[jira] [Commented] (DIRSTUDIO-958) Exporting schemas as OpenLDAP files outputs entries in random order

["Stefan Seelmann (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/DIRSTUDIO-958?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13835104#comment-13835104 ] 

Stefan Seelmann commented on DIRSTUDIO-958:
-------------------------------------------

Actually the order is not random, but the same order the schema objects were added to the project. For online project the schema objects are exported in the same order as received from the server. And attribute types are always exported first, then object classes.

That said I agree that it makes sense to export them using their hierarchical order.

I wonder if we can reuse existing code (e.g. Registries) to build the hierarchy and allows to traverse them for export? I see some difficulties because a single schema (in the sense of what we call schema in the Studio schema editor, a set of object classes and attribute types) on its own may not be valid as AT and OC not included in the schema may be referenced.


> Exporting schemas as OpenLDAP files outputs entries in random order
> -------------------------------------------------------------------
>
>                 Key: DIRSTUDIO-958
>                 URL: https://issues.apache.org/jira/browse/DIRSTUDIO-958
>             Project: Directory Studio
>          Issue Type: Bug
>          Components: studio-schemaeditor
>    Affects Versions: 2.0.0-M8 (2.0.0.v20130628)
>         Environment: Ubuntu 12.04 x86_64
> Java 1.7.0_45-b18
>            Reporter: Piotr Kubowicz
>            Priority: Minor
>
> To reproduce:
> 1. Create two separate schema projects; for example one online and one offline from the same server
> 2. Open schema editor, choose the online project, select 'apache' schema, right click, choose Export -> Schemas as OpenLDAP files. See the generated file.
> 3. Open the offline project, export 'apache' schema as described above
> 4. Files differ although the schema is the same!
> It seems that the order of entries during schema export is random. It would be correct if entries are sorted by type first, then by OID.
> Note that the random order may cause problems when the exported schema is used in some directory servers - you may have a parent object class appear after its child object class. As result a directory server will report schema as invalid due to 'missing' parent object class.



--
This message was sent by Atlassian JIRA
(v6.1#6144)