Minetest: Move repository

[numberZero]

GitHub is being acquired by Microsoft. That means, while it will probably remain usable for a year or so, it will be changing to discourage participating in free software projects like Minetest. EDIT: and to become more “user-friendly,” that is, featureless full of ads user-tracking, service. So we may have to move somewhere.

By the way, it would be nice if you tweet GH your disappointment on the acquisition (unless you like MS, of course—but then you play MC and not MT, don’t you?)

[shimmark]

It is unlikely that GitHub will do anything that would harm Minetest in the short term after the acquisition is completed and if they do in the long term Minetest could move then. So the acquisition is not a sufficient reason to move the repository to another platform right now.

It is possible though that there was already sufficient reason to move the repository somewhere else before the agreement was announced and that the announcement merely makes this a good time to move, as opposed to waiting until something bad actually happens and only moving then. And indeed I think this is the case! CI is already working on the GitLab mirror and moving the main repository there and updating links would not require so much effort as to be prohibitive.

GitLab happens to be superior to GitHub in features and openness, while GitHub is superior to GitLab in the size of its community. If the latter translates into more contributions for projects hosted on GitHub, that could be a reason to keep the status quo. But I have not seen evidence that projects hosted on GitHub get more contributions as a consequence of the GitHub community being larger and my impression is that if there is any effect it is so negligible (relatively to other factors that influence the number of contributions a project gets) that I have not been able to notice. Maybe others have a different impression, and maybe mine is wrong, but in any case it is not clear that this is a significant advantage of GitHub, while features and openness are certainly significant advantages of GitLab, not even only for an open source project.

So I do think the GitLab mirrors should become the main repositories and I don't think there will be any better moment to make this move than right now.

[paramat]

Some IRC discussion. Celeron55's opinion:
http://irc.minetest.net/minetest-dev/2018-06-04#i_5320728
http://irc.minetest.net/minetest-dev/2018-06-04#i_5320791
http://irc.minetest.net/minetest-dev/2018-06-04#i_5321804

Does seem to me now a case of 'when' instead of 'if', however there's no rush, we will likely have a few months to consider a move. The deal itself will not be completed for several months, 'by the end of the year' apparently. We also need to wait out the chaos caused by all the migration, and not add to that chaos as it is affecting GH and other Git sites.

[paramat]

I do feel we should move though, and not leave it too long, as we will certainly lose contributors as they move away from Github and it will cause problems with attracting new contributors and future core developers. Contributors could still contribute from a different site but it will make contributing more difficult.

So i have a suggestion: after 0.5.0 is released, which is only a few months away. In order to not disrupt that release and leave moving to a less busy time for core devs.
Of course we should start thinking about this now to be prepared.

Related issue #7413

[nerzhul]

The move is quite easy on gitlab, it just take time to run the importer again.
Currently we have no plans for this, our contributors are here, we can move, yes, and keep the github clone refering to gitlab, but it's too early for this. Just wait for GH to do some microsoft things, and we can talk about it in some months.
We don't need to rush on it

[SmallJoker]

I've already said it on IRC and will say it here again - wait and see what happens. A panic move at this point is nonsense since we don't know yet how much M$ is going to change.

[Wuzzy2]

I am very much in favor of this step and applaud you for this decision, as GitHub is full of non-free JavaScript and it rates an “F” grade on the GNU Ethical Repository Evaluation https://www.gnu.org/software/repo-criteria-evaluation.html. So the move to a more ethical repository is definitely the right direction.
So it makes only sense to move on to a server which is more respecting to freedom.
Also, Minetest should leave GitHub for strategic reasons as it's in direct competition to Microsoft-owned Minecraft.

The first question to answer is, where to migrate to. There aren't many viable options in my opinion. (Possibly incomplete) List of acceptable alternatives:

• GNU Savannah (savannah.nongnu.org) (driven by free software zealots, very unlikely to screw you over)
• GitLab (owned by a company, but the site itself is “acceptable for a GNU package”. Pray they won't get bought by $COMPANY in future)

Sourceforge MIGHT be an option, too, they are now owned by a very different company since about 2 years, and the horrible adware program has been terminated and the website no longer looks like a total mess.

