Hi Aditya,

One way we could use and is already used for the Gradle and Birt Flexible documentation in wiki is to create README.md files, uses Pandoc to generate a HTML file from it in tools\wiki-files and import this file in wiki using the HTML import macro.

Doing  so we follow both way you suggested. So users have it in 2 places while it's only maintained in one place where it's versioned (though Confluence also versions pages)

Maybe an overkill in this case though. Since we have already https://cwiki.apache.org/confluence/display/OFBIZ/Revisions+Requiring+Data+Migration+-+upgrade+ofbiz to start from

My 2cts

Jacques


Le 23/09/2017 à 10:41, Aditya Sharma a écrit :
I see two ways to do that

1. A page maintained on Confluence where all the data model changes are
updated. Here, we can have a page maintained for the upcoming release when
the release is out we make it sub child titled with the release name.
2. A file is maintained in ofbiz-framework code base itself that goes with
the package with information about data model changes. Whenever someone
downloads the package this file will help track data model changes that
come with the package.

We can make it part of contributor's practice to update it whenever such
changes are done.

Which way should we go?

Thanks and Regards,

*Aditya Sharma* | Enterprise Software Engineer
HotWax Systems <http://www.hotwaxsystems.com/>
<https://www.linkedin.com/in/aditya-sharma-78291810a/>
On Sun, Sep 10, 2017 at 4:46 PM, Michael Brohl <michael.br...@ecomify.de>
wrote:

+1 for the documentation of database changes.

We alreadyy do this for customer projects, along with (database specific)
data migration scripts.

I'm not sure if we can afford to provide sophisticated additional tool
support which is maintained continiously?


As a conclusion, I think we should setup a convention that any database
model change has to provide a proper change log and migration script for
the embedded database (if applicable).

Thanks,

Michael


Am 01.09.17 um 15:24 schrieb Taher Alkhateeb:

Groovy scripts are also great and can do the job. To comment on the
"advantage" of something like flyway or liquibase I can try to list
some:
- The scripts might get too big or complex to accommodate different
databases and platforms.
- Out of the box, these solutions are database independent
- Ability to redo / undo on schema changes
- Supporting declarative style for schema definitions based on
multiple formats (YAML, XML, JSON, etc ..)
- The DSL is easier to use (declarative and short)

So in short, both solutions are viable, and existing solutions might
be a bit easier to implement especially if you consider additional
features in those solutions.

On Fri, Sep 1, 2017 at 2:23 PM, Nicolas Malin <nicolas.ma...@nereide.fr>
wrote:

I'm in favor to tracking the different migration but at this time I
didn't
see the advantage to use flyway or other instead of manage easily by
ofbiz
throw groovy script.

I'm available to help for design or create a POC do realize it because
many
time in the past (and currently ow) I want to refactoring some code/db
schema with data migration but we haven't clean process to do that.

Nicolas


Le 31/08/2017 à 12:37, Jacques Le Roux a écrit :

Le 31/08/2017 à 12:32, Taher Alkhateeb a écrit :

I would personally prefer to not go any
earlier than 13, or preferably just 16 to trunk, which means we design
this solution for the future, not necessarily the past.

+1

Jacques





Reply via email to