01:30 <zan> hi, i'm encountering a prb where all my steps pass, but the scenario fail

01:31 <zan> is there a clue as to how to debug this?

01:56 <GitHub191> cucumber: 07master 03aslakhellesoy * 3897989 (9 files in 4 dirs): Bugfix: Tab indentation causes multiline step arguments to fail [#585 state:resolved] - http://bit.ly/b3NaHO

02:04 <GitHub169> cucumber: 07master 03aslakhellesoy * 4f2b799 (5 files in 1 dirs): Remove duplication - http://bit.ly/bUZk6O

02:11 <GitHub45> cucumber: 07master 03aslakhellesoy * 717780e (2 files in 2 dirs): Attempt to fix Cucumber::CODEPAGE error again for Windows [#561 state:resolved] - http://bit.ly/cpNY0g

02:23 <GitHub187> cucumber: 07master 03aslakhellesoy * de747cf (3 files in 3 dirs): #announce can now take non-String arguments just like Kernel#puts - #to_s is done implicitly. - http://bit.ly/coZLML

03:03 <Flydoire|Taktik> hi everyone

03:04 <Flydoire|Taktik> I have a problem with Unix symbolic links on "features" folder -> they are not followed

03:04 <Flydoire|Taktik> where do I report bugs on cucumber ?

03:08 <josephwilk> Flydoire|Taktik: https://rspec.lighthouseapp.com/projects/16211-cucumber/overview

03:09 <Flydoire|Taktik> thanks

03:44 <tils> hello everybody

03:44 <tils> does anyone know how to find out and change the browser version used by culerity/celerity/htmlunit?

03:45 <jarib> tils: Celerity::Browser.new(:browser => :firefox)

03:45 <jarib> (check the docs for alternatives)

03:45 <tils> jarib: ok thanks

03:48 <* tils> searches culerity code where it sets the browser options

06:07 <matthewford> can webdriver-selenium run headless, so my cucumber specs run on an intergration server ?

06:23 <hka> Matthew - Yes, you can use the HtmlUnit driver. This does, however, have a few limitations regarding javascript support.

06:31 <jarib> matthewford: either htmlunit (through the remote driver) or xvfb/vnc + firefox

06:33 <hka> You shouldn't need remote webdriver to run HtmlUnit. It can run directly from the build script on you ci server.

06:34 <jarib> hka: not with the selenium-webdriver gem

06:35 <jarib> you need java somewhere for htmlunit

06:35 <matthewford> jarib, thanks

06:36 <jarib> np

06:37 <matthewford> anyone tried steam, it seems to use HtmlUnit

07:08 <* Pistos> reasks: Does anyone here use Pivotal Tracker? :)

07:13 <josephwilk> Pistos Used to

07:14 <Pistos> josephwilk: What did you think of it? And why did you stop? :)

07:15 <josephwilk> Pistos Its great if you want an eletronic system, I stopped as I prefer a realy board with cards.

07:15 <josephwilk> *real

07:18 <Pistos> josephwilk: Okay, thanks for your comments.

07:18 <Pistos> I'm test driving it for my team. Feels like a great project tracker for development teams.

08:37 <malkomalko> hey all, another cuke4duke question... just running up against this right now on multiple machines

08:37 <malkomalko> when running mvn -Dcucumber.installGems=true cuke4duke:cucumber... it seems like I'm getting this stacktrace message: No such file or directory - File not found - null (Errno::ENOENT)

08:37 <malkomalko> with the failing line at : /Users/pair/.m2/repository/.jruby/gems/cucumber-0.6.3/lib/cucumber/feature_file.rb:36:in `initialize'

08:37 <malkomalko> this happens after the gems are installed.. I tried this with multiple ruby versions on a few different machines and I'm getting the same results.. any idea?

08:40 <msassak> malkomalko does this happen when you run cucumber by hand and pass in only the path to a feature?

08:40 <malkomalko> this is with mvn or running it by hand from inside the .m2 repo

08:41 <msassak> that's weird... you're sure the path to the feature is correct?

08:42 <malkomalko> pretty sure yes.. features folder in the root of the path I'm running the mvn command from

08:42 <msassak> what happens if you invoke cucumber with only the name of a single feature?

08:42 <msassak> like "cucumber features/my.feature"? does it read it?

08:43 <malkomalko> let me try that

08:43 <malkomalko> ~/.m2/repository/.jruby/bin/cucumber features/pull_campaign_entities.feature

08:43 <malkomalko> env: jruby: No such file or directory

08:44 <msassak> err sorry try "jruby -S cucumber features/my.feature"

08:48 <malkomalko> i have to install it through jruby.. I wasn't using normal cucumber.. I was using cuke4duke

08:48 <malkomalko> one sec

08:49 <malkomalko> that runs cucumber as though it's looking for ruby, not java

08:51 <msassak> the reason i ask is because that error is usually a pretty straightforward "That feature file doesn't exist"--it doesn't have much to do with anything in java land

08:57 <malkomalko> aight, thanks anyway

09:03 <preppy> has anyone worked with lists in cucumber before?

