I was typing this statement
‘DELETE FROM x WHERE id IN (SELECT id FROM y WHERE y1=‘unknown’ AND y2 =%1)’
Which needs to have the outer ’ changed to "
Spo I finish typing and I select the trailing ’
which selects the immediately preceding ':
'DELETE FROM x WHERE id IN (SELECT id FROM y WHERE y1='unknown" AND y2 =%1)"
So I fix the preceding ’ and get:
'DELETE FROM x WHERE id IN (SELECT id FROM y WHERE y1=‘unknown’ AND y2 =%1)"
I go on to fix the initial ‘:
"DELETE FROM x WHERE id IN (SELECT id FROM y WHERE y1="unknown’ AND y2 =%1)"
Now the 2nd y1=" is a problem. I fix that.
And I finally arrive at what I wanted, though more frustrated than I needed to be. This is an example when I should have the right thing, or done nothing, but did the wrong thing multiple times.
String quoting can be stack or sequences.
print ‘a’ + ‘b’ +“c” would have produced the correct behavior. But in this case, there are no ’ internal to the tokens.
What Komodo should do is treat the quoting as a stack and not a sequence so that “A’B’C” (quote stack: " ’ ) or “A’B"C"D’E” stack: “’”) only changes the corresponding other end of the quote stack, not the adjacent one.
This should also work for:
print ‘a’ + ‘b’ +“c” as the quote stack is either non-existent or ’ for a and b, and " for c.
I’m a bit confused, the examples you gave do not appear to be valid since you are closing the quote prematurely:
'DELETE FROM x WHERE id IN (SELECT id FROM y WHERE y1='unknown' AND y2 =%1)'
With that example the quote would end at y1=, unless you were to escape it. Are you using some language that does quoting differently?
The story cronicles my problem in correcting it to be valid.
I started with ’ as the terminal quotes, but wanted to change them to be the valid "
“DELETE FROM x WHERE id IN (SELECT id FROM y WHERE y1=‘unknown’ AND y2 =%1)”
However the editor kept messing with the quotes around ‘unknown’.
I noticed that if I shift-select the quote, it grabed the matching quote, but only the precedent matching quote, not the stack-based matching quote.
I see. But when it comes to correcting faulty code, how can you expect Komodo to get it right? It might get it right for this particular use-case, but it won’t always. If your code is faulty we can only do so much. I think the only thing you could really do in this situation is to disable the functionality if you run into it often. I don’t think Komodo can handle this any better than it is without creating new issues for “valid” use-cases.
Yeah, that’s why I put it in Discussion not bug. 
I don’t know if there is a right solution, so I’m soliciting feedback.
I do think a possible heuristic is by using the same logic as smart quotes, where
‘a is a start
a’ is an end
‘an example is ‘a terrible thing’ to waste’
‘an - waste’
‘a - thing’
Now they’ve been correctly paired.
My example would need to be tweaked to make this work though:
“DELETE FROM x WHERE id IN (SELECT id FROM y WHERE y1= ‘unknown’ AND y2 =%1)”
(space after y1=)
Which I think is acceptable.
I don’t think that’s feasible. You’re still going to have use cases that aren’t covered and thanks to the code now being “smarter” it becomes far more difficult for end-users to interpret what is happening.
Honestly I don’t think this is something we can reasonably solve.
I think it could be solved in a general case. While I suggested smart quoting I think there are a few characters who actually qualify for this stack-depth treatment: ", ’ , { } [ ], ( ), < > as those are generally paired.
Evaluating those gives:
And does not rely adjacency, and should work in general.
PS. The best/worst thing you can do is tell me it’s not solvable. 
Not “reasonably” solvable. Anything is solvable given enough time.
The problem is we would be applying logic to work with broken code, which will almost definitely break use-cases of valid code.
This just feels like a giant rabbit hole to me, there’s solutions that will cause problems that can then be solved with solutions that will then cause other problems … etc
