Really? That is a relief. Normally when people say they want to use Git, they really just mean that they want the trendy fun experience of Microsoft GitHub. XD
Especially since hosting your own Git server isn't exactly elegant these days (either massive VMs, shedload of python PIP deps, just bare minimum with ssh or of course the inbuilt git server with no authentication).
Looking at the URL's it suggests the primary git repositories will be on own servers, unless the plan is to move the repositories after the ending of the beta phase somewhere else.
Once upon a time, FreeBSD was a constitutional republic. Now it's a union, running by unelected bureaucrats!
Anyway, there's nothing we can do. Read a few emails from June 2014 to October 2014 (you can go further!) in the https://lists.freebsd.org/pipermail/freebsd-git/. Choice of words are interesting, e.g. "Totally [...]" "FreeBSD is getting with the program" and "just a few die hard people [...]". --- You've just need one of these mindsets in the upper echelons of any project.
By the way, It is hard to find any comprehensive lists of some discussions on the whole git/github thing, from genesis to revelation ... . To be specific, conversation between Core Team and/or FreeBSD foundation, who started it first, yays and nays, so forth and so on. If there's any, please share those links and possible blogs or archives. It's easier to find this kind of materials on Linux topics. There's more news and more eyes on that project. Because Linux has more user-base, and it leads me to my next point:
I completely disagree with mjollnir on [*]
But, there's a good point: widening the user base has a gigantic benefit. We're going to have more internet stalker (blogers, social neters!, youtuber, ...) and crowd will look more closely to FreeBSD to find topics and publishing more hit-pieces ... as a result gaining more view/attention and clicks/ad. It has a nice consequences for FreeBSD: more blogs and links on different topics, such as this one [**]
[*] I won't argue on it, it's just my personal opinion, feel free to criticise me for this one, on the other thread, over there, not here!
[**] Over there, Linus Torvalds picking on some dummy on mailing list, leads to world wide news headlines and detailed articles. Over here, git happens with feel of mystery and we have to read between the lines, speculate, guess ,...
Yes and that is why I wouldn't be completely up in arms even if they did move to MS GitHub. It does make it more accessible.
However, I think Theo hit the nail on the head when he stated in an interview (I think it was this one: https://undeadly.org/cgi?action=article;sid=20191231214356) that he didn't like "users". He prefers developers because they have more initiative to fix things if anything goes wrong. They are also more likely to know what is a good direction for a project (they understand limitations). They don't suck and drain a project.
So, does FreeBSD really want more users? I personally think the answer to that is a resounding yes! but only "good" ones XD
Trying to attract a load of Android/iOS tablet users for example will do more harm than good because they will never be happy. Whereas trying to attract a load of Sysadmin users because they happen to be more up-to-scratch with the Git workflow is ideal.
I can see why they want to get rid of portsnap itself. But SVN as a way to update source code needs to stay. The functionality of svnup for this is great. I wish a configuration file like this was in the base rather than local.
They would have to have a GIT software BSD alternative before attempting to use a GIT repository method for this. GIT due to its license can't go into the base without changing what FreeBSD is intended to be.
I'm pretty sure that is not a top of mind point when a developer wants to work on any software. It may be a positive for some but not a reason people would come here.
I use portsnap to update /usr/ports nightly with a cron job. Using poudriere for that would be overkill. I mean, poudriere just uses svn internally, so I can just use svn myself (or better yet, svnup) without needing poudriere.
Anyway, portsnap will continue to work for quite some time, so there is no reason to rush now. I would at least wait until the project has changed from svn to git. And then you can decide what’s the best way to update your ports collection in the future. Maybe at that time a new tool like “gitup” exists, or the “got” thingy of the OpenBSD folks is up to the task, or whatever. Or use svnup against the SVN-API of GitHub. There may be quite many possibilities.
Bottom line: Don’t worry. Everything will be fine.
I can see why they want to get rid of portsnap itself. But SVN as a way to update source code needs to stay. The functionality of svnup for this is great. I wish a configuration file like this was in the base rather than local.
They would have to have a GIT software BSD alternative before attempting to use a GIT repository method for this. GIT due to its license can't go into the base without changing what FreeBSD is intended to be.
I wonder if both git AND svn can be used for source and ports? That would make new installs easier for those used to svn and not git, for essential ports [ video, networking ] that have to be built at first, or even as part of the install script.
I wonder if both git AND svn can be used for source and ports? That would make new installs easier for those used to svn and not git, for essential ports [ video, networking ] that have to be built at first, or even as part of the install script.
I was wondering if SVN would be kept only for FreeBSD updates, and GIT for ports and documentation. I was looking into comparisons of the two. SVN was described as monolithic, and GIT as useful for smaller projects that didn't need to update the whole source tree.
For downloading FreeBSD documentation, SVN required downloading the whole tree, when only a part, or a subset was needed. It makes sense to keep SVN for FreeBSD base sourcecode. In comparison, SVN is way better than CVS or CTM. For many, at least for now, SVN would remain useful for sources needed to compile the kernel. I stopped compiling my base, as it often breaks things, that prevent consecutive build/install worlds. Recompiling World is too much work for that, when unwanted services can be blocked with PF.
I have doubts about GIT, and if OS's move onto it, they'll be reluctant to leave it for something better. The current move isn't for something better, it's a bandwagon movement.
Maybe they DON'T wan't to "work" on the software, don't know how FreeBSD's svn works, but they DO know how to make a fix and a PR. Enough of those and then they might want to "work" on it.
Doesn't sound like they are of any value then. How can you fix things on FreeBSD when you don't want to and don't know how it works? And being able to use git will change all that? I am sure that's not true.
If someone wants to work on an operating system or some software, git is not the thing that will make them do so and lack of git won't hold them back.
If the plan is to appeal to more programmers, and it's reasonable to assume they are targeting Linux and Windows developers, especially in latter case i.e. Windows developers, It's better to have a plan to port a full-feature GIT GUI front-end to FreeBSD beforehand. I'm not aware how things work on Linux teams, but in Microsoft/.NET world, most of the windows developers almost exclusively rely on Visual Studio to work with GIT and everything else. From integrating WSL with Visual Studio, to checking GIT commits. Don't be upset if you're using Visual Studio Code, but vscode is a joke on Windows developer environments and it's mainly used by web developers. Of course some install MSYS, others using WSL or GIT installer on windows. But It's a naive assertion that presume Microsoft developers who are using Visual Studio for many years, all of sudden change the lane to FreeBSD, only because it's source is available on GitHub/GIT.Also they have to learn to work with #!/bin/sh too!
I'm sorry to say, I don't buy it.
What I read in this report is consistent with what I can read in many posts on this forum and on the mailing lists.
It looks like FreeBSD is heading for the same destiny as Illumos, as described in this post (the 3rd one in the thread), for more or less the same reasons.
How sad!
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.
github.com
