If this is dead, please tell us

[KlomDark]

Haven't seen any pull requests accepted in years. This is my favorite, but if it's dead I'd better start looking for something else.

Anyone know of active/supported forks of this?

[dauheeIRL]

looks dead. I can see a fork that upgraded to .NET 4.8 a couple months ago but thats it

[polymo1]

Now at 3 years no commit, I'd say this is abandoned based off of the number of issues and pull requests.

[latop2604]

We miss the transition to dotnet core.

I did a poc some years ago but the razor part was too dependent on the external libs. And it was too much to recreate the ui part.

Also the AD feature make it hard to test.

[latop2604]

This fork seem to be the most recent one https://github.com/Pepperdine-University/Bonobo-Git-Server/tree/Code_Review_2023-09-26

[willdean]

This was not originally my project, but I think I'm one of the last people to commit anything or make a new release.

The story of why this is not maintained is not really anything exceptional - it's much the same for all OSS.
Here are some bullet points, in no particular order:

• The original maintainer is gone
• I'm very busy, and I long-ago fixed the problems which were bothering me and why I got involved in the first place - other contributors have drifted away too.
• The typical quality of unsolicited PRs is, to my mind, low - so if I was to merge one I would have two tasks: rewriting to a reasonable standard and/or negotiating with the original author (who feels they're doing you a favour). People who have not been maintainers on OSS projects often do not appreciate what a problem this is, particularly once the code has already become ill-disciplined: it becomes harder and harder to maintain any quality.
• My feeling is that there has been a huge reduction in interest in running your own on-prem Git server - most people are using Github or Bitbucket or whatever - this means there are far fewer potential contributors
• Past maintainers have not really been very focussed on either restraint when considering extra features or consistent coding approach - so there's a terrible mix of styles and standards.
• Prior to my involvement I think there was a really, really terrible mistake made where AD users where hived-off into their own homebrew database and stored in a completely different way to non-AD users - this has been an absolute pain, particularly given the next point...
• As @latop2604 says, AD is a nightmare to test- if you don't use AD you can't simply fire-up an AD server without screwing-up your network, and if you do use AD then it's probably running your business and you can't fiddle-about with it to support testing your hobby-project. I don't know how people (used to) test apps which integrated tightly with AD, and doubt anyone is writing any more of them nowadays anyway. I think this is why so much of the AD code in Bonobo has been drive-by: people did what they needed to make something work and then ran away, perhaps screaming as they went.
• There are a bunch of dependencies on Git and Git integration libraries which are now very out of date.
• We, totally unnecessarily, support multiple RDBMS for the (trivial) internal database, probably (it's before my time) for no good reason other than that someone who was enthusiastic said that would be a cool thing to do. Doubling the test matrix in the process.
• .NET Core happened - which first created some terrible years to build .NET applications where nothing worked properly - and now that the dust has settled and .NET is great again, there is a huge gulf of time separating us from our dependencies.

I am aware of all the reasons people like Joel Spolsky say you shouldn't re-write software from scratch, but if I needed something like Bonobo which wasn't the current version of Bonobo, then I would start again. My version might not have a web-UI at all. It would serve Git over HTTPS and use a JSON file to associate users and their repositories.

It's not my place to tell you that the project's dead, because it's not my project: you'll have to draw your own conclusions there. Just as you should for all other OSS projects.

[Abubakarstu]

@willdean @latop2604 @dauheeIRL @KlomDark @polymo1 it is not dead Now me from Pakistan Is trying to Update to latest Features and Improving Ui

[polymo1]

@willdean @latop2604 @dauheeIRL @KlomDark @polymo1 it is not dead Now me from Pakistan Is trying to Update to latest Features and Improving Ui

Good luck, but I doubt it's worth your time.

No commits for 3 years. It's very unlikely a pull request will get reviewed.

[Matt-17]

Right now I am upgrading Bonobo for learning purposes and for fun to Net8. For this I completely dumped AD and SQLServer support as of the reasons @willdean has written.

