Hi Pete and Bill,

Thanks, Pete, for finding and reporting this longstanding bug.

As Bill diagnosed, the real error was in gen9.f90: the call to 
packbits() passed argument 2 as "nsym2" where it should have passed 
"nsym2/3".  This caused the (undetected) out-of-bounds accesses to 
arrays dbits() and sym().  Writing to sym(i) with i>69 definitely 
clobbered something... though at present I'm not sure what was clobbered.

I do have a hypothesis, though it may be wishful thinking.  I think it's 
possible that this bug may be the root cause of undecodable JT9 
transmissions that sometimes occur in response to a CQ.  I will try to 
think of a way to test this hypothesis.

        -- 73, Joe, K1JT

On 2/1/2016 8:46 AM, [email protected] [wsjtgroup] wrote:
> On 01/02/2016 11:53, [email protected] [wsjtgroup] wrote:
>
>   We have Fortran array bounds checking turned on for our debug builds and I 
> am surprised that this defect has not caused a run time error as it is 
> clearly a bounds violation. Hi Pete,
>
> further to my comment above: the reason why the array bounds checking did not 
> pick this up is that the array dummy argument sym() was declared as assumed 
> size (sym(*)) which erases all type information from the caller, hence the 
> compiler cannot generate the required bounds checking code.
>
> I have fixed the original defect and also amended the packjt module code such 
> that it uses assumed shape arrays for dummy arguments and therefore would 
> detect this bounds error in debug builds.
>
> 73
> Bill
> G4WJS.

------------------------------------------------------------------------------
Site24x7 APM Insight: Get Deep Visibility into Application Performance
APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
Monitor end-to-end web transactions and take corrective actions now
Troubleshoot faster and improve end-user experience. Signup Now!
http://pubads.g.doubleclick.net/gampad/clk?id=267308311&iu=/4140
_______________________________________________
wsjt-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/wsjt-devel

Reply via email to