Hacker News Reader

Comment by suyash | Hacker News Reader

Yup all the free stuff is ultimately to lead us into using and paying for Azure. Need an open source alternative to GitHub.

View on HN

Sourcehut would like to have a word...

https://sr.ht/

I like the concept but the UI is so ugly

I find the UI much more beautiful and clean than github's, but this is mostly a personal preference. Just hope it will be feature complete really soon!

I had not seen that...interesting.

Source hut is cool. Among things, it prioritizes exposing and leveraging Git's existing functionality wherever possible, like email patches, rather than building new proprietary functionality on top of Git like Github and others do.

Everything about Github's PR workflow is killer to me, so to me personally the message I take away from this is that Source Hut isn't even competing with Github in the parts I care about.

If the one particular aspect I mentioned isn't for you then why even bother wasting your valuable time responding? Just go read their website instead and see if there is something for you. Maybe there is, maybe there isn't, but at least then you'd know definitively. Much better use of your time and effort. I'll even save you a few clicks:

https://drewdevault.com/2018/11/15/sr.ht-general-availabilit...

https://sourcehut.org/

I posted because I think my post would be useful to other people who are considering alternatives to GitHub but also like rich PR workflows like GitHub has, and because I think it's good to encourage would-be GitHub alternatives to invest in that area.

I've actually been following Sourcehut's mailing list and updates since forever because I find their philosophies interesting, but not right or practical for my own workflow.

Github flow is garbage.

Its broken by design and forces a subpar experience on top of something that actually fucking works.

Maybe you find it nice and dumbed down, good for you. The rest of us want git to actually be git.

Sharing dissenting opinions is useful and by no stretch a waste of time.

It was just on the front page last week for a new feature launch (project hubs). The design is very minimal, but it’s organized into several different composable features. The author is on here often.

https://news.ycombinator.com/item?id=23030489

I haven’t migrated over completely from GH yet, but it’s a compelling option.

Codeberg[1] also recently had some conversations here.[2]

[1] https://codeberg.org/

[2] https://news.ycombinator.com/item?id=22795930

Why is that a bad thing? If you found value in Azure and it solved your problem, why is that a negative?

Switching costs.

If the things are offered free, and you build a lot of it into your system, and now it costs money, you have to choose between paying that money, or paying a potentially high switching cost to move to a different system.

If you had known you were going to be faced with those costs, you may never have bought into the system in the first place.

The trap is comfy and warm. Why ask what the springs are connected to?

GitLab has been great for a while and actually had some features before GH.

Thanks for the mention. As for how the features today map:

1. GitHub Insights => GitLab Insights https://docs.gitlab.com/ee/user/project/insights/ Code Review Analytics https://docs.gitlab.com/ee/user/analytics/code_review_analyt... DevOps Score https://docs.gitlab.com/ee/user/instance_statistics/dev_ops_... and Value Stream analytics https://docs.gitlab.com/ee/user/analytics/value_stream_analy...

2. GitHub CodeScanning => GitLab SAST https://docs.gitlab.com/ee/user/application_security/sast/ and Secret Detection https://docs.gitlab.com/ee/user/application_security/sast/#s...

3. GitHub Codespaces => GitLab Live Preview with client side evaluation https://docs.gitlab.com/ee/user/project/web_ide/index.html#c... and upcoming server side evaluation https://gitlab.com/groups/gitlab-org/-/epics/167 which would work both with on-premises Kubernetes clusters and in any cloud.

Simple but big issue I have is getting solutions that suit the platform. Often when I'm looking for help on something it gives me the on-prem solution. Docs on your website are confusing for this reason and that's why I stopped using. Takes me 30 seconds to solve a problem on GH and 5min to solve the same problem on GL. Clean that up so when devs try out the platform it's slick and easy to buy into not annoying and difficult.

Why? It's not like you MUST deploy on Azure

I think it's more about all of the integration leading to efficiencies that are hard to ignore while they're being subsidized / promoted. It'll start out as something where the value proposition is too good to ignore, but, as time goes on, they'll tier off features and charge a premium as you scale up.

There was a time when I was hopeful and though Microsoft would take a developer first approach with the end goal being making it easy to deploy to Azure. After seeing the VSO (the old name for Codespaces) pricing I'm a lot more cynical. Microsoft's going to be altruistic until they're not.

Go look at the pricing for VSO. They have a full time dev (100 hours a month - lol) estimated at a cost of $23 USD per month. Just wait until there's some type of marketplace for extension developers. The local version of VSCode will disappear overnight and (IMO) we'll be paying $50+ / month for the same thing that's free right now.

Get ready to pay for everything per user per month for the rest of your life.

> The local version of VSCode will disappear overnight

That's the beauty of open source though. VSCode is prominent enough that someone will fork it and maintain it if Microsoft pulls the plug.

My fear is that an officially supported marketplace for selling extensions would be appealing enough that important extensions would put all future development effort into their marketplace versions and cease development of local versions.

Hi! PM on the Codespaces team here.

I wanted to quickly help calm any potential concerns about changes to VS Code or its model.

Codespaces is a separate, additional service from VS Code, very much focused on providing compute catered to developers and their workflows. We hope you'll use and love it, but we're also very happy if you'd prefer to continue to use VS Code in the ways that you may have previously done.

Either way, we will continue to invest heavily in VS Code as you know it. Codespaces will not change that.

Hopefully that helps clarify any concerns!

I believe you when you say your intentions are good, but I'm skeptical. Even if it's not the official stance, getting as many people as possible onto the SaaS offering has to be the goal and, IMO, there could be a point where maintaining the standalone product isn't worth it.

I'm worried that Microsoft's going to be successful in shifting a bunch of developer workload onto Azure and that'll create a power imbalance where developers will see increasing prices while they lose control their own workflows.

Is it true that Nadella frequents the gym and chastises people for running on treadmills instead of Azure? Lol.

> we will continue to invest heavily in VS Code

Just like MS heavily invest in Atom now? :)

"nobody gets fired for buying IBM"

Gitlab CE, Gitea, Gogs

Some background on these:

Gitea is a fork of Gogs. Both are written in Go and are pretty light on resources (they support SQLite as database backend).

Gitlab is heavier on resources and requires at least Postgresql and Redis.

Gitlab looks nicer and has a lot more features even in the open source edition. (Though some paid features seem to be arbitrary paid vs free, like scoped tags, a feature everyone could use is paid, while a really Enterprise feature like kubernetes support is free)

Doesn't gitlab fit the bill here?

Reply on news.ycombinator.com