HTTP and HTTPS example application for Mbed OS 5
Dependencies: mbed-http
This application demonstrates how to make HTTP and HTTPS requests and parse the response from Mbed OS 5.
It consists of six example applications, which you can select in source/select-demo.h
:
- HTTP:
- Does a GET request to http:httpbin.org/status/418.
- Does a POST request to http:httpbin.org/post.
- HTTPS:
- Does a GET request to https:os.mbed.com/media/uploads/mbed_official/hello.txt.
- Does a POST request to https:httpbin.org/post.
- HTTP with socket re-use.
- HTTPS with socket re-use.
- HTTP over IPv6.
- HTTPS with chunked requests.
Response parsing is done through nodejs/http-parser.
Note: HTTPS requests do not work on targets with less than 128K of RAM due to the size of the TLS handshake. For more background see mbed-http.
To build
- If you're using WiFi, specify the credentials in
mbed_app.json
. - Build the project in the online compiler or using Mbed CLI.
- Flash the project to your development board.
- Attach a serial monitor to your board to see the debug messages.
Defining the network interface
This application uses the on-board network interface for your board. If you use an external network interface (f.e. a WiFi module) you need to add the driver to this project. Then, open network-helper.h
and specify which network driver to use.
More information is in the Mbed OS documentation under IP Networking.
Entropy (or lack thereof)
On all platforms that do not have the TRNG
feature, the application is compiled without TLS entropy sources. This means that your code is inherently unsafe and should not be deployed to any production systems. To enable entropy, remove the MBEDTLS_NO_DEFAULT_ENTROPY_SOURCES
and MBEDTLS_TEST_NULL_ENTROPY
macros from mbed_app.json.
Flash size
Default flash size for HTTPS is very large, as the application is loading the default Mbed TLS configuration. To use a more optimized version, you can disable unused cypher suites and other Mbed TLS features with a custom configuration file. Create a new configuration file, then add in mbed_app.json
:
"MBEDTLS_CONFIG_FILE=\"mbedtls_config.h\""
to the macros
array.
Running tests
You can run the integration tests from this project via Mbed CLI.
- In
select-demo.h
set theDEMO
macro toDEMO_TESTS
. - Set your WiFi credentials in
mbed_app.json
. - Then run the tests via:
$ mbed test -v -n mbed-http-tests-tests-*
Tested on
- K64F with Ethernet.
- NUCLEO_F411RE with ESP8266 (not working on Mbed OS 5.12+)
- ODIN-W2 with WiFi.
- K64F with Atmel 6LoWPAN shield.
- DISCO-L475VG-IOT01A with WiFi (requires the wifi-ism43362 driver).
source/main-https-chunked-request.cpp
- Committer:
- Jan Jongboom
- Date:
- 2019-01-04
- Revision:
- 35:4b847971db1b
- Parent:
- 31:66704f6f17c5
File content as of revision 35:4b847971db1b:
/** * This is an example of doing chunked requests, where you do not need to load the full request body * into memory. You do this by adding a callback to the `send` function of the HTTP/HTTPS request. */ #include "select-demo.h" #if DEMO == DEMO_HTTPS_CHUNKED_REQUEST #include "mbed.h" #include "mbed_trace.h" #include "https_request.h" #include "network-helper.h" /* List of trusted root CA certificates * currently one: Comodo, the CA for reqres.in * * To add more root certificates, just concatenate them. */ const char