Good day Nicolas--
I had some time to test your changes (though not fully as I never could figure out how to reproduce getting org-note-abort to stick at 't) and they appear to be working.  In doing so, I noticed the following behaviors on the table addition usecase:
* regarding my earlier email that the table row was inconsistently added to the file...I now understand how to make it work consistently.  If  your "saved point" is *anywhere* in the file up to the start (inclusive) of the last line, your row will be added as you expect.  However, if your point is even one character position forward, a new table will be created.* adding a row and aborting it (Ctrl-X Ctrl-K) on aquamacs is always unable to clean itself up in the table case.Thoughts on both issues:
* I thought this might be due to an interaction with desktop mode (or maybe it's aquamacs specific) and that I could workaround it by excluding the target file by customizing desktop-buffers-not-to-save.  Unfortunately, that doesn't appear to work and it's possible a code change is necessary.* When things work, this isn't a big deal as you can easily fixup/delete the row yourself.  However, when you hit the previous issue, you need to separately edit the file by hand.

Thanks for the fixes.
--Brad    On Saturday, April 21, 2018, 6:24:01 AM PDT, Nicolas Goaziou <mail@nicolasgoaziou.fr> wrote:  
 
 Hello,

Brad Knotwell <bknotwell@yahoo.com> writes:

> Good day all--
> Earlier, I had written about the unreliability of adding
> a table-line.  Well, I ran into a similar problem with the
> (significantly simpler) entry capability.  Unlike the table-line, this
> was an easier debug and I determined that org-note-abort was set to
> t which ensures capture will always fail.
> While I don't know exactly how I did it, I suspect it was something
> like the following:* edit a template, test it and find it isn't quite
> what you want* make an (incorrect) change that sets breaks on
> insertion, sets org-note-abort to t and you're stuck until you reset
> it to nil by hand (remarkably, I've had to do the same thing with
> inhibit-trace as well).
> Searching the mailing list and looking at Google hasn't been any more
> illuminating on this.  It's important to note that while
> org-capture-kill sets org-note-abort to t, it's reset correctly when
> used in the normal way.
> Version: 9.1.7
> Thx and I'm guessing something similar happened with the table-line
> issue.

I pushed a fix in maint. Could you test it and report if it solves your
issue ?

Regards,

-- 
Nicolas Goaziou