Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Ask HN: What perfect software did you discover of recent?

news.ycombinator.com

Post reply on HN
Print view View original

31–40 of 42 posts

Re: Ask HN: What perfect software did you discover of recent?

#34

Post by shoo »

CrimsonCape wrote: post #28
sidmitra wrote: post #23

Earlier quoted context omitted.

There is also https://meldmerge.org/ which i've used on Linux and Mac before.

Does meld include any sort of git browser to dig into old commits of the same file? The ability to pick a file and compare it's states is a pain point I find with git UI apps. For example, TortiseGit will do this but it requires some painful non-intuitive menus and right clicking to resurrect the older data

I don't believe meld embeds any git specific support, but if you configure git to use meld as the tool for diffing and merging, then when you use the git CLI to display a diff, it should invoke meld

e.g. in the CLI you might do something like `git diff commit_a commit_b -- some/path/in/repo` to show the diff from commit_a to commit_b limited to some path

Refer to the "configuration" sections for

  https://git-scm.com/docs/git-difftool
  https://git-scm.com/docs/git-mergetool
these git documentation pages mention meld (along with many other diff tools).

Re: Ask HN: What perfect software did you discover of recent?

#35

Post by shoo »

kevinsync wrote: post #9

Beyond Compare — a visual diff tool. It’s so freakin good. https://www.scootersoftware.com

If you ever end up in a software project where you're helping maintain 100s or 1000s of "diff the output file against the expected file" style regression tests, and some of the files you're diffing are CSV files or so on, beyondcompare has some features that can help a lot. e.g. it can do CSV-aware diffing & help you suppress irrelevant details such as row order (supposing you don't care about row order), or you can mark the timestamp column as irrelevant and ignore differences from that column only. Often files differ superficially as streams of bytes but might be equivalent at a higher level (in terms of the data encoded in the bytes of text) & beyond compare can help with some of this.

If you're in a team with this kind of semi-automated semi-manual regression testing workflow for changes, where there's human-in-loop engineering effort to review regression tests that indicate 'there's some difference between actual output and expected output' and decide if the new output is "OK" or not, the cost of a beyond compare license likely pays for itself in less than a month in terms of freeing up hours of engineering effort.

That said, you can often do better than this by investing engineering effort to build a custom domain-specific file normalizer / differ tool, to e.g. deserialize values, normalize them to eliminate irrelevant details, compare them for equivalence in a domain specific way, show actionable diffs when they differ, which might give a much better result than a high quality general-purpose tool beyond compare.

Post reply on HN