[kde-community] Our new project metadata system

[Boudhayan Gupta]

On 1 April 2016 at 17:58, Thomas Pfeiffer <thomas.pfeiffer at kde.org> wrote:
> On Freitag, 1. April 2016 21:54:10 CEST Ben Cooksley wrote:
>
>> Out of curiosity, how are the Appstream files accessed by tools such
>> as the various app centers?
>> I presume some kind of repository exists?
>
> Nope, actually distributions extract the appstream data from the source code
> of each application and generate their own AppStream database.
>
>> If so, it may be worth pulling the information you need Boudhayan from
>> such a repository ( although the indexing process will undoubtedly
>> require either source code checkouts or compiled code, in which case
>> we'll probably have to use what the CI system has and pick a branch
>> group to represent - probably stable-kf5-qt5. More details on this
>> would be nice )

I was looking at the AppStream XML files and it has all the data I
need, even the translations. I can use that.

We could run a nightly job to pull in AppStream data from the repos
and keep it in a place the apps website code can access. Keeping it in
a standard place in the repo (say at the root) is probably a very good
idea.

> We could theoretically just tap into a distribution's AppStream database if
> that makes our lives easier, or just reuse their code to extract the data and
> create the database.

Using distributions' databases mean we don't have control over the
infra for our website. I don't think we can have that.

-- Boudhayan