SSL_CA_PEM[] = "-----BEGIN CERTIFICATE-----\n" "MIICiTCCAg+gAwIBAgIQH0evqmIAcFBUTAGem2OZKjAKBggqhkjOPQQDAzCBhTEL\n" "MAkGA1UEBhMCR0IxGzAZBgNVBAgTEkdyZWF0ZXIgTWFuY2hlc3RlcjEQMA4GA1UE\n" "BxMHU2FsZm9yZDEaMBgGA1UEChMRQ09NT0RPIENBIExpbWl0ZWQxKzApBgNVBAMT\n" "IkNPTU9ETyBFQ0MgQ2VydGlmaWNhdGlvbiBBdXRob3JpdHkwHhcNMDgwMzA2MDAw\n" "MDAwWhcNMzgwMTE4MjM1OTU5WjCBhTELMAkGA1UEBhMCR0IxGzAZBgNVBAgTEkdy\n" "ZWF0ZXIgTWFuY2hlc3RlcjEQMA4GA1UEBxMHU2FsZm9yZDEaMBgGA1UEChMRQ09N\n" "T0RPIENBIExpbWl0ZWQxKzApBgNVBAMTIkNPTU9ETyBFQ0MgQ2VydGlmaWNhdGlv\n" "biBBdXRob3JpdHkwdjAQBgcqhkjOPQIBBgUrgQQAIgNiAAQDR3svdcmCFYX7deSR\n" "FtSrYpn1PlILBs5BAH+X4QokPB0BBO490o0JlwzgdeT6+3eKKvUDYEs2ixYjFq0J\n" "cfRK9ChQtP6IHG4/bC8vCVlbpVsLM5niwz2J+Wos77LTBumjQjBAMB0GA1UdDgQW\n" "BBR1cacZSBm8nZ3qQUfflMRId5nTeTAOBgNVHQ8BAf8EBAMCAQYwDwYDVR0TAQH/\n" "BAUwAwEB/zAKBggqhkjOPQQDAwNoADBlAjEA7wNbeqy3eApyt4jf/7VGFAkK+qDm\n" "fQjGGoe9GKhzvSbKYAydzpmfz1wPMOG+FDHqAjAU9JM8SaczepBGR7NjfRObTrdv\n" "GDeAU/7dIOA1mjbRxwG55tzd8/8dLDoWV9mSOdY=\n" "-----END CERTIFICATE-----\n"; void dump_response(HttpResponse* res) { printf("Status: %d - %s\n", res->get_status_code(), res->get_status_message().c_str()); printf("Headers:\n"); for (size_t ix = 0; ix < res->get_headers_length(); ix++) { printf("\t%s: %s\n", res->get_headers_fields()[ix]->c_str(), res->get_headers_values()[ix]->c_str()); } printf("\nBody (%lu bytes):\n\n%s\n", res->get_body_length(), res->get_body_as_string().c_str()); } // Spread the message out over 3 different chunks const char * chunks[] = { "{\"message\":", "\"this is an example", " of chunked encoding\"}" }; int chunk_ix = 0; // Callback function, grab the next chunk and return it const void * get_chunk(uint32_t* out_size) { // If you don't have any data left, set out_size to 0 and return a null pointer if (chunk_ix == (sizeof(chunks) / sizeof(chunks[0]))) { *out_size = 0; return NULL; } const char *chunk = chunks[chunk_ix]; *out_size = strlen(chunk); chunk_ix++; return chunk; } int main() { NetworkInterface* network = connect_to_default_network_interface(); if (!network) { printf("Cannot connect to the network, see serial output\n"); return 1; } mbed_trace_init(); // This example also logs the raw request, you can do this by calling 'set_request_log_buffer' on the request uint8_t *request_buffer = (uint8_t*)calloc(2048, 1); // POST request to reqres.in { HttpsRequest* post_req = new HttpsRequest(network, SSL_CA_PEM, HTTP_POST, "https://reqres.in/api/users"); post_req->set_header("Content-Type", "application/json"); post_req->set_request_log_buffer(request_buffer, 2048); // If you pass a callback here, the Transfer-Encoding header is automatically set to chunked HttpResponse* post_res = post_req->send(&get_chunk); if (!post_res) { printf("HttpsRequest failed (error code %d)\n", post_req->get_error()); return 1; } // Log the raw request that went over the line (if you decode the hex you can see the chunked parts) // e.g. in Node.js (take the output from below): // '50 4f 53 54 20'.split(' ').map(c=>parseInt(c,16)).map(c=>String.fromCharCode(c)).join('') printf("\n----- Request buffer -----\n"); for (size_t ix = 0; ix < post_req->get_request_log_buffer_length(); ix++) { printf("%02x ", request_buffer[ix]); } printf("\n"); printf("\n----- HTTPS POST response -----\n"); dump_response(post_res); delete post_req; } wait(osWaitForever); } #endif