improve matching lines with different column value(s)
I think the algorithm used by Meld to find the differences within matching lines can be improved. In the comparison of the two attached files file_a.txtfile_b.txt Meld starts at line 39 not to notice any more that the lines differ only in their value in the fourth column. BTW the issue does not appear if the number of additional lines at the end of the second file (file_b.txt) w.r.t. the first file (file_a.txt) is reduced from fifty to only a few, e.g. three.