[Hamlib-developer] Send morse is broken

[Saku <oh...@sr...>]

Hi!

When upgrading from rigctl *Hamlib 4.6~git Jan 19* 05:32:18Z 2023 
SHA=268f44 to rigctld *Hamlib 4.6~git Jun 05* 13:34:59Z 2023 SHA=acc426 
64-bit
I found that morse sending does not work any more in way it was working 
before.

Setup:
      Icom IC7300 firmware 1.41
      Fedora release 37 (Thirty Seven)
*Sending morse uses different (2nd) telnet connection than what is used 
to control rig*

As there is no proper flow control I made (long ago) solution where I 
send text to rigctld letter by letter watching the result code (0 or -9) 
to see when rig does not accept new letters to buffer.

This has been working properly for long time even with texts length 
unlimited and speeds > 30wpm

Now, after upgrading Hamlib, this does not work any more sending text 
sometimes with long delays and repeated letters. I assume long delays 
are caused by repeated spaces in text.

Reason seems to be delays of getting RPRT results and sometimes getting 
-9 when letter is sent anyway.

Here is a dump of sent letters and RPRT codes taken from sending loop 
when using latest Hamlib:

c ->
q ->
   ->
c ->
q ->
   ->
d ->
e ->
   -> RPRT 0
O -> RPRT -9
O -> RPRT 0 RPRT 0 RPRT -9 RPRT 0 RPRT 0 RPRT -9 RPRT 0
O -> RPRT 0
H -> RPRT 0
1 -> RPRT 0
K -> RPRT 0
H -> RPRT 0
   -> RPRT 0
O -> RPRT -9
O -> RPRT 0
H -> RPRT 0
1 -> RPRT 0
K -> RPRT 0
H -> RPRT 0
   -> RPRT -9
   -> RPRT -9
   -> RPRT -9
   -> RPRT -9
   -> RPRT -9
   -> RPRT -9
   -> RPRT -9
   -> RPRT -9
   -> RPRT -9
   -> RPRT -9
   -> RPRT -9
   -> RPRT -9
   ->
k -> RPRT -9
k -> RPRT 0

And here is dump when using older Hamlib 4.6~git Jan 19. Nothing else 
changed than Hamlib version:

c -> RPRT 0
q -> RPRT 0
   -> RPRT 0
c -> RPRT 0
q -> RPRT 0
   -> RPRT 0
d -> RPRT 0
e -> RPRT 0
   -> RPRT 0
O -> RPRT 0
H -> RPRT 0
1 -> RPRT 0
K -> RPRT 0
H -> RPRT 0
   -> RPRT 0
O -> RPRT 0
H -> RPRT 0
1 -> RPRT 0
K -> RPRT 0
H -> RPRT 0
   -> RPRT 0
k -> RPRT 0

Something bad has happened during passed 6 months that was the gap 
between my versions.

I remember I did long and hard work to make morse sending work with 
longer strings than rig can accept at one go. Now it is all wasted time 
and it seems that I have to stay with old Hamlib that works.


-- 
Saku
OH1KH