Bug? Kdiff3 diffs seem to have gotten worse

Help
,
1

Asking here to decide whether to file a bug.

I have noticed recently that KDiff3’s diffs are pretty bad. I think they used to be much better than this, because Kdiff3 has been my go-to diff program for over a decade, and only recently I started thinking that something is wrong.

Relatedly, I started comparing to other tools and noticed that e.g. Meld actually nails the example in the Kdiff3 docs section about preprocessor commands - even though Meld doesn’t use those.

Is it possible that Meld is using some more recent diff algorithm? (There’s an open feature request for Kdiff3 to enable use of external diffs, which would help here).

Or maybe I did something to the settings that kneecapped Kdiff3? Nothing I can think of…

Has anyone noticed this? Found any solution? Should I file a bug?

2

hi, welcome.

if you can find clear and reproducible examples of two files that differ in a way that kdiff does not properly identify, or where it falsely identifies a difference where none exists, then yes—that would be report worthy bug at bug.kde.org.

please include these files as attachments to your report so developers and try to reproduce the error.

1 Like
3

OK, thank you. I opened the bug with the example files and results.

1 Like