Do you think the most common use case here would be in the context of a repo?

That is i am looking at a specific project, and want to search the code and issues of that project?

Not sure how useful a code search of all repos and forks at once would be.

Do you think the most common use case here would be in the context of a repo?

I would say initially yes and then we can expand.

Not sure how useful a code search of all repos and forks at once would be.

I think a good example is that you are working on a function that you remember the name of but forgot the repos name and a quick search would find it.

I would also like to add an issue search so someone is able to search through it all before creating a new ticket.

We can add text in the search bar noting that the search is from this repo

I'll see if i can create a mockup of this

Don't know if this can be of help but: https://github.com/google/codesearch

Don't know if this can be of help but: https://github.com/google/codesearch

I am been through this and had a discussion with @pingou about it so the problem is we don't want a lot of mix match in the codebase so something pythonic has more value.

So what are the chances to get code search in Pagure?

I would personally like to see at least some basic code search in pagure. I have to admit that this is something I find it quite convenient in GitHub, to quickly search through code, but also issues and PRs, to see if a problem I'm facing comes up somehow. Let's put it on the 6.0 Milestone and discuss later if / how to proceed.

I would personally like to see at least some basic code search in pagure. I have to admit that this is something I find it quite convenient in GitHub, to quickly search through code, but also issues and PRs, to see if a problem I'm facing comes up somehow. Let's put it on the 6.0 Milestone and discuss later if / how to proceed.

The issue we have for this is simply that we do not store the code in plain text
on the server. We only have the bare repo (which we clone - unbare - punctually
for merge PRs). So we would need to find a way to index the bare repo, or we'll
end up having quite an impact on disk usage.

Good point, I would think of an optional component that someone can spin-up even on a separate system, which then would be used if the feature get enabled. So to don't blow up pagure and massively increase the requirements, it should probably be optional and not part of the core.