fix(cache): add response delay to corrected age value

[RaphaelFakhri]

RFC 9111 section 4.2.3 defines a stored response's initial age as:

apparent_age           = max(0, response_time - date_value)
response_delay         = response_time - request_time
corrected_age_value    = age_value + response_delay
corrected_initial_age  = max(apparent_age, corrected_age_value)

CacheHandler captured a single timestamp at response-header time and computed
max(apparentAge, resAge). There is no request_time, so response_delay was always
zero and the corrected_age_value term reduced to the bare Age the origin sent.

The delay matters because the Age an origin reports is already out of date by the time the
response finishes arriving. Dropping it stores a slow response as younger than it is, and a
response near its freshness boundary is then served after it has actually expired.

Reproduction

Origin sends Cache-Control: public, max-age=600, Age: 100, and delays 3 seconds:

response_delay measured   : 3030 ms
upstream Age              : 100 s
RFC corrected_initial_age : >= 103 s
Age served from cache     : 100 s     <- understated by the full response delay

The understatement scales with upstream latency, so it is largest exactly where caching
matters most.

Change

onRequestStart records the request time, and corrected_age_value adds the resulting
delay. The field is reset on every onRequestStart, so a retried or redirected request
measures its own delay rather than the first attempt's.

Tests

test/interceptors/cache-corrected-age.js, three cases:

• the served age includes the response delay
• a response without an Age header is aged by at most the response delay itself, since
  section 4.2.3 treats a missing Age as age_value = 0
• a response whose corrected age already exceeds max-age is not reused

Cases 1 and 3 fail on main and pass with this change. Case 2 passes on both and is there to
pin the bound on responses that carry no Age.

npm run test:cache-interceptor passes 76/76, and lint is clean.

[mcollina]

can you avoid using setTimeout in tests but use fake-timers instead? timers are flaky.

[RaphaelFakhri]

Done, switched to @sinonjs/fake-timers with toFake: ['Date'], same pattern as test/interceptors/cache.js. The response delay is now injected via clock.tick() inside the origin handler, so the assertions are exact values instead of lower bounds and nothing depends on real timers. The http2-request-never-settles timeout on Node 25 looks unrelated to this change (http2 test, pre-existing flake).