Re: [Tlf-devel] TLF-1.3.1 early pre release

[Thomas Beierlein]

Hi Ervin (and others),

I just found time to add your last changes to TLF. Thanks for finding
the problem.

There is still one problem with your solution which you should sort out:
fldigi_get_log_call() is called from background_process which runs as a
separate thread parallel to the logging front end function logit().
That may be ok for just polling fldigi and displaying the
received text in the rtty window. But for manipulating hiscall and
displaying the according country information it may result in problems
as both - your function and callinput() may manipulate the same data.
Furthermore neither getctydata() nor searchlog() are checked to be
threadsafe, so calling it from the background thread may be
problematic.

I have added the code as is to get forward with the 1.3.1 release but
you should sort out the problems in next time.

And just another wish: Please do not add every global variable into
main.c (there are already much to much there). Put it where it
logically belongs (see my last commit).

73, de Tom DL1JBE


 Am Mon, 1 Oct 2018 08:12:32 +0200
schrieb Ervin Hegedüs <address@hidden>:

> On Mon, Oct 01, 2018 at 06:32:54AM +0200, Thomas Beierlein wrote:
> > Am Sat, 29 Sep 2018 23:17:55 +0200 schrieb Ervin Hegedüs
> > <address@hidden>: 
> > > Hi Thomas,  
> > ...  
> > > I just sat down to RIG since one hour ago, and tried to made some
> > > RTTY QSO's, testing the new version.
> > > 
> > > * when the callsign filled from Fldigi, the COUNTRY and ZONE in
> > >   bottom border or Worked window is wrong; that shows always
> > >   the last QSO info
> > > * sometimes the CQ zone filled with _that_ value, not the CQ zone
> > >   of current CALL
> > > * the BMAUTOADD doesn't works; I moved away with click in Fldigi
> > >   waterfall to the next signal, Tlf calculates the new FREQ, and
> > >   tune VFO, may be it defines...
> > >   
> > Hmm, that sounds not so good after all.
> > 
> > The first two problems may be connected to the change from Stephen
> > to enable the use of the cty.dat files from CT9. I checked it
> > working, but must admit only for CW.  
> 
> I think I've found source of all problems, and doesn't connect to
> Stephan's patch. The Fldigi fills the callsign, and there missed
> some extra function call in first two issues.
> 
> The third issue occured because the bmautoadd worked only when
> user types the callsign, the Fldigi callsign set was ignored.
> I've fixed it too.
> 
> There is a plus minor issue: in fldigixmlrpc.c Tlf sets the RIG's
> VFO based on its mode and the Fldigi's carrier. Those values are
> differs, eg: when OP uses USB or LSB, we need to calculate as
> other way the accurate freq. I didn't know about RTTY-REV, so
> I've wrote the code that calculate freq as same way than LSB/USB.
> 
> At this weekend I could work with a K3 in FSK mode, and
> checked RTTY-REV mode, but the calc was wrong: always tune the
> VFO to another direction. The first commit fixes this issue.
> 
> 
> > I will look into the patches you provided in meantime.  
> 
> right, thanks.
>  
> > P.S. Please note that I will be away most of the week, so responses
> > from me may be late.  
> 
> thanks for info,
> 
> 
> 73, Ervin
>  
> 
> _______________________________________________
> Tlf-devel mailing list
> address@hidden
> https://lists.nongnu.org/mailman/listinfo/tlf-devel



-- 
"Do what is needful!"
Ursula LeGuin: Earthsea
--