GNU Savannah does not offer automated exports, however. GitLab seems to offer this.

Thankfully, GitHub gives access to pretty much all raw data. Here I have written a quick and dirty guide on issue extraction to export them to JSON:
https://forum.minetest.net/viewtopic.php?f=56&t=20225#p320655

[nerzhul]

yeah let's move to shitgnunah without no features, or move to SVN shit on sourceforge yeah :yellow_heart:

No we will not move. And gitlab can be bought by apple or google at a point if it's strategic for them.

[sfan5]

GNU Savannah (savannah.nongnu.org)

I'd rather not move to something with an interface that looks like it was designed in 2006.

[nerzhul]

@sfan5 sorry but it's more a 2002 website :D http://git.savannah.gnu.org/cgit/administration/savane.git/plain/frontend/php/index.php

[benrob0329]

or move to SVN shit on sourceforge yeah 💛

You do know that Sourceforge has supported Git for years now, right?

[nerzhul]

it seems no, but for me sourceforge is just a SVN + ads website :)

[benrob0329]

Maybe you should get your head out of 2013 then 😛

[Wuzzy2]

yeah let's move to shitgnunah without no features, or move to SVN shit on sourceforge yeah 💛

Nerzhul, it seems to me that you are simply not willing to move out of principle and you don't even consider alternatives. Your sentence above is just FUD and you obviously did not even do the most basic research. Sourceforce actually supports Git (not that I really like Sourceforce a lot), and Savannah has clearly not “no features”, quite the contrary. You might argue “not enough features”, but that's different.

The possibility of GitLab being bought is real, but right now they seem like a very viable alternative, much better than GitHub. The migration also seems to be easy, so GitLab might be a low-hanging fruit.

The main reason why I am in favor of this is that we can finally kiss good-bye to proprietary JS. Am I really the only one who is annoyed by this?

I'd rather not move to something with an interface that looks like it was designed in 2006.

What kind of argument is that? Since when do websites have to be go with the latest web trend? Websites are not pieces of art. A website should first and foremost be functional. Great job in completely ignoring the site's function.

I don't want to defend Savannah, Sourceforge, GitLab or anything, I was mentioning them as possible options, but I do not like your attitude. Note I am not affiliated with any of these organizations, I never even hosted on any of these platforms (I am using repo.or.cz primarily), except GitLab for some “patches” many years ago.

I want to have a honest discussion, not baseless FUD.

The fact that Microsoft also owns Minecraft is ... concerning, to say the least. I think it is naive to believe that Microsoft will allow it's direct competition for long.
I strongly suggest you learn how to use the GitHub API to automatically backup all issues on GitHub daily, no matter if you actually move or not. Better safe than sorry. It's pretty easy, too: https://forum.minetest.net/viewtopic.php?f=56&t=20225#p320655
I have already backed up all issues for my MineClone 2.

I do agree, however, that there is no rush to move now.

[nerzhul]

I consider gitlab, but it's not time for this. If you don't trust me, just look at this:

https://gitlab.com/minetest/minetest/commits/master/.gitlab-ci.yml

Considering a tool as an argument for being competitive against minecraft is ridiculous. Minecraft uses git too, maybe we should use CVS ?

Other point, Gitlab can import a complete project, see MTG on gitlab, we have the issue and PR snapshot for the date it was imported.

[Wuzzy2]

Considering a tool as an argument for being competitive against minecraft is ridiculous. Minecraft uses git too, maybe we should use CVS ?

You completely missed the point, this is not about git vs CVS, it's about GitHub.

OK, I will explain it in more detail:

1. Minetest is a direct competitor to Minecraft
2. Minecraft is owned by Microsoft
3. GitHub is to be owned by Microsoft, soonish
4. Microsoft is a for-profit company
5. If you are a for-profit company operating a platform, it is a perfectly normal and natural business decision to ban at least your direct competition on it. This practice is quite common elsewhere
6. GitHub reserves the right to terminate service anytime if you are not a paying customer

Now add all of this together and there's a real chance of Minetest getting banned sooner or later. I'm not saying it's a guarantee, but it's possible. I'm saying it would be a good idea to be ready for the time if it happens.

