Smalltalk › Amber Smalltalk

Programmatically running tests

_‹ Previous Topic Next Topic _›

Classic

List

Threaded

7 messages Options

Bernat Romagosa

Programmatically running tests

Hi list,

How does one run a test and gather the results programmatically?

I'm trying the ways I know from GST but it looks a bit different, for instance you can't do:

(MyTestCase selector: #mySelectorForATest) run.

Cheers!

--
Bernat Romagosa.

Bernat Romagosa

Re: Programmatically running tests

And on the same topic... what do tests in red indicate in the test runner? I noticed I can't debug methods that show up red, but I can debug the ones in black.

2011/10/10 Bernat Romagosa <[hidden email]>

Hi list,

How does one run a test and gather the results programmatically?

I'm trying the ways I know from GST but it looks a bit different, for instance you can't do:

(MyTestCase selector: #mySelectorForATest) run.

Cheers!

--
Bernat Romagosa.

--
Bernat Romagosa.

Nicolas Petton

Re: Programmatically running tests

On Mon, 2011-10-10 at 17:14 +0200, Bernat Romagosa wrote:
> And on the same topic... what do tests in red indicate in the test
> runner? I noticed I can't debug methods that show up red, but I can
> debug the ones in black.

In black? I thought it was:
green -> passed
yellow -> failed
red -> error

Nico

>
> 2011/10/10 Bernat Romagosa <[hidden email]>
> Hi list,
>
>
> How does one run a test and gather the results
> programmatically?
>
>
> I'm trying the ways I know from GST but it looks a bit
> different, for instance you can't do:
>
>
> (MyTestCase selector: #mySelectorForATest) run.
>
>
> Cheers!
>
>
> --
> Bernat Romagosa.
>
>
>
>
>
> --
> Bernat Romagosa.

Nicolas Petton

Re: Programmatically running tests

In reply to this post by Bernat Romagosa

On Mon, 2011-10-10 at 16:49 +0200, Bernat Romagosa wrote:
> Hi list,
>
>
> How does one run a test and gather the results programmatically?

No idea :) I'll have a look.

Nico

>
>
> I'm trying the ways I know from GST but it looks a bit different, for
> instance you can't do:
>
>
> (MyTestCase selector: #mySelectorForATest) run.
>
>
> Cheers!
>
>
> --
> Bernat Romagosa.
>

Bernat Romagosa

Re: Programmatically running tests

That was weird, now I see them red and I can inspect the results. I guess I messed something up yesterday... :P

2011/10/11 Nicolas Petton <[hidden email]>

On Mon, 2011-10-10 at 16:49 +0200, Bernat Romagosa wrote:
> Hi list,
>
>
> How does one run a test and gather the results programmatically?

No idea :) I'll have a look.

Nico

>
>
> I'm trying the ways I know from GST but it looks a bit different, for
> instance you can't do:
>
>
> (MyTestCase selector: #mySelectorForATest) run.
>
>
> Cheers!
>
>
> --
> Bernat Romagosa.
>

--
Bernat Romagosa.

Stefan Krecher

Re: Programmatically running tests

In reply to this post by Bernat Romagosa

Hi Bernat,

Am Montag, 10. Oktober 2011 16:49:00 UTC+2 schrieb Bernat Romagosa:

How does one run a test and gather the results programmatically?

an interesting question in this context would be - how would an test-suite be executed from a CI Tool like Jenkins or Cruise Control?

There are so much interesting new features in html5, so it could be a new architectural approach to implement the business-logic in the browser.

Besides other points like Transactions, Exception-Handling, Session-Handling, architectural approach (MVC vs. MVP vs. PAC vs ...) etc. CI/ Testing is an important thing for enterprise applications.

regards,

Stefan

Bernat Romagosa

Re: Programmatically running tests

So, I've been taking a better look at it, and it seems this does the trick:

result := TestResult new.

StringTest buildSuite do: [:each | each runCaseFor: result].
result inspect

You can ask result for #runs, #failures and #errors, which is exactly what I needed :)

At first I was looking at the SUnit package and I couldn't figure out how it worked, but then I checked the TestRunner widget as gokr suggested, and it all became clear.

Cheers,

2011/10/11 Stefan Krecher <[hidden email]>

Hi Bernat,

Am Montag, 10. Oktober 2011 16:49:00 UTC+2 schrieb Bernat Romagosa:
How does one run a test and gather the results programmatically?

an interesting question in this context would be - how would an test-suite be executed from a CI Tool like Jenkins or Cruise Control?
There are so much interesting new features in html5, so it could be a new architectural approach to implement the business-logic in the browser.

Besides other points like Transactions, Exception-Handling, Session-Handling, architectural approach (MVC vs. MVP vs. PAC vs ...) etc. CI/ Testing is an important thing for enterprise applications.

regards,
Stefan

--
Bernat Romagosa.