How to get current path with query string using Capybara

Question

The page url is something like /people?search=name while I used current_path method of capybara it returned /people only.

current_path.should == people_path(:search => 'name')

But it fails saying

expected: "/people?search=name"
got: "/people"

How we can make it pass? Is there is any way to do this?

`"/people?search=name"` [is not a path](http://en.wikipedia.org/wiki/URI_scheme#Generic_syntax). `"/people"` is a path — Andrei Botalov, May 10 '13 at 10:52

nzifnab · Accepted Answer · 2016-09-08T17:13:20.213

236

I've updated this answer to reflect modern conventions in capybara. I think this is ideal since this is the accepted answer, and what many people are being referred to when looking for a solution. With that said, the correct way to check the current path is to use the has_current_path? matcher provided by Capybara, as documented here: Click Here

Example usage:

expect(page).to have_current_path(people_path(search: 'name'))

As you can see in the documentation, other options are available. If the current page is /people?search=name but you only care that it's on the /people page regardless of the param, you can send the only_path option:

expect(page).to have_current_path(people_path, only_path: true)

Additionally, if you want to compare the entire URL:

expect(page).to have_current_path(people_url, url: true)

Credit to Tom Walpole for pointing out this method.

edited Sep 08 '16 at 17:13

answered Mar 08 '11 at 04:40

nzifnab

15,876
3
50
65

4

I'll probably need this syntax very soon, I'm writing tests for a legacy application. Using `current_path.should ==` is working for now (though I need to add a trailing forward-slash as a string). I'm thanking you in advance for code I'll likely need. – Tass Aug 19 '11 at 21:16
3

`URI.parse(current_url).request_uri` is more succinct. See @Lasse Bunk's answer. – Richard Jones Nov 27 '13 at 01:16
As of Capybara 2.5, this is no longer the best answer. See @tom-walpole's answer below. – dkniffin Aug 25 '16 at 15:43
Since the asker is unlikely to change the accepted answer, I've updated the answer to reflect modern times. This should be more helpful to new users looking for a solution and seeing the first answer. Thanks @OddityOverseer for pointing it out – nzifnab Sep 08 '16 at 17:16
5

For new Capybara versions use `ignore_query: true` instead of `only_path: true` – Alexander Mar 19 '20 at 14:55

score 94 · Answer 2 · answered Nov 20 '11 at 16:52

94

I replaced the _path method with _url to achieve comparing the full urls with parameters.

current_url.should == people_url(:search => 'name')

answered Nov 20 '11 at 16:52

Robert Starsi

941
6
2

4

How did you deal with the host part? – Chris Nicola Apr 11 '12 at 22:58
11

You can also use `current_path` in newer versions of Capybara and match it against `people_path(...)` – Jason Stirk Dec 21 '12 at 22:46
When I dont have named path... I have only /users/register... then how I should use it? – Gopal S Rathore Mar 07 '13 at 07:08
What if it's happening on a render, so that the URL is different from the html page? – bigpotato Feb 03 '14 at 19:55

Thomas Walpole · Answer 3 · 2018-07-20T22:51:11.087

56

Just updating this question for modern times. The current best practice for checking current_paths when using Capybara 2.5+ is to use the current_path matcher, which will use Capybaras waiting behavior to check the path. If wanting to check against the request_uri (path and query string)

expect(page).to have_current_path(people_path(:search => 'name'))

If only wanting the path part (ignoring the query string)

expect(page).to have_current_path(people_path, only_path: true) # Capybara < 2.16
expect(page).to have_current_path(people_path, ignore_query: true) # Capybara >= 2.16

If wanting to match the full url

expect(page).to have_current_path(people_url, url: true) # Capybara < 2.16
expect(page).to have_current_path(people_url) # Capybara >= 2.16

the matcher will take a string which is compared with == or a regex to match against

expect(page).to have_current_path(/search=name/)

edited Jul 20 '18 at 22:51

answered Nov 20 '15 at 16:26

Thomas Walpole

48,548
5
64
78

4

in these modern times this should be marked as *the* answer. This will safe a lot of vague timing issues, thanks! – Axe Mar 06 '16 at 12:12
1

@Vanuan rspec has been deprecating the `should` syntax. You should strive to use `expect().to` in your specs going forward. – nzifnab Sep 08 '16 at 17:17
1

`only_path: true` is now `ignore_query: true` – srghma Mar 13 '18 at 19:25

score 20 · Answer 4 · answered Oct 16 '13 at 22:34

I know an answer has been selected, but I just wanted to give an alternative solution. So:

To get the path and querystring, like request.fullpath in Rails, you can do:

URI.parse(current_url).request_uri.should == people_path(:search => 'name')

You can also do a helper method in your test class (like ActionDispatch::IntegrationTest) like this (which is what I did):

def current_fullpath
  URI.parse(current_url).request_uri
end

Hope this helps.

Ciro Santilli OurBigBook.com · Answer 5 · 2015-01-31T22:39:50.290

2

EDIT: as Tinynumberes mentioned, this fails for URLs with port number. Keeping it here in case anyone else has the same brilliant idea.

current_url[current_host.size..-1]

golfs exactly as well (35 chars) as URI.parse(current_url).request_uri, but is potentially faster because there is no explicit URI parsing involved.

I have made a pull request to add this to Capybara at: https://github.com/jnicklas/capybara/pull/1405

edited Jan 31 '15 at 22:39

answered Sep 23 '14 at 15:28

Ciro Santilli OurBigBook.com

347,512
102
1,199
985

This doesn't work if the `current_url` might contain a port number. E.g. given a `current_url` of `http://foo.com:8888/some/path`, `current_url[current_host.size..-1]` will equal `:8888/some/path`. Also, behind the scenes `current_host` does the same sort of `URI.parse` logic that @nzifnab recommended in the accepted answer. – Tinynumbers Jan 29 '15 at 14:36

score 0 · Answer 6 · answered Dec 19 '20 at 01:51

0

Beware that author mentioned in this comment. That usage of current_url can lead to flaky tests.

answered Dec 19 '20 at 01:51

woto

2,893
1
29
24

How to get current path with query string using Capybara

6 Answers6

Linked