On Fri, Jul 26, 2019 at 02:12:39PM +0200, Johannes Schindelin wrote: > I was wary about this patch series ever since I got aware that it > refuses to use an already-available JSON schema validator (such as > `ajv`, a seemingly well-established all-purpose validator that even > claims to be the fastest validator) and instead tries to push down our > throat the first Go-implemented thing in git.git. Well, to be fair, the cover letters of all versions of this patch series included the following bit: I am happy to rewrite this validator in some other language, provided that the language has a JSON-Schema library supporting at least draft-04. Having said that, using an existing validator would make me happier, too. At the very least the commit message of the patch adding the validator should argue convincingly why rolling our own is worth it. > Besides, I have to admit that I am quite disheartened to see these lines > in the log: > > -- snip -- > 2019/07/26 11:39:28 Validated items: 0 > 2019/07/26 11:39:32 Validated items: 10000 These are necessary for Travis CI, which aborts the build if it doesn't receive output for 10 minutes, thinking that the tests hang somewhere. Yeah, less lines like these would be sufficient to avoid that timeout, but, of course, the real solution would be to not validate all of 'make test'.
