Page 1 of 4
Studies against AV false positives
Posted: Fri Apr 12, 2013 8:39 am
by Didelphodon
This thread shall be dedicated to the quest for finding the reason(s) why Purebasic executables so often raise false positive alerts in anti virus products.
See it as a joint venture to reveal those secrets to find a solution ...
Re: Studies against AV false positives
Posted: Fri Apr 12, 2013 8:43 am
by Didelphodon
Tried the following things:
Threw an absolutely empty default PB executable in virustotal:
See ...
https://www.virustotal.com/de/file/8793 ... 365751641/
=> 7 false positives
Again, threw an absolutely empty default PB executable in virustotal, but this time with any of the information fields filled out (compiler options):
See ...
https://www.virustotal.com/de/file/4cd0 ... 365752015/
=> only 3 false positives - note, empty information fields seem to trigger
Re: Studies against AV false positives
Posted: Fri Apr 12, 2013 8:46 am
by Didelphodon
Hm, DrWeb is one of the false positives - I'm going to ask a friend at DrWeb to do me a favour checking what triggers ...
EDIT: Just dropped my friend an email. To be continued ...
Re: Studies against AV false positives
Posted: Fri Apr 12, 2013 11:10 am
by IdeasVacuum
I applaud your intent, but fear you could be flogging a dead horse..........

Re: Studies against AV false positives
Posted: Fri Apr 12, 2013 11:38 am
by marc_256
é#@à%* I just installed NOD32 version6
And all my pb .exe files are erased ??!!
They are not even in quarantine ...
Marc,
Re: Studies against AV false positives
Posted: Fri Apr 12, 2013 11:42 am
by IdeasVacuum
That will be an option somewhere in NOD32.
Re: Studies against AV false positives
Posted: Fri Apr 12, 2013 11:52 am
by Didelphodon
Good news from DrWeb. They had a signature based on a false positive from somewhen in their database which is now removed. Hence there shouldn't be any more false positives with Purebasic executables (with good intention of course) in the future. We'll see ...
Re: Studies against AV false positives
Posted: Fri Apr 12, 2013 12:17 pm
by Didelphodon
However, I encourage everyone to fill out those detail informations in the compiler options (like company name, etc.) as it seems to be one of the aspects heuristics look for.
On one hand understandable that this might be an aspect but on the other hand a rather weak one

Re: Studies against AV false positives
Posted: Fri Apr 12, 2013 12:52 pm
by MachineCode
I read in another (software business) forum that if you digitally sign your executables (via Comodo or whomever) then no virus app will flag it as a false positive at all. Could solve the whole problem, but it's not cheap.
Re: Studies against AV false positives
Posted: Fri Apr 12, 2013 3:50 pm
by jpd
Hi Didelphodon,
is possible to recieve the first sample?
or the PB code
Thanks
jpd
is not necessary reproduced here with a simple debug line and create exe!!
Re: Studies against AV false positives
Posted: Fri Apr 12, 2013 4:11 pm
by jpd
Hi,
Here a Analyse from MAG2
Interesting result for a empty file
Process/Thread Events
Creates process: C:\windows\temp\test.exe ["C:\windows\temp\test.exe" ]
Network Events
Sends data to: 10.74.1.255:137
Best
jpd
Re: Studies against AV false positives
Posted: Sat Apr 13, 2013 1:41 am
by MachineCode
jpd wrote:Interesting result for a empty file
An empty file (exe) is not actually blank. It still contains headers, setup and initialisation code. Any of that can trigger a false positive.
Re: Studies against AV false positives
Posted: Sat Apr 13, 2013 9:10 am
by Didelphodon
MachineCode wrote:jpd wrote:Interesting result for a empty file
An empty file (exe) is not actually blank. It still contains headers, setup and initialisation code. Any of that can trigger a false positive.
True! That's actually the footprint of a compiler.
Re: Studies against AV false positives
Posted: Sat Apr 13, 2013 10:58 am
by DoctorLove
jpd wrote:Hi,
Here a Analyse from MAG2
Interesting result for a empty file
Process/Thread Events
Creates process: C:\windows\temp\test.exe ["C:\windows\temp\test.exe" ]
Network Events
Sends data to: 10.74.1.255:137
Best
jpd
Is that your local IP? as in its trying to connect to your debugger?
Re: Studies against AV false positives
Posted: Sun Apr 14, 2013 5:35 pm
by doctorized
Once I had this problem with NOD32 and a dll file that contains a DataSection.
I solved my problem with this:
Code: Select all
DataSection
Data.b .......... ; 20 bytes
Data.l $12345678
Data.b ........ ; rest of the bytes
EndDataSection
When I read the data section I read from 1 to 20 and then from 25 and after.