You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@tamaya.apache.org by "Oliver B. Fischer (JIRA)" <ji...@apache.org> on 2017/02/04 10:48:51 UTC

[jira] [Assigned] (TAMAYA-236) Reconsider PropertySource ordering and getOrdinal()

     [ https://issues.apache.org/jira/browse/TAMAYA-236?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Oliver B. Fischer reassigned TAMAYA-236:
----------------------------------------

    Assignee: Oliver B. Fischer

> Reconsider PropertySource ordering and getOrdinal()
> ---------------------------------------------------
>
>                 Key: TAMAYA-236
>                 URL: https://issues.apache.org/jira/browse/TAMAYA-236
>             Project: Tamaya
>          Issue Type: Improvement
>          Components: API
>    Affects Versions: 0.2-incubating
>            Reporter: Werner Keil
>            Assignee: Oliver B. Fischer
>             Fix For: 0.3-incubating
>
>
> As marked {{TODO rethink whole default PropertySources and ordering}} in {{PropertySource}} we should reconsider the whole {{getOrdinal()}} method.
> The Chief Architect of Lightbend, key contributor to Typesafe Config mentioned in a Microprofile discussion about ordinals, that he's 
> {quote}
> unconvinced that the use of ordinals is a preferred solution for the problem at hand—it just ends up becoming magic numbers which sadly are not absolutely ordered. It is also extremely finicky since a change in an ordinal by a third party will lead to surprising effects downstream, most specifically for config value overloads.
> A deterministic, absolute, order at declaration site means that the >consumer is always in control and order cannot change due to the whims of third-party libraries.
> {quote}
> We may take a look at how e.g. Typesafe Config (and others like Apache Commons Config) handle this, but I am not aware either of them use the concept of an "ordinal" or "order" to solve this problem.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)