[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Savannah bug tracker disabled?

From: Jordan Schidlowsky
Subject: Re: Savannah bug tracker disabled?
Date: Mon, 4 Nov 2019 12:36:19 -0600

Is everyone here forgetting git is not svn?  git is a DISTRIBUTED versioning system.  As in, when you use checkout a repo you have a copy of the ENTIRE repository.  Github is really just another remote for the repo.  It's absolutely TRIVIAL to setup a mirror on whatever other service you like (including hosting your own solution).  

If people are really concerned about the bug/issue tracking being tied to github's platform, then just do in-repository tracking.  Pretty simple.  If you wanna take advantage of the GitHub platform for pull requests and issue tracking then do it.  If you are concerned about ANY part of the project being tied to a service/hosting provider then just add some process to track those things in-repository as well...    (include the wiki and the website as well?)

https://github.com/MichaelMure/git-bug  (GPL BTW)

TBH most of this thread is FUD.  git itself solves almost all these problems.  IMHO, use GitHub as a remote for the repo (the advantages are clear).  If some don't like GitHub or have issues with it, mirror.  If there are concerns about vendor lock-in, then track (or additionally track) in-repository.

On Nov 4, 2019, at 4:46 AM, Ivan Vučica <address@hidden> wrote:

I’ll answer as if you are asking the question and not asking rhetorically.

Whoever hosts the project takes over the responsibility for:

- storing user credentials
- abuse management
- security and service upgrades
- data migration when moving from service A to service B
- possibly even GDPR and other privacy requests (data portability, right to be forgotten)
- backups AND restores
- service uptime and monitoring (remember, no SLO means 100% uptime which is a terrible impossible-to-reach target)
- administrative overhead and user support
- complaints when things aren’t going right
- deciding who gets administrative credentials to execute all of the above
- deciding how to finance all this ($5 droplet on digital ocean or something more?€
- technical planning for all the above

They take this on even if they are not aware of it.

I’m hosting my own stuff. Aside from development work, for three weeks a quarter hold a pager and/or respond to tickets in my professional life and/or shepherd automated or semiautomated software deployment processes. Something like GNUstep hosting would be less labor than that — but it’s still quite some stuff to be responsible for.

This is even before we get into technical choices you should make and how long it takes to deploy them.

apt-get install $LISTOFPACKAGES won’t cut it.

On Mon 4 Nov 2019 at 10:26, Umberto Cerrato <address@hidden> wrote:
Why don’t you self host your project in your own website? Something like a self hosted Savannah or similar (there are few around). Then you could: modify the UI to make it more GitHub-like and user friendly and leave some landing projects on GitHub, GitLab etc. that redirect to your e.g. subdomain with hosted files and bug tracker etc.

Hello there anyway…
Sent from Gmail Mobile

reply via email to

[Prev in Thread] Current Thread [Next in Thread]