[ 
https://issues.apache.org/jira/browse/CB-7431?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14115479#comment-14115479
 ] 

Martin Gonzalez commented on CB-7431:
-------------------------------------

Hi Ian, well the changes that you've done, I'm not sure are going to work fine.
This failure is aimed for the  window.resolveLocalFileSystemURI process, that 
would only happen over the exec success callback,
{code}
failed.bind(null, done, 'window.resolveLocalFileSystemURI - Error resolving 
URI: file://' + encodeURI(localPath)));
{code}
However this statement on the error callback, as 'done':
{code}
}, done, 'File', '_getLocalFilesystemPath', [entry.toURL()]);
{code}
it establish that the test is completed, and no error were found, and if it 
reaches there without failed expectation, it would pass a failed test.

The right move it would be fail it, providing the proper information, besides 
of the errocallback.
{code}
failed.bind(null, done, 'cordova.exec -_getLocalFilesystemPath - Error getting 
local system path: '  +entry.toURL()));
{code}

That would fail the text providing the information, and failing the text if 
errocallback as well.
 

> File.spec.109 causes Jasmine2 to start failing
> ----------------------------------------------
>
>                 Key: CB-7431
>                 URL: https://issues.apache.org/jira/browse/CB-7431
>             Project: Apache Cordova
>          Issue Type: Bug
>          Components: iOS, Plugin File
>            Reporter: Ian Clelland
>            Assignee: Ian Clelland
>            Priority: Critical
>
> The translation of file.spec.109 from mobile-spec with Jasmine 1 to 
> in-plugin-tests with Jasmine 2 produced a situation in which the {{done}} 
> function could be called twice inside of a single test.
> This appears to put Jasmine in some kind of internally-inconsistent state, an 
> many subsequent tests fail (it appears to be all tests using custom 
> matchers). This behaviour was noticed on iOS, ant not on other platforms, but 
> may affect other platforms differently.
> Marked as critical because we can't continue testing iOS without this being 
> resolved.



--
This message was sent by Atlassian JIRA
(v6.2#6252)

Reply via email to