09:04 <msassak> preppy what do you mean by that?

09:12 <preppy> well lets say i'm trying to write a cucumber story around a RESTful model

09:12 <preppy> the index page is usually a list of AR's

09:12 <msassak> ok

09:12 <preppy> how would i structure a scenario to say that "I follow 'link' for 'row_item'"

09:14 <msassak> you can select specific things to follow using xpath or css selectors

09:15 <preppy> could you give me an example?

09:15 <msassak> that's the general solution--if you look in the generated web_steps.rb in features/step_definitions, you can see how it is used

09:15 <msassak> basically, you specify a selector that narrows the scope down

09:16 <preppy> so use "follow 'link' within 'xpath'"

09:16 <msassak> like within("#active_record_model_1") do click_link("View") end

09:16 <msassak> yup

09:16 <preppy> hmmm...

09:16 <preppy> okay i'll try taht

09:16 <preppy> thanks!

09:17 <msassak> what feature are you driving out that requires this?

09:17 <preppy> updating an item from the list

09:17 <preppy> i need to click the link to edit

09:17 <msassak> (in my experience verifying that an index page has a bunch of items on it isn't usually where the value is)

09:18 <msassak> in that case I would go straight to the edit page

09:18 <preppy> and skip the page where you click from?

09:20 <msassak> if you scenario revolves around editing a single item, start at that item

09:21 <msassak> it all depends on what it is you are developing with that scenario

09:22 <msassak> the question is whether the scenario is about showing that items in a collection can be navigated to from the collection listing, or that items can be edited and the changes saved

09:23 <preppy> well i have a scenario for a single item already

09:24 <preppy> so this scenario is for the first part of your statement

09:24 <msassak> okay cool :-)

09:24 <preppy> yea my original scenario was for one item is the user's list

09:24 <preppy> and this one creates 10 items

09:25 <preppy> so i'll try out the "within" and see if i can get the link to work

09:25 <preppy> thanks!

09:26 <msassak> np

11:06 <lucasbz> debugging in cucumber, how do I do that?

11:06 <lucasbz> anyone?

13:59 <matt2s> I have an external service to the web app I'm testing. Do you guys mock it when running cucumber, or test through the complete stack?

14:00 <josephwilk> Initally against the service, but using that to provide examples to create a fake service.

14:01 <josephwilk> So not mocking as such, just making the service hotswopeable for a fake version

14:06 <matt2s> I see, I'll think a bit on how I should do this. It's just xml-rpc calls to a server listening, so it should be fairly trivial to "fake" it

14:07 <matt2s> I was thinking of abstracting the entire service behind an interface / adapter and just mock the adapter or something. Then I can create more specialized tests for the adapter afterwards or something

14:07 <msassak> aslakhellesoy got a minute?

14:07 <aslakhellesoy> msassak: of course

14:08 <msassak> i'm ready to go ahead with the changes to gherkin

14:08 <msassak> but i wanted to talk a bit about the builder interface with cucumber to make sure i'm thinking the same as you

14:09 <aslakhellesoy> msassak: sounds great. what are you wondering about?

14:09 <msassak> i imagine that when we go to one event per row, the builder in cucumber will need to cache rows like it does with tags, and then pass that into Table.new, is that correct?

14:11 <aslakhellesoy> msassak: that was my thinking yes

14:11 <aslakhellesoy> and on the way, comments can happen too

14:11 <msassak> yup

14:11 <msassak> what do you think about table_start and table_end messages?

14:11 <aslakhellesoy> so i think we'll create row objects - with an optional comment hanging on top

14:12 <aslakhellesoy> and a table consists of rows

14:12 <aslakhellesoy> i'm not sure we need a table_start and table_end

14:12 <aslakhellesoy> would be nice to do without i think - simpler api?

14:13 <msassak> greg thought that, too, and now i'm leaning in that direction

14:13 <msassak> i think this will make the ragel much simpler, actually

14:13 <msassak> so that's good

14:13 <aslakhellesoy> great. let's try that first

14:13 <aslakhellesoy> even better

14:13 <msassak> will do

14:14 <aslakhellesoy> anything else?

14:14 <msassak> nope!

