Pratik Mankawde 399308b417 fix(telemetry): Require an https endpoint whenever use_tls=1 ...

The scheme guard only ran when tls_client_cert was set, so use_tls=1 against an
http endpoint booted: the CA file was checked for readability, then the exporter
read TLS off the URL scheme alone and sent spans in the clear. An operator who
asks for TLS should get TLS or a startup error.

Widen the guard to every use_tls=1 node. tls_client_cert already requires
use_tls=1, so this strictly widens the old condition. The error message and the
function's description now name use_tls rather than the certificate key.

one_way_tls_on_a_plain_http_endpoint_is_accepted pinned the old behaviour and is
flipped to _throws, asserting the scheme message, the endpoint key and the URL.
Three cases that sent use_tls=1 at an http endpoint while asserting a file error
now set an https endpoint, so the scheme guard cannot be what throws.

2026-09-23 14:29:22 +01:00

..

benchmarks/libxrpl

refactor: Replace boost::filesystem with std::filesystem across the codebase (#7012)

2026-08-12 13:40:39 +00:00

libxrpl

fix(telemetry): Require an https endpoint whenever use_tls=1

2026-09-23 14:29:22 +01:00

test

fix(rpc): Do not let span naming change the RPC error a client sees

2026-09-22 19:06:07 +01:00

tests

fix(telemetry): Require an https endpoint whenever use_tls=1

2026-09-23 14:29:22 +01:00

xrpld

Merge branch 'pratik/otel-phase4-consensus-tracing' into pratik/otel-phase5-docs-deployment

2026-09-23 13:57:47 +01:00