(Tcl) curl Dependency Engine Multi-Step Execution Plan

See more CURL Examples

This example expands on the basic concept by demonstrating a multi-step execution plan, where several curl commands are chained together through their inputs and outputs.

Instead of resolving a single missing input with one helper command, the target curl command in this example depends on values that are produced by other commands, which may themselves depend on additional inputs. This creates a dependency chain that requires multiple steps to fully resolve.

When the DoYourThing method is called with the target curl command, the following occurs:

• Scans the target command for required inputs using the {{variable}} syntax
• Determines which inputs already have assigned values
• For any missing inputs:
  • Identifies a curl command that can produce the required value
  • Recursively evaluates that command to determine its own required inputs
  • Continues this process until all dependencies are resolved down to known values

Once all dependencies are identified, a multi-step execution plan is constructed. Each curl command is ordered so that it runs only after the values it depends on have been produced by earlier steps.

The execution plan is then carried out step by step, with intermediate results flowing between commands, and the final step being the target curl command originally passed to DoYourThing.

Note: This example requires Chilkat v11.5.0 or greater.

For more information, see https://www.chilkatsoft.com/curl_dependency_engine.asp

Chilkat Tcl Extension Downloads Chilkat Tcl Extension Downloads

load ./chilkat.dll

set success 0

set httpCurl [new_CkHttpCurl]

# The final curl command we want to execute.
# It depends on {{drive_id}}, which is not yet known.
set targetCurl "curl -X GET https://graph.microsoft.com/v1.0/drives/{{drive_id}}/root/children"

# Define a helper function that can produce drive_id.
# This function lists drives for a given site, which requires {{site_id}}.
set fnName "getDrives"
CkHttpCurl_AddFunction $httpCurl $fnName "curl -X GET https://graph.microsoft.com/v1.0/sites/{{site_id}}/drives"

# Define the output of getDrives.
# The JSON response contains an array named "value".
# We take the first element (value[0]) and extract its "id" as drive_id.
set jsonPath "value[0].id"
CkHttpCurl_AddOutput $httpCurl $fnName $jsonPath "drive_id"

# drive_id depends on site_id, which is also not yet known.
# Define another helper function to obtain site_id.
set fnName "getSite"
CkHttpCurl_AddFunction $httpCurl $fnName "GET https://graph.microsoft.com/v1.0/sites/root:/sites/{{site_name}}"

# Map the "id" field from the getSite response to the variable site_id.
set jsonPath "id"
CkHttpCurl_AddOutput $httpCurl $fnName $jsonPath "site_id"

# Provide the initial known input.
# site_name is the starting value that allows the dependency chain to be resolved.
CkHttpCurl_SetVar $httpCurl "site_name" "test"

# ----------------------------------------------------------------------------------------------------------------------
# Configure OAuth2 authentication using the client credentials flow.
# Secrets (client_id, client_secret, token_endpoint) are retrieved from the
# local secrets manager because EnableSecrets is enabled.
set jsonOAuth2 [new_CkJsonObject]

CkJsonObject_put_EnableSecrets $jsonOAuth2 1
CkJsonObject_UpdateString $jsonOAuth2 "oauth2.client_id" "!!sharepoint|oauth2|client_id"
CkJsonObject_UpdateString $jsonOAuth2 "oauth2.client_secret" "!!sharepoint|oauth2|client_secret"
CkJsonObject_UpdateString $jsonOAuth2 "oauth2.scope" "https://graph.microsoft.com/.default"
CkJsonObject_UpdateString $jsonOAuth2 "oauth2.token_endpoint" "!!sharepoint|oauth2|token_endpoint"
CkHttpCurl_SetAuth $httpCurl $jsonOAuth2
# ----------------------------------------------------------------------------------------------------------------------

# Execute the target curl command.
# Internally, the system will:
# 
# 1) Detect that {{drive_id}} is required but not defined.
# 2) Find that "getDrives" can produce drive_id.
# 3) Detect that getDrives requires {{site_id}}, which is also not defined.
# 4) Find that "getSite" can produce site_id.
# 5) Verify that getSite requires {{site_name}}, which is already known.
# 6) Build a multi-step execution plan:
#      - First:  run getSite (produces site_id)
#      - Second: run getDrives (produces drive_id)
#      - Third:  run the target curl (uses drive_id)
set success [CkHttpCurl_DoYourThing $httpCurl $targetCurl]
if {$success == 0} then {
    puts [CkHttpCurl_lastErrorText $httpCurl]
    delete_CkHttpCurl $httpCurl
    delete_CkJsonObject $jsonOAuth2
    exit
}

# Retrieve and display the HTTP response status code.
set statusCode [CkHttpCurl_get_StatusCode $httpCurl]
puts "response status code: $statusCode"

# Load and display the JSON response from the final curl command.
set responseJson [new_CkJsonObject]

CkJsonObject_put_EmitCompact $responseJson 0
CkHttpCurl_GetResponseJson $httpCurl $responseJson
puts [CkJsonObject_emit $responseJson]

delete_CkHttpCurl $httpCurl
delete_CkJsonObject $jsonOAuth2
delete_CkJsonObject $responseJson