[lsb-discuss] FYI on parsing something rather large...

Denis Silakov silakov at ispras.ru
Thu Apr 23 05:48:30 PDT 2009


Scott Baeder wrote:
> Well, the good news is that it worked!  Took a bit of time to consume it
> all...Already giving me all sorts of useful feedback on things that our
> developers (at least in this one unit) could do to improve things
>   

Glad to hear this:)

> -rw-rw-r-- 1 linux-app-checker linux-app-checker 944146736 Apr 22 02:09
> ZZZZZZZZZZZZZZZZZZZZZ_x86_upload_data
>   

Yes, *upload_data files created by the tool can be very large. Usually
their size don't exceed the total size of executables and shared objects
in the application, but can be rather close to this value.

One thing that is not currently mentioned anywhere is that it is better
to provide the tool with stripped files, that don't contain debug
information. The thing is that such information can greatly increase the
size of output logs, while it is actually of no value for the analysis.

We are working on handling such situations to drop unnecessary data and
thus decrease the size of output files. However, we can't promise that
the size will decrease significantly for all possible cases in the
nearest future...

> Notice that the test_log.html is over 37M, and as you can guess, trying
> to "navigate" or open up some of the different screens takes a bit of
> time ;-)
>   

Uff, indeed:)

> My guess-timate is that it took about 2:45 min to run, since I started
> it just before I left for the day, which would have been registered as
> 00:00 given the settings on this VM...Yes, this was running on Centos5.3
> VM using VMWare Player on a Lenovo T61P laptop where I only gave the VM
> about 768M of ram...

Actually, not so bad for a huge app and AppChecker inside VM:)

-- 
Regards,
Denis Silakov

Linux Verification Center, ISPRAS
web: http://www.linuxtesting.org
e-mail: silakov at ispras.ru



More information about the lsb-discuss mailing list