compare files whose lines have been broken

This is a very difficult issue that I have been attempting to build a resolution on since the question was originally asked. It seems to me that the logic used in the absnse of line breaks makes this extrememly painful. If you remove all line breaks then the two files are all single line files, all diff utils that I have used to comparisons based upon line numbers, so they would not work. I beleive to otain your goal it would be best for you to write a bash script that first removes the line breaks, then increments one character at a time and reports the differences. However this would still be flawed because one a single space is detected it would then report all further characters and changes since the characters in the specific locations do not match. You may want to add some logic stating that if the character was replaced with a space to evaluate the non-space character against the next character in the opposite file.

you could use output each of those files into a files that have
a single line only, then compare the one-line files.
In other words, you replace the new line character by one space and you
remove extra spaces.

A seder could do that for you.
Example

###copy and paste to seder.sed below this line

#!/bin/sed -f

:loop
N
$!b loop

s?[ \t\n]\{1,\}? ?g

#### end of copy ####

After you pasted this to a sed script file (eg seder.sed) you make the
script executable:

chmod +x seder.sed

Now, let's say the two files you want to compare are file1.txt and file2.txt.
You run these commands:

./seder.sed file1.txt > firstfile
./seder.sed file2.txt > secondfile
diff firstfile secondfile