Other point, Gitlab can import a complete project, see MTG on gitlab, we have the issue and PR snapshot for the date it was imported.

That's a start, nice. :+1:

[rubenwardy]

Survannah and repo.or.cz both have terrible user unfriendly designs - 2000 web design is not good because it doesn't have consistent and clear navigation. As for features, the very fact that you're looking for an issue tracker is telling.

I think we should either use gitlab.com or a self hosted gogs instance. Gitlab has CI and a community there, but gogs seems to have tons more polished. I don't think that it's necessary for a git platform to have CI directly part of it, so I'd prefer to go with gogs if selfhosting. I think it's worth waiting a month or so given that github hasn't even been sold yet, and maybe gitlab will be more appealing when it's less underload

[nerzhul]

gogs is far less complete than gitlab, and the gogs support is far away from gitlab.
It's nice for self hosting and not being very exigeant but Gitlab is more and more better on each release (monthly)

[sfan5]

What kind of argument is that?

It's not, it's my personal opinion on savannah's design.

Your sentence above is just FUD
I want to have a honest discussion, not baseless FUD.

In the interest of a honest discussion, you shouldn't be so quick to dismiss what other people say as FUD.

[Fixer-007]

Moving away from github at this time will cause more harm than good, there is no evidence of large-scale degradation, it is not hostile to minetest or opensource, has very large devbase. If it goes down the tube - there are viable alternatives.

[Fixer-007]

Minetest is a direct competitor to Minecraft

On android - I can agree (via ad-infested Multicraft), on PC - minetest is engine lacking some critical features with game, that looks like minecraft alpha, but without mobs, armour, interactivity, has unpolished gameplay and unsure goals, the gap is insane, there is no way it offers dangerous competition to minecraft (especially in modded). It is still good for avg folk, but MTG devs must do their act to close the gap.

If you are a for-profit company operating a platform, it is a perfectly normal and natural business decision to ban at least your direct competition on it. This practice is quite common elsewhere

This will ruin platform appeal and would be a major PR disaster.

GitHub reserves the right to terminate service anytime if you are not a paying customer

Their bus. model relies on closed repos (for money) and providing github-like solution for big enterprises (very big money), closing off FOSS projects will end up in negative PR and outflow of developers, makes no sense (unless FOSS part takes up too much server resources).

[Fixer-007]

Forgot most critical part: everyone is on github, devs are there, users are there, everyone in one place, easy to speak, post bugs, add code, moving away repos to gitlab will annoy many (needs yet another annoying registration), not everyone will register, we will end up with small dev/user participation.

[Wuzzy2]

On android - I can agree (via ad-infested Multicraft), on PC - minetest is engine lacking some critical features with game, that looks like minecraft alpha, but without mobs, armour, interactivity, the gap is insane, there is no way it offers competition to minecraft (especially in modded).

Yet our community is growing, lots of players are on servers (for a free software). Minetest is no longer a small side project. Also, Minetest Game ~= Minetest, many of these features are implemented in games or mods.

This will ruin platform appeal and would be a major PR disaster.

Maybe, but I wouldn't bet on that.

Please note I have absolutely no problems with moving to GitLab, that might be even the best option.

I never even proposed to move to repo.or.cz, I knew you would instantly hate on that. :-)

I still don't see huge problems with Savannah. Sure, it's old-fashioned, but by no means unusable. Amazing how nowadays everyone seems to be an usability expert.
Probably the biggest problem of Savannah is the lack of export features, not the interface.

As for features, the very fact that you're looking for an issue tracker is telling.

Savannah has an integrated bugtracker, right now I'm simply evaluating the possibilities. Wait, why do I even have to defend my personal use of software? :O

[Fixer-007]

Yet our community is growing

We need some proper metrics for that, serverlist has at most 300-600 players since I remember it, that is okayish for FOSS game, but it can quickly die with wrong moves (aka Urban Terror disasters).

[stujones11]

I honestly don't know what all this fuss is about. Would we be reacting the same way had google bought GH? IMO, that would have actually been worse.

I have never really understood all the MS bashing among the FOSS community, I personally have nothing against them except I choose not to use their software.

