Are you guys fine with these new shenanigans from Github. I found a bug and wanted to check what has been the development on that, only to find out most of the discussion was hidden by github and requesting me to sign-in to view it.
It threw me straight back to when Microsoft acquired Github and the discussions around the future of opensource on a microsoft owned infrastructure, now microsoft is exploiting free work from the community to train its AI, and building walls around its product, are open source contributors fine with that ?
My hot take is the quickest way out of this quagmire is to abandon Git. With the education system & bootcamps raising the next generation to think MS GitHub is Git, it would probably be less work to start rolling with another VCS as megacorp Microsoft won’t have the agility to pivot away from Git. Git isn’t even that great—arcane CLI, patches don’t commute, basically permanently locks in your name & email, large files require a separate tool, etc. And most of the popular alternative forges are literally just trying to clone MS GitHub rather than invent something new or solve the shit problems it doesn’t like threading, pull request model sucks, source code doesn’t need to be a social media platform with gamified stars & anxiety-inducing activity charts to encourage that MIT code in your free time the corpos will use & never contribute back while demanding you use it to build your résumé… or it’s built on email as the common denominator with fingers in ears as if mailing lists are the optimal workflow for all projects when a majority of folks don’t even know how to bottom post & keep their mail with the same evil Microsoft or the other evil Google. Nothing is being bold enough to actually have a better user experience—currently the best lure is… free software, but worse UX? Being a better UX the Microsoft GitHub is not a even that high of a bar. Some folks claim “network effect” but it seems clear that a lot of folks already want out.
Nah. Start anew. Check out Darcs, Pijul, Fossil, Mercurial, Bazaar, or whatever else is out there. Build on the ideas that improve version control.
Wtf just use gitlab/gitea/whateverthefuck instead of Microsoft garbage
Your project will inevitably get forked onto MS GitHub & the SEO will rank that fork above you—changing VCS adds a layer of friction that discourages forking back onto MS GitHub. Best you see is these pretty please attempts to encourage not forking to Microsoft’s platform. Microsoft has a massive pull in the direction Git goes, & whatever MS GitHub does, the alternative forges seems to want to clone even if it’s bad (look at Forgejo diverging from Gitea to copy Actions verbatim even tho we all know working with YAML for CI is a bad idea that scales poorly). When you look at the latest release notes for Git, often the first publisher is GitHub’s blog—this is what gets shared around the link aggregators. Part of the strategic purchase of GitHub on Microsoft’s part was getting access to that project sway (& upselling services—it’s not pure conspiracy—with some of the changes definitely being for the better).
And again, Git is not the best DVCS—but folks are hesitant to try other platforms since there is less forges & tooling. If Microsoft is controlling the Git ecosystem like it is, that effort, in my opinion, would be better spent choosing a better DVCS system that isn’t already infected by Microsoft or Google or Apple or similar.
I self-host GitBucket, and honestly your reasoning behind giving up arguably the best version control application, just because of one hosting site, is downright ludicrous.
Yeah, I stand by my stance that Microsoft has poisoned the whole Git experience, where everyone will be comparing all forges to MS GitHub & the direction of the Git project ship is being steered heavily by Microfsoft. I also disagree with “best” VCS—I will agree with Git having currently the most/best tooling around it which can lead to a better overall experience, but Git’s fundamentals are not without some obvious flaws.
It hasn’t. There are literally thousands and thousands of developers using Git daily without having nothing to do with GitHub.
You are entitled to your opinion, but that’s a fact. What MS does or doesn’t, with GitHub, has no effect on these devs. You can see how egregious it is to read a random person sayint we should stop using a certain tool, because Walmart also uses it? Jeesh.
I would venture while thousands use Git without MS GitHub, 98% have an account since you largely can’t contribute to many projects without due to lock-in.
If Walmart was the biggest funder, making the most calls to the project, & optimizing it to be sold in their stores, I would 100% have hesitance against something if I could find an alternative (physical vs. digital goods working a bit different).
It’s not that I don’t see your point, it’s just that I’m pessimistic that the open source community at large would in practice move off of MS GitHub or otherwise offer alternative contribution channels before we would see another tool + platform supplant Git as the status quo in the next generation of VCSs. I would rather accelerate that future—unless like Google or Facebook is the clear leader of that new tool, but many projects right now not named Jujutsu seem to be independent.
The great majority of developers never contribute, that’s a false expectation. Majority of programmers work in the private sector and use local git hosts/solutions instead of GitHub.
Again, expecting those devs to not use git because of one hoster, is a ludicrous idea in itself.
Open source drives that ecosystem & the private sector follows suit as the developers try new tools & suggest platforms from their open source experience. There are a lot of companies on MS GitHub Enterprise as well as opening some of there source in the form of libraries & they all choose Microsoft for marketing reasons over technical ones (Git is just Git after all).
You are acting like I’m saying “leave Git now or else”. I suggest it could be an option, if IMO a faster option, to get out Microsoft’s of monopoly. It’s easier to pitch as better tool with better features that solve problems than argue “well, ethically, you just self-host or move to Codeberg since user freedom matters”. That philosophical argument is harder to preach than Fossil ports the whole forge + tooling ecosystem with your project or Darcs/Pijul solve the merge conflict order of patches not commuting or Mercurial offers a more user-friend CLI.
What? Th E fuckv
Have you tried anything contemporary to or newer than Git (i.e. not CVS or Subversion)? It’s quite an anomaly that Git has held it’s reign as long as it has compared to other software & to assume it’s not worthy of criticism or isn’t regularly criticized online is delusional. There has been a lot of interesting work in VCS space that many have ignored since their heads are just stuck in the Git bubble. Was Git better than things before it? Mostly yes, but there are options now (& around the same time frame) with more ergonomic CLI, better conflict resolution, handling of large file blobs, better project management, & so on.
You seem to know a lot about VCS. Recently I’ve been looking for a VCS that handles binary files well, has deduplication, allows for deletion/forgetting of older versions without too much hassle, deals well with binary file conflicts, and allows for storing the old binaries on another server (like git LFS). Do you happen to know something that fits that description?
My memory could be failing me but I believe the gaming industry prefers Perforce for large files. Pijul’s FAQs seems to point in the direction that it could be good enough for some use cases. I haven’t put too much effort into researching solving that specific hard problem, but if I was to create a video game, you really need to look at how to best handle your assets.