Re: extend diff3 to ignore whitespace differences - here's a patch

[Paul Eggert]

"John Lumby" <address@hidden> writes:

> Could one simply say that the result will be identical to each of the
> files as reported by diff -w?

I don't think that would explain the situation well enough.  The
actual situation is something more like "we'll output one of the three
lines that diff -w thinks is 'equal', but we won't tell you which line
it is".  And things get more complicated if we also consider other
diff options that ignore differences, e.g. -B or -I, as these options
can ignore entire hunks and can cause line numbers to unexpectedly
change (from diff3's point of view).