Entry

Definition

A Hurl file is a list of entries, each entry being a mandatory request, optionally followed by a response.

Responses are not mandatory, a Hurl file consisting only of requests is perfectly valid. To sum up, responses can be used to capture values to perform subsequent requests, or add asserts to HTTP responses.

Example

# First, test home title.
GET https://acmecorp.net
HTTP 200
[Asserts]
xpath "normalize-space(//head/title)" == "Hello world!"

# Get some news, response description is optional
GET https://acmecorp.net/news

# Do a POST request without CSRF token and check
# that status code is Forbidden 403
POST https://acmecorp.net/contact
[Form]
default: false
email: [email protected]
number: 33611223344
HTTP 403

Description

Options

Options specified on the command line apply to every entry in an Hurl file. For instance, with --location option, every entry of a given file will follow redirection:

$ hurl --location foo.hurl

You can use an [Options] section to set option only for a specified request. For instance, in this Hurl file, the second entry will follow location (so we can test the status code to be 200 instead of 301).

GET https://google.fr
HTTP 301

GET https://google.fr
[Options]
location: true
HTTP 200

GET https://google.fr
HTTP 301

You can use the [Options] section to log a specific entry:

# ... previous entries

GET https://api.example.org
[Options]
very-verbose: true
HTTP 200

# ... next entries

Cookie storage

By default, requests in the same Hurl file share cookie storage, enabling session-based scenario. The shared cookie store can be disabled with [--no-cookie-store] option.

Redirects

By default, Hurl doesn't follow redirection. To effectively run a redirection, entries should describe each step of the redirection, allowing insertion of asserts in each response.

# First entry, test the redirection (status code and 'Location' header)
GET https://example.org
HTTP 301
Location: https://www.example.org

# Second entry, the 200 OK response
GET https://www.example.org
HTTP 200

Alternatively, one can use --location / --location-trusted options to force redirection to be followed. In this case, asserts are executed on the last received response. Optionally, the number of redirections can be limited with --max-redirs.

# Running hurl --location foo.hurl
GET https://example.org
HTTP 200

Finally, you can force redirection on a particular request with an [Options] section and the --location / --location-trusted options:

GET https://example.org
[Options]
location-trusted: true
HTTP 200

Redirections can be tested either by:

• running and asserting each step of redirection:

GET https://example.org/step1
HTTP 301
[Asserts]
header "Location" == "https://example.org/step2"


GET https://example.org/step2
HTTP 301
[Asserts]
header "Location" == "https://example.org/step3"


GET https://example.org/step3
HTTP 200

• using --location / --location-trusted, testing each step with redirects query:

GET https://example.org/step1
[Options]
location: true
HTTP 200
[Asserts]
redirects count == 2
redirects nth 0 location == "https://example.org/step2"
redirects nth 1 location == "https://example.org/step3"

url query can also be used to get the final effective URL:

GET https://example.org/step1
[Options]
location: true
HTTP 200
[Asserts]
url == "https://example.org/step3"

Retry

Every entry can be retried upon asserts, captures or runtime errors. Retries allow polling scenarios and effective runs under flaky conditions. Asserts can be explicit (with an [Asserts] section), or implicit (like headers or status code).

Retries can be set globally for every request (see --retry and --retry-interval), or activated on a particular request with an [Options] section.

For example, in this Hurl file, first we create a new job then we poll the new job until it's completed:

# Create a new job
POST http://api.example.org/jobs
HTTP 201
[Captures]
job_id: jsonpath "$.id"
[Asserts]
jsonpath "$.state" == "RUNNING"


# Pull job status until it is completed
GET http://api.example.org/jobs/{{job_id}}
[Options]
retry: 10   # maximum number of retry, -1 for unlimited
retry-interval: 300ms
HTTP 200
[Asserts]
jsonpath "$.state" == "COMPLETED"

Control flow

In [Options] section, skip and repeat can be used to control flow of execution:

• skip: true/false skip this request and execute the next one unconditionally,
• repeat: N loop the request N times. If there are assert or runtime errors, the requests execution is stopped.

# This request will be played exactly 3 times
GET https://example.org/foo
[Options]
repeat: 3
HTTP 200

# This request is skipped
GET https://example.org/foo
[Options]
skip: true
HTTP 200

Additionally, a delay can be inserted between requests, to add a delay before execution of a request (aka sleep).

# A 5 seconds delayed request 
GET https://example.org/foo
[Options]
delay: 5s
HTTP 200

delay and repeat can also be used globally as command line options:

$ hurl --delay 500ms --repeat 3 foo.hurl

For complete reference, below is a diagram for the executed entries.