Does it really matter who owns GH anyway and how is that likely to effect us? Sorry, I'm not buying into this mass hysteria, I will be remaining with GH, at least until I have good reason to move.

[TacoTheDank]

besides, it's not even that bad, I seriously don't get the huge fuss considering Microsoft isn't actually going to change anything
read this to see what I mean:
https://techcrunch.com/2018/06/04/microsoft-has-acquired-github-for-7-5b-in-microsoft-stock/

(originally got the notification about this via the "inshorts" app, and was initially concerned, link to read the whole thing led to this)

[rubenwardy]

Microsoft has done a lot of bad stuff to open source in the past, which is why people are wary of using anything with their name on it

[stujones11]

Microsoft has done a lot of bad stuff to open source in the past

That may be true, however, their attitude towards open-source has changed a lot in recent years. They are just a highly successful business doing what successful businesses do. I do not believe they have any kind of hidden agenda.

[Fixer-007]

Microsoft has done a lot of bad stuff to open source in the past

Github is private entity and not open source at all from very beginning. Only time will show the true intentions of MS.

[stujones11]

Only time will show the true intentions of MS.

Their true intentions are to make a profit (like any business). How that effects GH only time will tell. Either way, changing hosts right now is probably a very bad idea, at least wait until the dust settles then make an informed decision on what is best for this project.

[nerzhul]

and if we wanted to move, let gitlab and other concurrents finish their infrastructure upgrade, it's a big move :p

[paramat]

We are planning to get 0.5. out within a few months anyway. However 'easy' migration may seem it will be disruption and an adjustment for core devs, it's best we leave that until after 0.5 is released, we already have our hands full enough preparing 0.5 which is a big release.

[ghost]

I'd suggest to make the move as soon as possible. It sounds reasonable to me that it's not possible in practical terms to do it before 0.5, but focusing efforts on the move soon after it's released sounds like a good idea.

@Wuzzy2:

1. Microsoft is a for-profit company
2. If you are a for-profit company operating a platform, it is a perfectly normal and natural business decision to ban at least your direct competition on it. This practice is quite common elsewhere
3. GitHub reserves the right to terminate service anytime if you are not a paying customer

MS has learned from their mistakes of the PR disaster that is doing things so blatant and clear, and these days they are much subtler. They have secretly modified Windows APIs with respect to the beta, to gain advantage over the competition, for example. In this case, e.g. searches could favour their products and make MT stuff much harder to find. Or there could be an "accident" where data was selectively erased. Hard to tell, but they will no doubt be in a position where they are actually capable of damaging the project, and I wouldn't wait for that to happen.

@rubenwardy:

Microsoft has done a lot of bad stuff to open source anyone who stood in their way in the past, which is why people are wary of using anything with their name on it

FTFY. And they keep doing it. They have a PR machine spreading the meme that they have changed, but that's not true and I doubt it ever will.

[Wuzzy2]

Hey, here's a fun thought:

How about self-hosting?

[TacoTheDank]

don't y'all already have a gitlab mirror?
https://gitlab.com/minetest

considering that, i dont think it would be THAT hard to move if you guys wanted to... but it's up to you :|

[rubenwardy]

gitlab isn't that much better than github in terms of Microsoft, given that it's hosted on Microsoft's Azure

[stujones11]

given that it's hosted on Microsoft's Azure

Didn't they drop them for google cloud, or was that just a rumour? (or maybe a crazy dream)

[TacoTheDank]

tl:dr for this whole thread: our world sucks, corporations suck, everything sucks, people are assholes

but actually tho, this is why we cant have nice things ;-;

[numberZero]

You’re not the first to suggest that :P

I think we should either use gitlab.com or a self hosted gogs
instance.
 — rubenwardy

The idea is great, but it requires some money (CI needs CPU power,
so cheapest VPS’es are certainly not suitable; not sure what our forum
runs on, though), as well as time for maintenance. OTOH more control
may help management. (BTW, I have some experience in server
administration). Also, forum/wiki/dev accounts could be merged in this
case.

So, I’m totally for such idea, but are we ready for that?

В Wed, 06 Jun 2018 13:38:42 -0700
Wuzzy [e-mail] пишет:

Hey, here's a fun thought:

How about self-hosting?

