go-diff
go-diff copied to clipboard
Munged text leads to incorrect diffs
https://github.com/sergi/go-diff/commit/db1b095f5e7c905e196ff6bfd56189a41aa76309 introduces a bug in its change from diffLinesToRunesMunge
to diffLinesToStringsMunge
. Since each line is represented by 1 or more ascii characters, it's possible for the diffing algorithm to split hashed lines incorrectly whereas before the rune indexed lines were indivisible.
For instance, DiffLinesToChars
could return hashed strings such as:
"1,2,3,4,5,6,7,8,9,10,9,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,9,27,28,9,29,30,31,32,33,9,34,35,36,37,38,39,40,41"
"42,9,10,9,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,9,27,28,9,29,30,31,32,33,9,34,43,36,37,38,39,44,45,46,47"
DiffMain
may then split the leading 42
such as:
[{Delete 1,2,3,} {Equal 4} {Delete ,5,6,7,8} {Insert 2} {Equal ,9,10,9,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,9,27,28,9,29,30,31,32,33,9,34,} {Insert 4} {Equal 3} {Delete 5} {Equal ,36,37,38,39,4} {Delete 0} {Insert 4} {Equal ,4} {Delete 1} {Insert 5,46,47}]
And the resulting diff after hydration is completely wrong.
This affects users of the DiffLinesTo*
APIs as well as any user that passes true
for checklines
in DiffMain
or DiffMainRunes
.