I've just upgraded to SONAR Professional from 8.5 and also updated my PC from Windows 7 to 10.  Using a Focusrite 18i8 USB.
 
One thing I do in the heat of creativity is to select a section of some track I've laid down, set that to loop, then record another track doing a solo or something.  So yesterday I tried this and wound up with 17 8-bar takes stacked on top of each other.  What I would have done in previous versions would be to copy/paste the looped section to the length I really wanted it and then try to reassemble the clips into something resembling my solo "performance" so I could either save some ideas to do it again, or if it was good enough, comp those sections together for the final bit.
 
This does not seem to be a supported or easy thing to do in SONAR as the loop recording modes seem to expect that whatever you looped was in fact how long you wanted the end result to be, and there's no easy way to reassemble a stack of clips recorded over a loop back into a linear bit.  OK, I can accept that, even though stopping to copy/paste stuff out seriously interrupts my creative (music playing) flow.
 
However I think I have either discovered a bug or don't understand some setting which is making the following happen.
 
When I went to move the last take out to the right, it appears to be missing the first 1/8th note worth of audio.  So when I went to butt take 16 up to take 17, there was a 1/8th note gap at the beginning of take 17.  Then all the 1/8th note sections of each loop were stacked in their respective positions at the start of the looped section, but I don't recall splitting the first 1/8th note off of each clip.  Disaster!
 
So, two questions.
 
1) Is the workflow I have described possible in SONAR?  What would really be great is:
a) Select "n" bars for looping.
b) Record n * x bars of a solo on a new track.
c) Easily extend the looped section "x" times starting at its current position.
d) Unroll the clipped solo back to one contiguous section of audio.
 
2) Anyone else seen this phenomenon where the first little bit of each take is split off, making reassembly really difficult?
 
I'm willing to believe I did something stupid, but the above describes the way I often work.

On my cell now, but Google speed comping. It is actually a fast workflow for what you are doing, but be mindful of the appearance of the smart tool (hot spots on clips determine its function/appearance).

The residual lead in is a partial take, but you can expand take lanes and delete it. Creating a comp is nice, I typically will flatten the comp then copy/paste that to another track if more "intense trickery" is going to be done, but "speed comping" is where to start searching.

Sorry I cannot type more, but I really hate typing on a cell.

Thanks for the note about speed comping, I'll check that out.
 
I tried it again on a new project to make sure I wasn't hallucinating.  All of the earlier takes get split at the point I stop the recording.  I just did a 4-bar loop and recorded 3 takes, stopping at about beat 2 + 1/16.  So take 3 stops at beat 2 + 1/16 and takes 1 and 2 are split at beat 2 + 1/16.
 
If I "undo" then the entire recording disappears, so it doesn't look like this "split all takes at the point where I stopped" was something I did.  I would call this a bug unless someone can point out why I would ever want to do this and how to make it stop.
 
Ok, I chose to store takes in successive tracks and it does not split the earlier takes.  So I'll do that from here on out.

Are the take lanes expanded? I forget if that remnant actually inserts a split point, but the top level track will not overlap (like layers) by default.

Expand take lanes (+ in lower left if track), and you can delete that take. I suspect you are looking at the main track only, which looks like a split but is actually an overlap.

To answer your question, I just had it at the default which just overlaps all takes on one track.  Probably not the most efficient way to do this, but I'm pretty sure I used to do it that way in 8.5 or earlier.  I'm not sure what the rationale is for splitting all earlier takes at the stop point, but I can avoid this going forward by using take lanes and/or recording each take on a new track.
 
Thanks!

Here is a good overview of speed comping.

The 1/8 note break in all the clips probably corresponds to the point where you stopped recording. Check the documentation link from mettelus to see if you can "heal" or reposition that cut point.
 
In previous versions of Sonar, the entire audio recording taken during a loop appears as a single file in either the per-project audio folder or the default Sonar audio folder. Loading that file into the project gives you access to the whole recording as a single clip. Then you decide how to place it and trim it. Things could have changed since then so I don't know if that file is still a straight line recording. Also, this is not a deliberately supported feature, and it's not as simple as what you might have wished Sonar would do. But to rescue your effort, you could give it a try.

Thanks for the tips.
 
I think what's going on is that when I do a solo, I have no idea what I'm really doing and it starts somewhere and goes somewhere else.  So this makes the idea of comping together a take from clips of overlaid 8-bar sections not quite where I want to go.  I think the take concept would be more for something like doing a vocal section over and over and then assemble the most awesomest version of it.  
 
The end result is going to be some multiple of 8 bars but probably less than what I really played because I'm going to trim out the lame sections.  I have a pretty high threshold for lame otherwise the whole thing would be blank!
 
Looks like I'll just do what I did before which is to reassemble the clips into one long clip (or get the single file from the wave folder), copy and paste the looped section into the arrangement and then trim sections out.

The easiest way to get the whole clip into one linear take is to import the recorded audio file from the project's audio folder to a new track. The looped takes are really just SONAR "wrapping" the waveform display visually, and streaming different chunks of the file over the same time range; the file is all one continuous recording.
 
The only caveat is that when you import it, no record latency compensation will be applied, so you'll have to check your Input latency (Preference > Audio > Driver Settings), and slide/nudge the take back that many samples (plus any Manual Offset you might have entered to get sample-accurate latency compensation).