[stujones11]

@Wuzzy2 I guess one good thing about moving to something as user friendly as bitbucket is that it would deter a lot of shit-posters and nuisance issues.

[nerzhul]

Gitlab CI is free with unlimited pipelines minutes for OSS projects. CI runs on DigitalOceans and has autoscaling
Another point, Gitlab is on Azure but also on AWS. Mostly on Azure because it's the cheapest cloud currently (with some infrastructure drawbacks at a point but they mitigate it)

[lhofhansl]

FWIW, I am not too concerned about MS owning GitHub. While I'm not fan of MS, this is a move to embrace open source, not to piss everybody or anybody off.

If MS didn't like MineTest they have way more effective means at their disposal to shut it down.

[tacotexmex]

this is a move to embrace open source

What if I told you their strategy is called Embrace, Extend, Extinguish?

[tacotexmex]

What’s so bad about selfhosting? The forums are selhosted, are they not? Traditionally it’s been a hassle, but that ceased to be a problem when Docker came along.

[ghost]

What's so bad about selfhosting?

I don't think it's intrinsically bad. It has some rough edges, though. To make pull requests, people would need to register to the self-hosted platform and fork there, do their patches there, and submit from there. I imagine that that in turn means that the platform needs to allow creation of third-party repos (not sure if it's possible not to, with the current self-hosting options).

(scratched "register" per the comment/s below)

[tacotexmex]

I’m thinking that one upside of selfhosting is that SSO can be implemented throughout the MT infrastructure, which would mean that users automatically get a Gitlab/Gogs/Gitea account when signing up for the forums.

[LunaSquee]

Can confirm @tacotexmex's answer. I did just that. https://gitlab.icynet.eu (see that it has no register section, just an external login button)
I'd be totally up to lending an helping hand with setting things up if needed.

[tukkek]

Please don't move Minetest away from GitHub. I am a newcomer to the project, having just gotten my first pull request merged and I absolutely wouldn't have done so if Minetest wasn't on GitHub, which is the platform that I (and everyone else) uses to host their other projects. The same way I'm here right now, others are sure to follow - and they won't (like me) unless the project is easily accessible on GitHub.

Some thoughts:

1. How many of you are here right now because GitHub made it easy for you to participate? Everyone...?
2. What other medium-to-large sized projects are feeling the need to panic and abandon GitHub immediately?
3. There isn't actually any rational reason to leave GitHub since the news was announced, except FUD over things Microsoft did decades ago, which is ancient history in software time-scale. Kneejerking is no way to manage a project, let alone base major technical decisions on.
4. Which other platform offers the same quality as GitHub does today? None, really. Self-hosted... joke?

Finally: the main appeal of GitHub is not how many developers, testers and contributors it has brought Minetest to this day (although that is neat also) - but how many more it can bring you in the near future. Suppose Minetest 0.5 makes a big hit in a popular YouTube channel: being on GitHub can possibly bring you many dozens of modders, testers, developers and hundreds of players practically overnight - while if you're on Sourceforge, Gitlab or self-hosted (lol) then it's pretty much guaranteed that potential new members will see that as a big barrier to entry. Some here have said stuff like "small community is good" - I think that way of thinking is extremely short-sighted and does much more harm than good to the project.

All in all, if the thumbs up and downs in these comments are a signal of how people feel, I think MT stays.

[ghost]

Well, looking at the thumbs up and down, we may be split.

[tacotexmex]

Overall strong arguments by tukkek. Github is strong in a social network sense and the MT community is still dabbling with phpBB, so needless to say we might have underestimated the social factors for contributions as a whole.

[tukkek]

Well, looking at the thumbs up and down, we may be split.

No idea what thumbs you're looking at, friend. I'm seeing 6-9 thumbs up against 1 thumbs down on a handful of comments in favor of staying. I'll be the first to admit we don't have a proper tally either way but in no way things are looking split, even or neutral... not even close (as far as thumbs go).

Thank you @tacotexmex, I've also noticed you're following me since yesterday :) send me a hello if you're on Discord!

[rubenwardy]

[tukkek]

That is one comment friend, there are many others for you to look at showing a lot more support towards staying. I'm not going to post screenshots, anyone can see for themselves.

