Re: [layerindex-web] PR / PE in layer index

[Paul Eggleton <paul.eggleton@intel.com>]

Hi Pranay

On Tuesday, 24 September 2019 10:00:46 AM NZST Pranay Mankad wrote:
> Thank you for supporting the recipesExtended view over the last couple of
> days. Here’s a small patch that includes values from two more variables from
> the Bitbake cooker – PR and PE.
> 
> I read through Paul’s explanation of why PE made sense but not PR, and
> thinking about it, the PR in layerindexs’ case shows if the untouched recipe
> had a revision associated with or not. Source metadata is important for
> information, and hence I’m submitting a change with the same.

I suppose so yes. Out of interest I did a search of OE-Core and there were 
more recipes than I would have guessed that set PR, so it could perhaps be 
useful. I'm not sure I'd bother adding it to the UI yet though (and you 
haven't, that's fine.)
 
> Here’s the patch (this is my first time submitting one, hopefully it’s the
> right way to do it):

So unfortunately no, you need to send it as an actual patch and not just 
pasted into an email - git-format-patch / git-send-email are the most 
effective ways to do that (though it assumes you are able to send email 
directly from your linux system). Here's some info:

http://www.openembedded.org/wiki/How_to_submit_a_patch_to_OpenEmbedded

(Of course you'd need to change mailing lists / prefixes in the examples, and 
some of the metadata requirements don't apply, but the basics are the same.)

One other thing - you're editing the initial migration (layerindex/migrations/
0001_initial.py) - that's not right. You need to undo the change to 
0001_initial.py and then generate a new migration using the following command:

./manage.py createmigrations layerindex

That will give you a new migration in layerindex/migrations/. You can then 
apply that migration to the database in the normal way:

./manage.py migrate layerindex

Of course if your database already has those new fields the latter will fail, 
in this instance you could simply add --fake since the result would have been 
the same.

Cheers,
Paul

-- 
Paul Eggleton
Intel Open Source Technology Centre