Runner now runs unfiltered; #error guard couples BASIC_AUTH to custom
transport configs; CMake include and REQUIRES cleanup; doc corrections and
tightened asserts from the final review.
Add max_write_chunk to the mock to force real short writes and fix its
capture ordering. Pin the open/write/fetch_headers/read sequence, short-write
retry, and the header read timeout (ESP_ERR_HTTP_FETCH_HEADER).
Enable BASIC_AUTH in both test configs (test-only). Pin the WWW-Authenticate
retry path, the credential-less 401 retry capped by max_authorization_retries,
and the independently gated preconfigured Basic auth.
Pin auto-follow inside one perform() call, disable_auto_redirect surfacing
the 302, and max_redirection_count tripping ESP_ERR_HTTP_MAX_REDIRECT (the
counter is checked before each redirect).
Add read_bytes_before_error to the mock (read-side truncation budget) so the
FIN lands inside a chunk body; pins decode output and
ESP_ERR_HTTP_INCOMPLETE_DATA on truncation.
Pin transport reuse across perform() calls, the Connection: close reconnect,
and the exact get_state() sequence seen from the event handler (the getter
lags its triggering events).
Pin async connect and read would-block retry paths, and that a would-block
on the initial header send aborts the request (close + ESP_ERR_HTTP_WRITE_DATA)
instead of retrying.
Mock TCP transport plus 13 basic cases that pin current client behavior.
Fixture Content-Length values corrected; all mock-based cases compile only
with CONFIG_ESP_HTTP_CLIENT_ENABLE_CUSTOM_TRANSPORT.
A blocking transport read can return zero when no data arrives before
timeout. Passing that zero length to http_parser_execute signals EOF while
a response is incomplete, puts the parser in HPE_INVALID_EOF_STATE, and
causes later response bytes to be discarded. The shortened response can
then be treated as successful.
Skip parser execution for every zero-length transport read, not only async
reads. Also compare the raw esp_transport_read result against raw
ERR_TCP_TRANSPORT values so timeout and peer-close failures retain their
documented HTTP error classifications.
A standalone reproduction built with the unmodified HTTP parser showed the
blocking timeout transition to HPE_INVALID_EOF_STATE and loss of the
remaining 15 bytes. Skipping the zero-length parser call delivered the full
chunk and message-complete callback.
Disclosure: this fix was prepared with AI assistance (Claude) and reviewed by me before submission.
Constraint: esp_http_client_get_data returns raw transport result values before esp_transport_translate_error.
Rejected: Keep the async-only zero-length guard | blocking transport reads also return zero on timeout.
Confidence: high
Scope-risk: moderate
Directive: Do not pass a transient zero-length transport read to the HTTP parser as EOF.
Tested: standalone blocking mid-chunk timeout reproduction; source-only duplicate-comment cleanup; git diff --check
Not-tested: hardware TLS transport integration
Signed-off-by: yi chen <94xhn1@gmail.com>
When CONFIG_ESP_HTTP_CLIENT_STRICT_HEADER_BUFFER is enabled,
esp_http_client_request_send() fails fast when a single request header is
larger than buffer_size_tx, instead of silently emitting a truncated request.
The condition is reported with a dedicated ESP_ERR_HTTP_HEADER_TOO_LONG error
code rather than overloading ESP_ERR_HTTP_WRITE_DATA, so callers can tell a
permanent header-sizing failure apart from a transient write error. errno is
cleared on this path so the async caller (which maps errno == EAGAIN to "retry
later") does not spin retrying a request that can never succeed.
Closes https://github.com/espressif/esp-idf/issues/18639
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Add ESP_KEY_SOURCE_BUFFER and ESP_KEY_SOURCE_PSA key sources so all
hardware backends (DS, ECDSA, secure element) are accessed via PSA
key IDs through a single esp_tls_cfg_t.client_key field.
Refactor the esp_err_to_name() system to decouple esp_common from
higher-level components. Instead of a monolithic generated table,
each component registers its error codes into a dedicated linker
section (.esp_err_msg_table) via idf_define_esp_err_codes() in its
CMakeLists.txt.
New files:
- tools/err_codes_extract.py: extract ESP_ERR_* defines from headers to CSV
- tools/err_codes_to_c.py: generate C source placing entries into linker section
- tools/err_codes_to_rst.py: generate RST documentation from error codes
- tools/cmake/err_codes.cmake: CMake module providing idf_define_esp_err_codes()
- components/esp_common/include/esp_err_codes.h: esp_err_msg_t typedef
- components/esp_common/src/esp_err_to_name_new.c: new lookup using link-time array
- tools/test_apps/build_system/err_codes_check/: CI test app
Changes:
- Remove all optional component dependencies from esp_common/CMakeLists.txt
- Add .esp_err_msg_table section to all 5 linker scripts
- Register error codes in 18 components via idf_define_esp_err_codes()
- Add new scripts to .gitlab/ci/rules.yml build_check patterns
- use new scripts to generate doc and add CI validation
- Update esp_err.rst to add description of composable code registration
esp_http_client_set_redirection() now rejects any redirect target whose
scheme is not https:// when the origin is HTTPS. This catches http, ftp,
ws and any other scheme before client state is mutated. Same-host /
https-to-https redirects are unaffected. Apps that intentionally want
mixed-scheme redirects can set disable_auto_redirect=true and handle
HTTP_EVENT_REDIRECT.
According to RFC 7616, nonce-prime and cnonce-prime is used for SHA-256-sess only and not for SHA-256.
This commit updates the check and uses nonce only for "-sess" algorithms.
Regression from 66995965e7
Fixed vulnerability where malicious HTTP servers could trigger OOB reads
by sending empty or very short algorithm fields in WWW-Authenticate headers.
Changes:
- Replace unsafe memcmp() with strcasecmp() for algorithm comparison
- Add algorithm NULL validation at function entry point
- Fix duplicate md5-sess check, add missing SHA-256 check
