Loading ...
Sorry, an error occurred while loading the content.

Re: [XP] Book Review: Pragmatic Ajax

Expand Messages
  • Luiz Esmiralha
    ... Are you in a bad mood today, Phlip? ;) A project can get by with testing for one browser only. Intranet projects are usually like this. I have to admit
    Message 1 of 24 , May 1, 2006
    • 0 Attachment
      On 5/1/06, Phlip <phlip2005@...> wrote:
      >
      > Luiz Esmiralha wrote:
      >
      > > Selenium is compatible with these browsers:
      > >
      > > Windows Mac OS X Linux Internet Explorer 6.0 IE 5.2 will not be
      > supported not
      > > available Firefox <http://www.mozilla.com/firefox/> 0.8+, 0.9+, 1.0+
      > 0.8 to
      > > 1.5 0.8+, 0.9+, 1.0+ Mozilla
      > > Suite<http://www.mozilla.org/products/mozilla1.x/>
      > > 1.6+, 1.7+ 1.6+, 1.7+ 1.6+, 1.7+
      > > Safari<http://www.apple.com/macosx/features/safari/> not
      > > available 1.3+ not available Netscape
      > > Navigator<http://browser.netscape.com/ns8/>
      > > ? not available not available Opera <http://www.opera.com/> ?
      > Incompatible
      > > ? OmniWeb <http://www.omnigroup.com/applications/omniweb/> not
      > available 5.1
      > > partial <http://wiki.openqa.org/display/SEL/OmniWeb> not available
      > > Konqueror <http://www.konqueror.org/> not available not available ?
      > > Galeon<http://galeon.sourceforge.net/> not
      > > available not available ? Camino <http://www.caminobrowser.org/> not
      > > available 1.0a1 <http://wiki.openqa.org/display/SEL/Camino> not
      > available
      >
      > Beautiful! Now some evil sticky questions...
      >
      > - has anyone ever tested the _entire_ set,
      > simultaneously, after each code change?
      >
      > - has anyone ever figured out how One Test
      > Button can drive tests into all those browsers?
      >
      > - has anyone ever figured out how to navigate
      > to any fault after such a test fails in one
      > of those browsers?
      >
      > These sound like TDD zealotry questions, and of course a healthy
      > project can get by with only spot-checks at test-button-time. However,
      > it seems those questions emerge from the TDD principles...
      >
      >
      Are you in a bad mood today, Phlip? ;)

      A project can get by with testing for one browser only. Intranet projects
      are usually like this. I have to admit that 80% of the projects I was part
      of were intranet projects.

      You can use Selenium RC (http://www.openqa.org/selenium-rc/tutorial.html) to
      automate tests in multiple browsers, as it integrates with JUnit easily.

      Cheers,
      --
      Luiz Esmiralha


      [Non-text portions of this message have been removed]
    • Phlip
      ... Oops. Are you Selenium s maintainer?? I m actually just thinking of all the trouble _I_ d get into trying to invent Selenium. ;-) -- Phlip
      Message 2 of 24 , May 1, 2006
      • 0 Attachment
        Luiz Esmiralha wrote:

        > Are you in a bad mood today, Phlip? ;)

        Oops. Are you Selenium's maintainer??

        I'm actually just thinking of all the trouble _I_'d get into trying to
        invent Selenium. ;-)

        --
        Phlip
        http://www.greencheese.us/ZeekLand <-- NOT a blog!!
      • Forrest Chang
        If you use selenium with Ruby On Rails, you can do a rake test_acceptance and it will kick off the selenium tests with each browser you ve specified in the
        Message 3 of 24 , May 1, 2006
        • 0 Attachment
          If you use selenium with Ruby On Rails, you can do a "rake
          test_acceptance" and it will kick off the selenium tests with each
          browser you've specified in the config file. This probably comes
          close to meeting your 1st two questions. Not surprisingly since these
          drive browsers, they go by quite a bit slower than normal unit tests.

          For the last question, are you asking for a navigation into offending
          code? You can navigate to the part that didn't pass in the test the
          way it is, and I've thought about writing something that goes through
          the ROR stack traces to hit that.



          On 4/30/06, Phlip <phlip2005@...> wrote:
          > Luiz Esmiralha wrote:
          >
          > > Selenium is compatible with these browsers:
          > >
          > - has anyone ever tested the _entire_ set,
          > simultaneously, after each code change?
          >
          > - has anyone ever figured out how One Test
          > Button can drive tests into all those browsers?
          >
          > - has anyone ever figured out how to navigate
          > to any fault after such a test fails in one
          > of those browsers?
          >
          > These sound like TDD zealotry questions, and of course a healthy
          > project can get by with only spot-checks at test-button-time. However,
          > it seems those questions emerge from the TDD principles...
          >
          > --
          > Phlip
          > http://www.greencheese.us/ZeekLand <-- NOT a blog!!
          >
        • Phlip
          ... Kent Beck calls that topic test faults are syntax errors . That means _all_ errors, whatever their source, should appear in the same list in your editor,
          Message 4 of 24 , May 1, 2006
          • 0 Attachment
            Forrest Chang wrote:

            > For the last question, are you asking for a navigation into offending
            > code? You can navigate to the part that didn't pass in the test the
            > way it is, and I've thought about writing something that goes through
            > the ROR stack traces to hit that.

            Kent Beck calls that topic "test faults are syntax errors". That means
            _all_ errors, whatever their source, should appear in the same list in
            your editor, and you should hit <F4> or similar to optionally navigate
            to them.

            Once when I pushed for a system like that, my colleagues using Vim
            demonstrated that it was impossible, because Vim indeed had an option
            to jump to the first syntax error. But its problem was it _always_
            jumped; you couldn't just opt to jump. Under TDD, 5!% of the time when
            a test fails your cursor should already be near the cause.

            --
            Phlip
            http://www.greencheese.us/ZeekLand <-- NOT a blog!!
          • Ian Collins
            ... Quite hard with JavaScript because your tests run in a different environment (browsers) for your editor. The browser my even be on a different box when
            Message 5 of 24 , May 1, 2006
            • 0 Attachment
              Phlip wrote:

              >Forrest Chang wrote:
              >
              >
              >
              >>For the last question, are you asking for a navigation into offending
              >>code? You can navigate to the part that didn't pass in the test the
              >>way it is, and I've thought about writing something that goes through
              >>the ROR stack traces to hit that.
              >>
              >>
              >
              >Kent Beck calls that topic "test faults are syntax errors". That means
              >_all_ errors, whatever their source, should appear in the same list in
              >your editor, and you should hit <F4> or similar to optionally navigate
              >to them.
              >
              >
              >
              Quite hard with JavaScript because your tests run in a different
              environment (browsers) for your editor. The browser my even be on a
              different box when testing in multiple environments.

              >Once when I pushed for a system like that, my colleagues using Vim
              >demonstrated that it was impossible, because Vim indeed had an option
              >to jump to the first syntax error. But its problem was it _always_
              >jumped; you couldn't just opt to jump. Under TDD, 5!% of the time when
              >a test fails your cursor should already be near the cause.
              >
              >
              >
              Not being able to link form the editor is a good way to force small
              changes and frequent test runs.

              Ian.
            Your message has been successfully submitted and would be delivered to recipients shortly.