[bug#54241] [PATCH 0/4] 'github' importer gracefully handles rate limiting

[Maxime Devos <maximedevos@telenet.be>]

[-- Attachment #1: Type: text/plain, Size: 1264 bytes --]

Ludovic Courtès schreef op za 05-03-2022 om 23:03 [+0100]:
> Maxime Devos <maximedevos@telenet.be> skribis:
> 
> > Ludovic Courtès schreef op do 03-03-2022 om 22:14 [+0100]:
> > > +(define (update-rate-limit-reset-time! headers)
> > > +  "Update the rate limit reset time based on HEADERS, the HTTP response
> > > +headers."
> > > +  (match (assq-ref headers 'x-ratelimit-reset)
> > > +    ((= string->number (? number? reset))
> > > +     (set! %rate-limit-reset-time reset)
> > > +     reset)
> > > +    (_
> > > +     0)))
> > 
> > When can this second case happen?
> 
> I don’t know if it’s supposed to happen.  It’s defensive programming:
> better keep going than crash if the server starts behaving slightly
> differently.

If it's not supposed to happen, can it at least be reported with a
warning, such that we then know that 'update-rate-limit-reset-time!'
needs to be extended or GitHub needs to be contacted?

FWIW, I think crashing in case of bogus HTTP answers is fine, as long
as it crashes with a _nice_ error message ("guix: error: HTTP server
foo.com returned an unrecoginised X-Ratelimit-Reset $SOME_STRING" or
something like that) instead of some vague backtrace.

Greetings,
Maxime.

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 260 bytes --]