[rubenwardy]

That's not one comment, that's the opening post and the reactions for this issue

[tukkek]

No, that is cherry-picking instead of looking at all the data. I'm not going to dabble on this subject anymore, I have already said that we don't have an official tally and I'm not sure thumbs are going to make this decision or not (or what is). My point has been made and I'm not gonna dance around the issue for ten replies in a row.

[ghost]

It depends on how you read them. I read many of the most voted as "let's do it but without rushing it". Which is very reasonable.

[tacotexmex]

I’m leaning towards ”let’s wait a year and see if another git front-end with social features will emerge and gain any sort of momentum and evaluate that platform then”.

[benrob0329]

crapping on self hosted solutions as being a joke

That is just uncalled for, and quite rood to those of us who like to self-host in the spirit of owning our own data.

[rubenwardy]

Reactions on the issue are people showing their support or not for it. Reactions on replies are whether they agree with that reply, not the issue itself. More people will react to the issue rather than individual coments

[nerzhul]

gitlab has social features, and it has confidential issues for security issues, github hasn't, but github has discussions for devs. Gitlab has imported 150k projects from github in 3 days, we will see what happen, but they are very agressive, and no we iwll not move tomorrow

[tukkek]

Gitlab has imported 150k projects from github

GitHub had 67 million repositories as of 2017 so that's just a blip compared to the humongous size of the GitHub community. How many of those are just safe-keeping mirrors while the main development will continue to be hosted on GitHub (or just inactive projects) is anyone's guess too.

Regarding self-hosting: I don't mean it as an insult for people who like to self-host their stuff, I have no idea why people are taking it as a personal offense. I do think that it's ludicrous to suggest a project like Minetest should migrate from GitHub to a self-hosting solution though (based on a knee-jerk reaction to an acquisition, of all things). I fail to see a single benefit from doing so for Minetest and I do see a lot of disadvantages.

[nerzhul]

can i login with my google account on GH ? no :) and it's the most important userbase. Yeah GH has 67 million, but how many forks they have ? Don't forget a single user PR for a project means a fork, Gitlab.com doesn't have those forks currently because it's fresh upstream moves

[Wuzzy2]

Oh boy, I have so many issues with tukkek's comments. There we go:

GitHub, which is the platform that I (and everyone else) uses to host their other projects.
THIS attitude is why I hate the web so much nowadays. Everything is now called a “platform”, as if it makes something very special and this word is used to automatically deligitimize any alternative. I call it the “platform fallacy”.
GitHub is not a “platform”, it is a privately owned proprietary service built on top of Git. The word “platform” gives an impression of generality and standardization, which is absolutely not the case. The real platform would be git itself. We are not proposing a switch from Git to SVN, but just a move to a different server. This is not the end of the world.

I am a newcomer to the project, having just gotten my first pull request merged and I absolutely wouldn't have done so if Minetest wasn't on GitHub,

I don't like your attitude. It's quite arrogant. What is this, a threat?
People who are interested in contribution should learn Git, not just GitHub or any other particular server. In general, I am very pissed off when people claim they know Git when they actually just learend how to use one single website.
If you're unable to deal with repositories outside of GitHub, then you don't know how to use Git at all. There are a few good Git tutorials out there. Just learn Git and stop complaining. I mean, seriously, WTF?

The same way I'm here right now, others are sure to follow - and they won't (like me) unless the project is easily accessible on GitHub.

Guess what? Programming is NOT “user-friendly”. Knowing how to use git is an essential skill to participate, just like knowing to to code C++. And Minetest sure isn't the only project.

How many of you are here right now because GitHub made it easy for you to participate? Everyone...?

No idea, but not me.

There isn't actually any rational reason to leave GitHub since the news was announced, except FUD over things Microsoft did decades ago, which is ancient history in software time-scale. Kneejerking is no way to manage a project, let alone base major technical decisions on.

Of course there are reasons, but some of these were there before Microsoft. For me the main one is the over-use of proprietary JS. Minetest simply should not have been hosted on GitHub in the first place. I have my doubts that Microsoft is going to fix this.
Another one being the bugtracker of GitHub really sucks, especially for large projects. It does not even support severity/priorities.

