I'm using this regex:
\n\n|(?<=[^,][."*?!\^])(?!\.)[ ](?=["*]?[A-Z])
To match the empty space after a sentence (I'm using [ ] to show you the matches):
First sentence.[ ]Middle a sentence..[ ]Last sentence.
First sentence?[ ]Middle a sentence!.[ ]Last sentence.
"First sentence," Middle a sentence..[ ]Last sentence.
"First sentence."[ ]"Middle a sentence.".[ ]Last sentence.
It's worked fine so far. But now I want the regex not to match when there are .., ?., !., ."., etc. at the end of a sentence. In other words, when there's an extra period:
First sentence.[ ]Middle a sentence.. Last sentence.
First sentence?[ ]Middle a sentence!. Last sentence.
"First sentence," Middle a sentence.. Last sentence.
"First sentence."[ ]"Middle a sentence.". Last sentence.
I thought adding an (?!\.) would do the trick:
\n\n|(?<=[^,][."*?!\^])(?!\.)[ ](?=["*]?[A-Z])
But as you can see here, it's still matching the empty spaces after the extra period.
Why is this and how to fix it?
You can extend this negated class [^,] with all other characters that you don't want to occur as one-before-last character in a sentence.
So like: [^,.]
If however you want to allow ." to end a sentence, and only want to avoid two points (or !. and ".), then put your look around condition there as look-back:
(?<=[^,][."*?!\^])(?<![.!?"]\.)