paultaykalo / cedarasync

asynchronous testing for Cedar (and others)

CedarAsync lets you use Cedar matchers to test asynchronous code. This becomes useful when writing intergration tests rather than plain unit tests. (CedarAsync only supports Cedar's should syntax.)

Instead of

client.lastResponse should contain(@"Google");

use

in_time(client.lastResponse) should contain(@"Google");

to force contain matcher check client.lastResponse multiple times until it succeeds or times out.

Example

#import "CedarAsync.h"
#import "HTTPClient.h"

SPEC_BEGIN(HTTPClientSpec)

using namespace Cedar::Matchers;

describe(@"HTTPClient", ^{
    __block HTTPClient *client;
    
    beforeEach(^{
        client = [[[HTTPClient alloc] init] autorelease];
    });
    
    it(@"can fetch google's homepage", ^{
        // uses NSURLRequest internally
        [client fetchURLString:@"http://google.com"];
        
        // plain Cedar matcher use - does not wait
        // (passes immediately since it takes sometime to fetch google)
        client.lastResponse should be_nil;
        
        // async matcher use - waits for lastResponse to contain 'Google'
        in_time(client.lastResponse) should contain(@"Google");
    });
});

SPEC_END

Timeout & Polling Interval

Temporarily change timeout and polling interval:

CedarAsync::Timing::current_timeout = 4; // seconds
CedarAsync::Timing::current_poll = 0.3;  // seconds

or

with_timeout(10, ^{
    in_time(valueThatTakesForever) should equal(@"so large...");
});

Change default timeout and polling interval (these values are used to populate current_timeout and current_poll before every test run):

CedarAsync::Timing::default_timeout = 4; // seconds
CedarAsync::Timing::default_poll = 2;    // seconds

Todo

• clean up queued up actions on NSRunLoop, GCD, etc. after every test