Hmmm, yes, this JS bothers me. I could write an issue on the official GitHub repository to start a community project to review all the JS and getting rid of proprietary code and …! Oh, wait! GitHub itself is not free software and has no public repository, isn't it?

Also, about Microsoft …
https://www.jwz.org/blog/2018/04/in-case-youve-forgotten-microsoft-is-still-a-vile-garbage-fire-of-a-company/
https://www.theinquirer.net/inquirer/news/2409808/microsoft-bullied-mps-over-government-switch-to-open-source-standards

Which other platform offers the same quality as GitHub does today? None, really. Self-hosted... joke?

It depends what you mean with “quality”, but I guess GitLab has lots of fancy features, but I haven't looked at it carefully yet. Also, GitHub is just a website like many others. Stop calling it “platform”!
Why do I have the suspicion that no matter what alternative I am going to throw at your head, you are going to reject it if it is not a near-perfect GitHub clone?

Self-hosting was a serious suggestion. I have no idea what the hell is wrong with that.
Self-hosting has only benefits:

• Independence
• No company can screw you over
• Complete control over your own server
• Can install any software you like (hint hint: GitLab is free software!)
• Can install a real bugtracker

I give the final words to JWZ: https://www.jwz.org/blog/2018/06/lol-github/

[nerzhul]

@Wuzzy2 you just forget one thing about self-hosting: maintenance costs & hosting costs.

[ghost]

I'm on GitHub for the same reason I'm on Bitbucket and on Sourceforge (edit: or in the KDE bugtracker or the GNOME web or the Mozilla bugtracker or others): because in order to give back to certain free software projects by filing bug reports or providing patches, those projects required me to register in the corresponding webs. Now that there's a more powerful reason to leave than to stay to contribute to the projects hosted here, I'm just ceasing to submit contributions in this particular website. I'll just keep my patches all for myself instead, unless the projects provide a means for contributing that doesn't need me logging into GH.

I've already withdrawn my active PRs; fortunately Minetest was the only project where I had PRs still open. I've moved my projects and I'm now in the process of reviewing them and leaving moving notices. After some time, I'll delete them and my user.

[sfan5]

People who are interested in contribution should learn Git, not just GitHub or any other particular server

I'm pretty sure the intended argument here is about network effects.
If I want to report fix a trivial bug in a tiny project, but am required to register on some random persons personal gitlab instance, I will not bother.

This applies to larger projects too, while core contributors will move over to a new platform*, the chance of someone "stumbling upon" the project or already having an account to easily report bugs are significantly lowered.

Of course there are reasons, but some of these were there before Microsoft. For me the main one is the over-use of proprietary JS.

Not everyone subscribes to the radical definition interpretation of free software, and you can't except nor force anyone to.

I'll just keep my patches all for myself instead, unless the projects provide a means for contributing that doesn't need me logging into GH.

Previously the README(?) suggested sending your patches to celeron55 by email, not sure if it still does. But I don't think coredevs would mind getting patches sent by e-mail or posted on IRC.

*: If you continue to insist that Github and others are not a "platform", I suggest you to review a dictionary.

[Fixer-007]

Future Github CEO Ask me anything:
https://www.reddit.com/r/AMA/comments/8pc8mf/im_nat_friedman_future_ceo_of_github_ama/e0a26fk/

[Fixer-007]

After some time, I'll delete them and my user.

Delete your user on every proprietary platform or just GH?

[ghost]

How do the devs feel about this? I think there opinion matters the most.

[nerzhul]

@ChristianSirolli using selfhosting will just break the link with potential contributors and the git platform, which is important for MT. Using Github or gitlab is fine for me, i have a very high preference for gitlab because they are more competitive and their tool is more powerful than github

[tacotexmex]

I'm pretty sure the intended argument here is about network effects.
If I want to report fix a trivial bug in a tiny project, but am required to register on some random persons personal gitlab instance, I will not bother.

This applies to larger projects too, while core contributors will move over to a new platform*, the chance of someone "stumbling upon" the project or already having an account to easily report bugs are significantly lowered.

