brundlefly
Looks to me like legato.cal is working correctly to extend the selected notes, and they just end up overlapping and hiding the unselected notes in between. In any case, it makes no musical sense to apply legato to notes of the same pitch, and not select all of them in the range.
 

Now I know how it works. :-)
 
In the image below, I randomize notes in a bass track (to humanize them, but I'm intentionally doing it a bit too much in the image just to make it easy to see), then select a set of notes one by one to apply legato. This is the correct way to do it, isn't it?
 

 
 
I now wonder what the OP did to cause the issue he explained.
 

icontakt
I now wonder what the OP did to cause the issue he explained.

 
I can reproduce the OP's behavior.
 
Steps to reproduce:
1.  Enter several 2 or 3 note chords in the PRV
2.  Select only the top notes of the chord
3.  Run Legato.cal
 
Expected results:  Only the top notes of the chords should be extended
 
Actual results:  All notes of the chords are extended and become selected
 
I haven't looked at the code in the Legato.cal script yet, but it appears to select and extend all notes in the selected time range regardless of individual note selection.
 
-- Ron

Yes, if you want individual pairs of notes to be played legato, you have to process them individually like that. BTW, you can keybind .CALs to speed up a process like that.
 
I suspect the OP was either mistaken about what was selected or just misinterpreted the result somehow.

brundlefly
I suspect the OP was either mistaken about what was selected or just misinterpreted the result somehow.

The OP's behavior can be reproduced.
Did you miss my post?
 
-- Ron

Yes, Ron, I did miss your post. 
 
I've since played around a little, and was able to repro the problem, but only when using hard-quantized MIDI where "melody" and "chord" note start times were identical. I also noticed that notes that were not selected initially become selected after processing. So, though I haven't looked at the CAL, myself, it appears that the CAL may be adding to the selection as it processes, and it has problems with notes that have identical start times.
 
This latter issue isn't too surprising because obviously there's a logical problem about what to do when the "next" note starts at the same time as the "current" one. Obviously legato makes more sense for a monophonic voice/line where it's impossible for two notes to start at the same time. Selecting only melody notes should avoid that problem, but if CAL is adding to the selection as it processes so that it can encounter other notes that start at the same time that's going to be a problem.
 
Some things I saw also seemed to suggest that the CAL might not be re-initializing variables properly when it's run a second time, or maybe pointers aren't getting disposed of by SONAR. Undoing, re-selecting and running it again can give different results the second time.
 
Since CAL is deprecated, and CAL scripts tend not to be very thoroughly debugged, especially with regard to unexpected input, anything is possible.
 
All that said, I haven't seen a problem when selecting a melody line and a running legato.cal in an unquantized piece that didn't have any selected notes starting on the exact same tick as some other note.
 

brundlefly
All that said, I haven't seen a problem when selecting a melody line and a running legato.cal in an unquantized piece that didn't have any selected notes starting on the exact same tick as some other note.

Completely agree, Dave. 
 
In the case of the OP, it seems the best approach is to separate the right and left hand parts before processing, since we don't know how often melody notes in the right hand occur at the same time of the harmony notes in the left hand.
 
-- Ron

brundlefly
BTW, you can keybind .CALs to speed up a process like that.

 
Yes, I use keyboard shortcuts for other CALs I regularly use. I'll keybind legato too.
 

brundlefly
I also noticed that notes that were not selected initially become selected after processing. 

 
Ah, that's why I though it was a bug when I captured the first screenshot. I performed undo and saw the notes I didn't select were selected.

Hi everyone.

Yes, the problem is that thr CAL selects (and affects) unselected notes in the vertical plane of the selected horizontal range. So the best solution for this case is separate the melody from the harmmony in individual clips or do the legato manually when there are just a few notes.

Thank you everyone!!

I suggested that in post #7

Bristol_Jonesey
I suggested that in post #7

 
 And so did I in post #3.