[imiric]

I wouldn't say it's even that. To me it reads like they're using sane defaults for Deno that any developer would use. The fact performance can be improved by leveraging knowledge of Deno internals is irrelevant. Comparing default OOB performance is an apples-to-apples comparison.

This "exposé" seems like mud slinging from a disgruntled Deno maintainer with inherent bias to see Deno perform better. If anything, it paints a poor picture of Deno, not Bun.

[girvo]

No, I disagree. A couple of those things pointed out are fair: they’re unfair comparisons that aren’t apples to apples, which matters for benchmarks. The SQLite one for example is like me benchmarking a game on two PCs to test graphics performance, but with wildly different settings on both. It tells the end user nothing useful about the comparative performance.

All of those things are easy mistakes to make though. Where it’s a bit not great is if there was apparently communication about it and the Bun team didn’t at least change the SQLite comparison — it’s not “OOTB” behaviour, it’s a library.

[imiric]

That's fair. It seems both issues have now been fixed[1].

I still think the approach of reporting this by the Deno developer was unnecessarily hostile.

[1]: https://news.ycombinator.com/item?id=33199911

[girvo]

That I absolutely agree with, this could’ve been resolved a lot more maturely.