Right now it works but's still buggy in some places. But if anybody is interested I will try to polish. Other changes I made so far is switching to EF Core Migrations and IConfiguration appsettings.json so an upgrade will not be straight forward.

But I guess, I won't create a PR as it's unlikely it will be reviewd.

[latop2604]

We can also update the readme file with a disclaimer to point to your fork. At least I can do it.

Also I think not having AD support is not Bonobo-Git-Server anymore. But something else. In my previous attempt to migrate to dotnet core I also dropped it.

I will be happy to help you on the new version

[Matt-17]

Thanks. I know, AD support is a key-feature of Bonobo. But the reason is, I have no access to an AD server and personally I always used (and use) it with ef only. If I could figure out how to easily test this, I might reintegrate this.
If I publish it, would it be better to keep it as fork or create a complete new repository? The latter would decrease the git repo significant as I would not include the git binaries.

[willdean]

I think keeping it as a fork seems sensible, so that the history follows through, but we can still be clear that it's a break with compatibility, or at least with feature-parity. I wonder how many people actually need native AD nowadays? In my professional life I look after a SASS system where basically 100% of our corporate users use "Microsoft Login" which goes via their corporate O365 accounts. That's in the UK - I realise other countries may differ. As I've already said, I think the AD support in Bonobo is problematic.

…

________________________________ From: Matthias Voigt ***@***.***> Sent: Wednesday, October 2, 2024 6:57:11 PM To: jakubgarfield/Bonobo-Git-Server ***@***.***> Cc: Will Dean ***@***.***>; Mention ***@***.***> Subject: Re: [jakubgarfield/Bonobo-Git-Server] If this is dead, please tell us (Issue #905) Thanks. I know, AD support is a key-feature of Bonobo. But the reason is, I have no access to an AD server and personally I always used (and use) it with ef only. If I could figure out how to easily test this, I might reintegrate this. If I publish it, would it be better to keep it as fork or create a complete new repository? The latter would decrease the git repo significant as I would not include the git binaries. — Reply to this email directly, view it on GitHub<#905 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AAJ63ZOPAXTLFABYOVQFIB3ZZQXXPAVCNFSM6AAAAABJSOTMISVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGOBZGI4DCMBRGU>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[rothgecw]

For what it's worth, I do still use/need AD auth, but I can understand if it's a less used feature that is dropped. I just probably wouldn't go with that fork if that's the case. Still, I'm glad to see work still being done on Bonobo.

edit: I should note, if I can auth against AAD, I think that would be just fine. I don't recall that being an option, but I may have overlooked it when I first set up my instance of it. Going to have to review that.

[Matt-17]

Hey, I finally pushed an alpha version of the upgrade. As mentioned, AD support has been removed for now, but I plan to add it later, without making it mutually exclusive with EF Authentication. I've also experimented with Microsoft login, which seems promising.

I’ve managed to figure out how to support both Sqlite and SQLServer again, so hopefully, that feature will come back soon. The database structure has changed a bit though, and I’ve moved server configuration into the database, which makes upgrading more challenging. The old web.config has been replaced by appsettings.json, which is still a work in progress.

I’ve made significant architectural changes, so it won’t be easy for contributors, and it’s still not final. Additionally, I’ve tightly integrated a new GitHub Pages website within the repository.

Because of all these changes and the new ideas I want to implement in the future, I’ve decided to rebrand the project as Gibbon Git Server (I’m bad at names ;-). One of the upcoming changes includes moving from group-based repository management to project-based management. I’m trying to maintain a Lite version, but no promises on that yet.

Since this is still an alpha version, there may be bugs, and I’m actively working on improving things. I’d love feedback or suggestions, and any reviews or issues opened on GitHub are highly appreciated!

Here’s the new project: Gibbon Git Server

[rothgecw]

Awesome - I'm going to follow the new project. I did have one question related to Bonobo and possibly the new Gibbon project: can someone have read without write permissions without making a repo public? I've struggled with this quite a bit. I have several repos I need to make available to specific (AD) individuals or (AD) groups, but I've found no way to limit someone's access to read-only. Let me know if I'm looking at this wrong. Thank you!