Buckets leave and reenter the bucket list.

[okibcn]

@rasa, I see that the directory is not really getting new buckets, but it has a list of about 15-20 buckets bouncing from in and out of consideration every one or two days. How I saw it? well it requires a little intro:

---

created a while ago a Scoop search framework for myself. I finally decided to share it with everyone a few days ago given the lack of sound (IMHO) CLI scoop search tool.

The framework has two parts:

1. A database generator that creates a compact database with all the required info for the search tool.
2. A search tool that gets data from that database.

The database generator is, at this time, unable to explore GitHub to find new buckets. So, it uses Scoop-Directory to retrieve a list of known buckets.

---

After that (long?) intro, you can see in that project the history of two files: ActiveBuckets and DeprecatedBuckets. I implemented this tracking a couple of days ago to avoid missing any past buckets ( I am planning on doing the same for apps in the known scoop repos, now tracked as artifacts.).

What is the logic behind the bucket discovery-removal-recover policy in Scoop-Directory?

Thanks

[rasa]

@okibcn There is no policy, per se, but we should not drop an included bucket just because GitHub didn't return it in the most recent search result.

I will consider how much effort it would be to work around this issue. Thanks for bringing it to our attention.

[okibcn]

It shouldn't be a big deal. In my case, It was just a line, but it all depends on the code. I haven't gone through your code. Is it the Crawl code in Python forked in other of your GH repos?