
> [global]:~/slave/wget2-solaris10-i386/build/tests/.test_16130 >
> 
> 
> It would be nice if this could be reduced as it breaks my automatic build
> machine.

Hi Dago,

sorry about that... but thanks for the report - something that does not happen 
on 64bit architectures. This is by purpose a bad chunk size corner case, I 
thought I fixed the code for 32bit architectures (at least on a 32bit Debian 
VM), but maybe there are differences. I'll check that soon.

Regards, Tim

signature.asc
Description: This is a digitally signed message part.