* [email protected] <[email protected]> [2026-03-29 13:27]:
> On Sun, Mar 29, 2026 at 11:58:35AM +0200, Dr. Arne Babenhauserheide wrote:
> > Jean Louis <[email protected]> writes:
> > > True certainty can be obtained by testing functions and seeing if they
> > > are doing what is meant to be.
> > 
> > 1. That’s only true if the tests are completely exhaustive. Which very
> >    likely makes them a lot longer than the code itself.
> > 2. Who reviews that exhaustive test code?
> 
> Exactly. That "testing" thing sounds as if someone was trying to
> decide undecidables :-)
> 
> Jean Louis, please read [1] and come back then.
> 
> Cheers
> 
> [1] https://sqlite.org/testing.html

Tomas, you were right. SQLite has 590x more test code than the
library itself — 155 KSLOC of code vs 92,053 KSLOC of tests. 100%
branch coverage, crash simulations, fuzzers doing a billion mutations
per day. They literally simulate power failures. So yes, exhaustive
testing is insane — but SQLite did it anyway because billions of
devices depend on it. For the rest of us? Pragmatic testing still
wins. But I stand corrected.

-- 
Jean Louis

P.S. I did not read it. My computer did.

---
via emacs-tangents mailing list 
(https://lists.gnu.org/mailman/listinfo/emacs-tangents)

Reply via email to