toggle quoted messageShow quoted text
Well I use github cli for some stuff in elektron forge but I also
have github desktop which I use all the time.
Its easy to clone repos and stuff though I need to get their git
links off the site first.
As for a centralised place, I do wander why all the centralised
information couldn't be stored in ini files or text files.
Not sure how to do xml stuff but that is a commen way of storing
all sorts of data.
On 4/02/2021 5:30 pm, Reef Turner
The proposal for the NVDA addon store submissions is
intentionally aiming to keep all metadata in a centralised
location. This allows us and the add-ons community to moderate
it. The plan involves making the list of add-ons available
directly from NVDA, so NV Access must be cautious about what
will be available via that list.
For add-on authors who either dislike using or are
unfamiliar with GitHub I can recommend the Github CLI tool,
which allows creating Pull requests from the command line. I
believe this will streamline the process for many users.
I think we are basically at a stage where we can start
accepting the json metadata for add-ons to the store. And
start offering it from NVDA. One thing I'd like first is a
published checklist for reviewing updates to the json data.
To start with we can perform the checks manually, then I
hope that some of the people in our community who are
experienced with Github Actions can help to automate the
checks. If all the checks are automated we will be able to
have a very rapid update process.
I should also point out, I expect that the add-ons website
is updated to use the same data source. However I believe that
this will remain blocked until we have a system in place for
translating the metadata.
I hope this helps to explain the approach we are taking.
Software Developer - NV Access
I think we should have
asked the community at the beginning to come up with a unique
JSON for the update. Do if the "update_url" is not specified
in the manifest, use some default "https://addons.nvda-project.org/files/update.php?file=mySuperScript",
did you ever think about this when you parsed the HTML code of
the site and the version numbers in the file name?
And if "update_url" is specified in the manifest, then it is
already checked, and since the JSON structure would be the
same for everyone, the next steps would be identical,
regardless of whether JSON is downloaded from the community
site or from any other third-party site.
I imagined it in my head and I think it's the perfect option!