The WebSocket echo server tests connect immediately after seeing
'Starting server on port:' in the device log, but URI handlers
(/ws, /auth) are registered asynchronously after the server starts,
taking 40-660ms depending on config and CI load.
This causes two failures:
1. WebSocket handshake returns 404 Not Found because the URI
handler is not registered yet when the client connects.
2. WebSocket echo returns wrong data because the server is in a
partially initialized state.
Wait for 'Returned from app_main()' before connecting, which
guarantees all URI handlers are registered. Add connection retry
with WebSocketBadStatusException handling to WsClient.
Extract _wait_for_server_ready() helper to deduplicate the WiFi
credential input and server readiness logic.
(cherry picked from commit 7a50e3ab04)
fix(examples): make https_request rom_impl CI test resilient to transient external server failures (backport v5.5)
See merge request espressif/esp-idf!48417
The binary size (0x100110) slightly exceeds the default 1MB factory
partition, causing a build failure with 0x110 bytes overflow.
Use CONFIG_PARTITION_TABLE_SINGLE_APP_LARGE to provide sufficient
space, consistent with other TLS-based mqtt examples.
After startup is done and app_main is reached, all calls to
heap_caps_add_region_with_caps (adding a new heap to the list of registered
heaps) will not affect the minimum free heap size value.
For heaps created before startup, this won't apply and their size will
end up being taken into account in the calculation of the minimum.
The esptool version bundled with the release/v5.5 IDF environment only
accepts the underscore subcommand name 'write_flash'. The hyphenated
alias 'write-flash' is rejected by argparse on this esptool version.
Use the canonical 'write_flash' name accepted by all esptool versions.
The write_time_to_nvs helper is introduced by parent MR !47528's sibling on master
(!47201). On release/v5.5 the parent hasn't merged yet, so this MR's rom_impl test
resilience patch references an undefined name. Include the helper inline here so
pre-commit (ruff F821 + mypy) passes; if the parent's backport (!48415/!48416)
lands first, the merge will need a trivial dedup of this helper.
The test_examples_protocol_https_request_rom_impl test on ESP32-C2 with
the ROM mbedTLS implementation hits a public TLS endpoint
(tls13.browserleaks.com / www.howsmyssl.com) for five sequential
sub-tests. CI runs frequently fail mid-test with one of:
* mbedtls_ssl_handshake returned -0x7780
(MBEDTLS_ERR_SSL_FATAL_ALERT_MESSAGE - server-side fatal alert)
* esp-tls: [sock=NN] select() timeout
* esp-tls: Failed to open new connection in specified timeout
In a representative log the first three sub-tests succeed and only the
fourth/fifth fail, which is consistent with rate-limiting or transient
server unavailability rather than a device bug. Mark the test as
flaky with up to two reruns (10 s delay) so a single transient remote
failure does not fail the CI job.
Also align this test with the rest of the file by injecting the host
timestamp via NVS (write_time_to_nvs) so the device clock no longer
depends on a previous boot's persisted SNTP result, removing one more
source of CI flakiness on first-boot or after NVS erase.
The test_examples_protocol_https_request_rom_impl test on ESP32-C2 with
the ROM mbedTLS implementation hits a public TLS endpoint
(tls13.browserleaks.com / www.howsmyssl.com) for five sequential
sub-tests. CI runs frequently fail mid-test with one of:
* mbedtls_ssl_handshake returned -0x7780
(MBEDTLS_ERR_SSL_FATAL_ALERT_MESSAGE - server-side fatal alert)
* esp-tls: [sock=NN] select() timeout
* esp-tls: Failed to open new connection in specified timeout
In a representative log the first three sub-tests succeed and only the
fourth/fifth fail, which is consistent with rate-limiting or transient
server unavailability rather than a device bug. Mark the test as
flaky with up to two reruns (10 s delay) so a single transient remote
failure does not fail the CI job.
Also align this test with the rest of the file by injecting the host
timestamp via NVS (write_time_to_nvs) so the device clock no longer
depends on a previous boot's persisted SNTP result, removing one more
source of CI flakiness on first-boot or after NVS erase.
The esptool version bundled with the release/v5.5 IDF environment only
accepts the underscore subcommand name 'write_flash'. The hyphenated
alias 'write-flash' was added in later esptool releases, so the CI
target test failed with 'argparse.ArgumentError: invalid choice:
write-flash'. Use the canonical 'write_flash' name, which is accepted
by all esptool versions.
