[Rd] Tcl socket server (tcltk) does not work any more on R 4.3.2

[peter dalgaard]

I don't think we're going to fix this before 4.3.3. Given that it has gone unnoticed since June 2022 (yes '22) and that tampering in this area has a history of popping up complications in other areas, I think we should leave it alone until 4.4.0. 

(I see that Ivan and Tomas has been on the issue since I started writing, but the above probably still holds true.)

- Peter D.

> On 21 Feb 2024, at 08:01 , webmail.gandi.net <phgrosjean using sciviews.org> wrote:
> 
> Thank you, Ivan for this investigation. I inspected the R changes file (https://cran.r-project.org/doc/manuals/r-devel/NEWS.html) and found nothing about this. I should inspect the sources too!
> 
> It could possibly break other Tcl/Tk related stuff. The doc about Tcl_ServiceAll and Tcl_DoOneEvent is confusing. On one hand, it says when Tcl is used from an external program, Tcl_ServiceAll should be used in its event loop instead of Tcl_DoOneEvent (and the change in the latest R versions goes in that direction). But in the other hand, it is indicated that Tcl_ServiceAll does not always handle all Tcl events and extra Tcl_DoOneEvent should be called in this case. I think we spotted one case where Tcl_ServiceAll is not doing its job correctly. There may be others.
> 
> Since the {tcltk} package was working fine with  "while (Tcl_DoOneEvent(TCL_DONT_WAIT) && max_ev) max_ev—;", unless there is a clear performance enhancement with "while (i-- && Tcl_ServiceAll())", it would perhaps be wise to revert this back.
> 
> Indeed, when I use this on the server side with R 4.3.2:
> 
> library(tcltk)
> cmd <- r"(
> proc accept {chan addr port} {         ;# Make a proc to accept connections
>   puts "$addr:$port says [gets $chan]" ;# Receive a string
>   puts $chan goodbye                   ;# Send a string
>   close $chan                          ;# Close the socket (automatically flushes)
> }                                       ;#
> socket -server accept 12345             ;# Create a server socket)"
> .Tcl(cmd)
> .Tcl("vwait myvar")
> 
> It works again as expected. And vwait is known to call Tcl_DoOneEvent. Of course, in this case, R is blocked and waits for the `myvar` variable on the Tcl side. Anyway, the problem seems to be really in Tcl_ServiceAll not catching all Tcl events.
> 
> All the best,
> 
> Philippe
> 
> ..............................................<°}))><........
> ) ) ) ) )
> ( ( ( ( (    Prof. Philippe Grosjean
> ) ) ) ) )
> ( ( ( ( (    Numerical Ecology
> ) ) ) ) )   Mons University, Belgium
> ( ( ( ( (
> ..............................................................
> 
>> Le 20 févr. 2024 à 17:13, Ivan Krylov via R-devel <r-devel using r-project.org> a écrit :
>> 
>> В Tue, 20 Feb 2024 12:27:35 +0100
>> "webmail.gandi.net" <phgrosjean using sciviews.org> пишет:
>> 
>>> When R process #1 is R 4.2.3, it works as expected (whatever version
>>> of R #2). When R process #1 is R 4.3.2, nothing is sent or received
>>> through the socket apparently, but no error is issued and process #2
>>> seems to be able to connect to the socket.
>> 
>> The difference is related to the change in
>> src/library/tcltk/src/tcltk_unix.c.
>> 
>> In R-4.2.1, the function static void TclSpinLoop(void *data) says:
>> 
>>   int max_ev = 100;
>>   /* Tcl_ServiceAll is not enough here, for reasons that escape me */
>>   while (Tcl_DoOneEvent(TCL_DONT_WAIT) && max_ev) max_ev--;
>> 
>> In R-devel, the function instead says:
>> 
>>   int i = R_TCL_SPIN_MAX;	
>>   while (i-- && Tcl_ServiceAll())
>>       ;
>> 
>> Manually calling Tcl_DoOneEvent(0) from the debugger at this point
>> makes the Tcl code respond to the connection. Tcl_ServiceAll() seems to
>> be still not enough. I'll try reading Tcl documentation to investigate
>> this further.
>> 
>> -- 
>> Best regards,
>> Ivan
>> 
>> ______________________________________________
>> R-devel using r-project.org mailing list
>> https://stat.ethz.ch/mailman/listinfo/r-devel
> 
> 
> 	[[alternative HTML version deleted]]
> 
> ______________________________________________
> R-devel using r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel

-- 
Peter Dalgaard, Professor,
Center for Statistics, Copenhagen Business School
Solbjerg Plads 3, 2000 Frederiksberg, Denmark
Phone: (+45)38153501
Office: A 4.23
Email: pd.mes using cbs.dk  Priv: PDalgd using gmail.com