Low-friction, scriptable HTTP testing in the terminal. Inspired by Postman -- yet ultimately, very different.
- Save commonly-used HTTP requests as
.httpfiles and run them with:repost foo.http - Store shared variables in an environment file and reference them in your requests.
- Create a hook file to augment the request execution with custom JS code.
Warning: Repost is an experimental tool not intended for production usage. Use at your own risk.
In particular, do not execute untrusted requests, hooks or environments, as these can (by design) run arbitrary JS code on your computer.
# install repost CLI from Github
npm i -g luketurner/repost
echo "GET https://luketurner.org HTTP/1.1\n" > example.http
repost example.httpBasic Usage: repost FILE
When called with this form, repost will execute the HTTP request contained in the specified FILE.
Multiple files can be specified, in which case the requests will be executed sequentially in the specified order.
The file(s) must have a .http extension to be executed. Files with unknown extensions will be included in the list of results as a failure.
The content of the file(s) should be an HTTP request to execute, including:
- (required) Status line
- (optional) Headers
- (optional) Body
The headers and the body must be separated by an empty line, as in a true HTTP request:
POST https://example.com/foobar HTTP/1.1
Content-Type: text/plain
Hello, world!
Note on parsing HTTP requests
Within
repost, HTTP requests are not read directly from the file over the network. They're first parsed into a Javascript data structure, so Hooks can programmatically alter the contents of requests prior to execution.To convert the HTTP requests to JS, I wrote a naive parser. It works for simple requests, but is far from a true, compliant HTTP parser.
This means some HTTP features (e.g. Multipart, non-utf8 encodings) are not supported. Caveat emptor!
For more usage examples, see the example directory.
By default, the repost CLI outputs results using a concise, line-oriented format.
For each request that's run, a line is printed with the following information about the response:
runStatus httpStatusCode filename responseSize responseDuration
e.g.
> repost example.http
succeeded 200 example.http 345 72
This format is intended to work well with shell pipelining. For example, to execute all requests and only see the ones that failed, run:
repost *.http | grep ^failedRequests may contain embedded Javascript code that will be executed at runtime by the ejs templating engine. For example, you can use a <%= ... %> tag to inject a Javascript expression into your request:
POST https://example.com/foobar HTTP/1.1
Content-Type: text/plain
One plus one is <%= 1 + 1 %>!
See ejs docs for a list of all the tags you can use in your templates.
Embedded JS has access to all the variables and functions present in the Environment(s) that are configured.
This section is for those trying to make changes to Repost itself.
For starters, you need to clone the Git directory:
git clone https://github.com/luketurner/repost.gitThen install npm dependencies:
npm iRepost exposes its development commands as npm scripts. The following scripts are available:
npm run cli # runs the repost CLI
npm run debug # Runs the repost CLI with --inspect-brkCopyright 2021 Luke Turner - Published under the MIT License.