[buckminster-dev] Re: Version override with type Triplet vs. Omni Versio

[Henrik Lindberg <henrik.lindberg@xxxxxxxxxxxxxx>]

Can you file a bugzilla issue for this?

Sounds like it translates the input to OmniVersion, and then the 
translated version once again :)

- henrik

Ken Wenzel wrote:
> Hallo,
>
> I suppose that I encountered a bug regarding Buckminster's capabilities for
> version overrides using advisor nodes. I'm using the latest version of 
> Buckminster
> for Eclipse 3.5.0.
>
> If type "Triplet" is selected then Buckminster automatically converts a 
> given version number
> to an Omni Version expression. For example the version number 2.3.0 gets 
> converted to "raw:2.3.0.M/format(n[.n=0;[.n=0;]][d?S=M;]):2.3.0" - this 
> is controversial to the selected version type "Triplet" (by the way the 
> type Omni Version is missing and therefore cannot be selected).
>
> If I try to save the CQuery then the following exception is raised:
> Save Failed
> Format "format(n[.n=0;[.n=0;]][d?S=M;])" was unable to parse 
> raw:2.3.0.M/format(n[.n=0;[.n=0;]][d?S=M;]):2.3.0
>
> It seems that Buckminster tries to parse the version number as a triplet 
> and not as an Omni Version expression.
>
> Either Buckminster should not automatically transform a triplet version 
> number to an Omni Version expression or the attribute type should be 
> omitted for version overrides and all versions should be handled as Omni 
> Version expressions.
>
> The current implementation of version overrides with type "Triplet" is 
> unfortunately unusable.
>
> Best regards,
> Ken Wenzel