[jira] Commented: (TRINIDAD-1547) @agent in skinning should allow for a range of agent versions

["Blake Sullivan (JIRA)" <de...@myfaces.apache.org>]

    [ https://issues.apache.org/jira/browse/TRINIDAD-1547?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12794969#action_12794969 ] 

Blake Sullivan commented on TRINIDAD-1547:
------------------------------------------

Oh yeah.  The patch also fixes the following bug in the old code:

@agent webkit and (version:1.8) and (version:1.9.0)

would treat the extra 'and' as an 'or'.  It was identical to writing:

@agent webkit and (version:1.8), webkit and (version:1.9.0)

With this patch, such a rule will be parsed correctly to match the media query specification and will always return false as a result (since the version can never match 1.8 and 1.9.0 at the same time).  The solution is to use the correct syntax or, in this case, the new max-version support (since that was the intent of the rule)

@agent webkit and (max-version:1.9.0)



> @agent in skinning should allow for a range of agent versions
> -------------------------------------------------------------
>
>                 Key: TRINIDAD-1547
>                 URL: https://issues.apache.org/jira/browse/TRINIDAD-1547
>             Project: MyFaces Trinidad
>          Issue Type: New Feature
>            Reporter: Jeanne Waldman
>         Attachments: JIRA_1547_1_2_12_2.patch
>
>
> related to https://issues.apache.org/jira/browse/TRINIDAD-799 Add agent version support in skinning
> We currently have @agent support, @agent ie and (version:5)
> We need to add version ranges - the min-version and max-version.
> See thread from April, 2008:
> http://www.mail-archive.com/dev@myfaces.apache.org/msg31841.html
> -- from thread --
> max- means less-than-or-equal-to:max-version:8 means
> version <= 8.0 == true
> --
> option 5)  the version feature is a String that matches the native "major.minor.whatever" format of the browser's engine.  If the browser's engine uses non "." for separating versions, "." is used instead.
> For matches, the "*" character is allowed in any version section.
> For comparisons, the "*"  is always a valid match regardless of <, >, or =  comparison
> For comparisons where the comparison side contains fewer version sections than the actual browser version, the comparison side is padded with * version sections and the comparison occurs as above.
> For comparisons where the comparison side contains more version sections than the actual browser version, the browser version is padded with 0 version sections and the comparison occurs as above.
> // user wants to match IE 5, actual browser version ie 5.5
> @agent ie and (version:5)
> matches because version:5 expands to version 5.* and 5.* matches 5.5
> @agent ie and (min-version:5)
> matches because version:5 expands to version 5.* and 5.*  < 5.5 = true
> @agent ie and (max-version:5)
> matches because version:5 expands to version 5.* and 5.* > 5.5 = true
> // actual browser version gecko 1.9
> @agent gecko and (min-version:1.9.2)
> does not match because the browser version 1.9 expands to 1.9.0 and 1.9.2 is > than 1.9.0
> // actual browser version gecko 1.9
> @agent gecko and (min-version:1.9.*)
> matches because the browser version 1.9 expands to 1.9.0 and 1.9.* == 1.9.0 
> --
> I have implemented and attached a Version class with the desired behavior and tested that the following works as expected
>    System.out.println("toString:" + new Version("5.0.3", "*"));
>    System.out.println("hashCode:" + new Version("5.0.3", "*").hashCode());
>    System.out.println("not equals:" + new Version("5.0.3", "*").equals(new Version("5", "*")));
>    System.out.println("compareTo == 0:" + new Version("5.0.3", "*").compareTo(new Version("5", "*")));
>    System.out.println("compareTo == +:" + new Version("5.0.3", "*").compareTo(new Version("5", "0")));
>    System.out.println("compareTo == -:" + new Version("5.0.3", "*").compareTo(new Version("5.0.4", "*")));
> Given the way that compareTo works, I changed the wildcard behavior from what is listed before to make the wildcard behavior only affect compareTo() results of 0.  However, this still has the desired effect because min- and max- comparisons are <= and >= respectively.  Anway, I believe looking at the code in compareTo() will give everyone a better idea of what I intend. 

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.