14:15 <aslakhellesoy> ok, i have something

14:15 <msassak> ok :-)

14:15 <aslakhellesoy> refactoring strategy

14:15 <aslakhellesoy> or: what to do to marry gherkin with cucumber

14:15 <aslakhellesoy> we've explored a lot of ideas on the gherkin branch

14:16 <msassak> yeah

14:16 <aslakhellesoy> and i'm thinking we're trying to bite over too much at the same time

14:16 <aslakhellesoy> new parser, new ast, plugins etc

14:16 <aslakhellesoy> so here is an idea i'd like to try

14:16 <aslakhellesoy> 1) refactor master (treetop) to behave like the current gherkin

14:17 <aslakhellesoy> leave the ast mostly unchanged

14:17 <aslakhellesoy> 2) at this point swapping in gherkin will be rather easy - and an isolated task

14:17 <aslakhellesoy> when this works we can ditch treetop, but we still have the current ast

14:18 <aslakhellesoy> once that's done we can work with a new model

14:18 <msassak> that sounds like a good idea

14:18 <aslakhellesoy> and the goal of a new model in my view is:

14:18 <aslakhellesoy> 1) a simpler formatter api

14:19 <aslakhellesoy> 2) less bugs in the whole model lifecycle - we have some funky bugs related to backgrounds, hooks etc.

14:19 <aslakhellesoy> and in general a simpler and better model

14:19 <aslakhellesoy> however, a fast parser has higher pri than a clean model

14:19 <aslakhellesoy> and we can work on them once at a time

14:20 <aslakhellesoy> in smaller chunks :-)

14:20 <aslakhellesoy> are you with me?

14:20 <msassak> that sounds very reasonable

14:20 <msassak> most definitely with you

14:20 <aslakhellesoy> great

14:20 <msassak> it's been a long time, and i think getting something out the door will be good for everyone's spirits

14:20 <aslakhellesoy> so when we have gherkin in, we can pick the good parts from the gherkin branch' experiments

14:20 <aslakhellesoy> yes, absolutely

14:21 <aslakhellesoy> i need something small and achievable to attack

14:21 <msassak> well, refactoring the treetop API is a good place to start

14:21 <aslakhellesoy> yes. haven't thought about how exactly

14:21 <aslakhellesoy> logic will have to move out of the parser to somewhere else

14:22 <aslakhellesoy> probably some kind of builder that sits between the parser and the ast

14:22 <aslakhellesoy> -like we're planning to do with gherkin

14:22 <aslakhellesoy> yeah - a builder that has the gherkin listener api

14:22 <aslakhellesoy> but builds the current master ast

14:22 <msassak> i'd have to look at the gherkin branch closely again, but i don't think that building the current ast would be *that* hard

14:22 <msassak> there would be a lot of caching of things, if i remember

14:23 <msassak> but you can depend on the well-orderedness of the messages, so that helps a lot

14:23 <aslakhellesoy> i may have some time to work on this tomorrow night and sunday

14:24 <aslakhellesoy> feel free to start if you have time

14:24 <msassak> okay

14:24 <msassak> i'll probably work on gherkin first

14:24 <aslakhellesoy> but tbh it would be great if you could work on the gherkin refactoring so that stabilises

14:24 <aslakhellesoy> yeah

14:24 <msassak> that's my focus at the moment, yes

14:25 <msassak> what time is tomorrow night for you?

14:25 <aslakhellesoy> then we should make new releases and upload .net .dlls to github's new file upload section

14:25 <aslakhellesoy> i'm 6 hours ahead of you

14:25 <aslakhellesoy> you're eastern time right?

14:25 <msassak> yup

14:25 <aslakhellesoy> so 9ish my time and 3-4 hours

14:25 <msassak> i mean, when you say "tomorrow night", do you mean 10PM your time, 8PM...?

14:25 <msassak> ahh okay

14:26 <msassak> well if you start at 9 tomorrow, I can be around for a few hours at least

14:26 <aslakhellesoy> with the possibility of getting caught up with other stuff, but i'll resist :-)

14:26 <msassak> if you need someone to bounce ideas off of

14:26 <aslakhellesoy> let's just see - if i'm online i'm online :-)

14:26 <aslakhellesoy> yes - need a pair

14:27 <aslakhellesoy> or at least someone to discuss with

14:27 <aslakhellesoy> i'm a little demoralised by the many failed attempts to integrate gherkin

14:27 <aslakhellesoy> but pretty sure this will work smoothly

14:27 <msassak> same here

14:28 <msassak> there were too many moving parts at once

14:28 <aslakhellesoy> we never learn do we

14:28 <msassak> and i think it points to the need for some stable interfaces deeper inside cucumber

14:28 <msassak> :-)

14:28 <aslakhellesoy> gotta go review the peepcode screencast

14:29 <aslakhellesoy> it's slick!

14:29 <msassak> i watched the first 10 minutes or so

14:29 <msassak> peepcode is always so impressive

14:29 <aslakhellesoy> extremely well put together

14:29 <msassak> definitely

14:30 <aslakhellesoy> kk - catch you tomorrow hopefully

14:30 <msassak> sounds good

14:30 <msassak> ttyl