For me this is the argument at its core. Network effect vs platform purity. While an avid supporter of selfhosting personally I’m not sure I can support a move of this project’s hosting. Also, shaming coders for not being adept with Github and not git I guess has its place, but it won’t increase network effects towards Minetest.

Regarding using (hosted) Gitlab.com instead of Github is, to me, equal to going down the selfhost route. There’s no network effect on Gitlab.com anyway.

[Wuzzy2]

Of course there are reasons, but some of these were there before Microsoft. For me the main one is the over-use of proprietary JS.

Not everyone subscribes to the radical definition of free software

It's the only definition of free software and it has existed for decades. Also, the definition of open source is basically the same. Besides, if Minetest is not about free software, then what is? I don't get it why so many core devs seem to be almost opposed to the idea. But I digress.

Even if we'd go by the very low standards of “source-available”, proprietary JS goes against that.

I'm just ceasing to submit contributions in this particular website. I'll just keep my patches all for myself instead, unless the projects provide a means for contributing that doesn't need me logging into GH.

Wow, respect. I should probably take the same steps. I have accepted this GitHub nonsense for far too long now.

[TacoTheDank]

watching this play out with valid arguments on both sides and no end in sight

[TacoTheDank]

oh i just found this article, its a very good read about GitHub's future, i suggest you guys read it
https://www.theregister.co.uk/2018/06/04/nadella_tells_worried_github_devs_judge_us_by_our_actions/

some quotes from it: "No one spends billions of dollars – $7.5bn in this case – on something because they think it's great and they just want to protect it. Microsoft wants a hefty return on investment and it is counting on devs' inertia in shifting to a new platform to deliver it."

"Microsoft completely rearchitectured Skype and tried to force everyone into paying it more money for existing services that are slowly being pulled behind a paywall. LinkedIn provides valuable data. And GitHub is a combination of the two – a source of data and an opportunity to make money by pushing Microsoft services."

oh also Nat Freidman (future GitHub CEO) is doing a Reddit AMA
https://www.reddit.com/r/AMA/comments/8pc8mf/im_nat_friedman_future_ceo_of_github_ama/
Although he may not be completely unfiltered, because Microsoft's team likely wouldn't let him be, it's still reassuring to see what he wants for the community.

[tukkek]

Now that there's a more powerful reason to leave than to stay

Can you tell us what that reason is?

[ghost]

Can you tell us what that reason is?

That's somewhat off topic for this thread, as that's a personal decision, but basically the same reasons I never ever considered joining CodePlex or contributing to projects hosted there: it's owned by the same predatory, promises-breaking, bullying, antitrust-laws-violating, monopolistic company and I don't want to be part of their statistics, so I'll cease my contributions here.

More on-topic, I'm also concerned about the potential harm they can do now that they are in such a powerful position, so I also wholeheartedly recommend MT moving to other more free platforms, or at least that aren't currently owned by such company.

[sapier]

I've been inactive by some time but I feel same way github owned by Microsoft is not a place I want to be any longer.

I don't know if it does help but I'm already moving to gitlab and can confirm the project import does quite a good job. Including wiki and issue import.

[tacotexmex]

I have to admit, seeing modders disappear from Github and scatter all across the internet sure feels like defeat and one can imagine the disruptive effect it may have on the Minetest community. Luckily Rubenwardy’s Content DB arrives just in time to possibly bring all mods with various git hosting sources together.

[stujones11]

@sapier Sorry to see you go, you have done a lot of great things for this project. Just remember, you can always still email your patches to celeron55 ;^)

Seriously though guys, get over yourselves, this stuff happened like 20 years ago...

IDC what anyone says, GH has served us all well and I see no 'sane' reason why that would not continue. A few old sayings come to mind.

• Cutting off the nose to spite the face.
• Biting the hand that feeds.
• Throwing your toys out the pram.
• The grass is always greener.

I'm sure I could think of more :)

@tacotexmex Sadly, that is the worst part of all this. We are sure to lose a good deal of contributors and maybe even devs whichever way things go :'(

[tukkek]

I'm also concerned about the potential harm they can do now that they are in such a powerful position

So the ""powerful reason"" you have stated for abandoning GitHub is that you personally don't like a company very much? Sounds about right for this thread. I wish you guys could hear how absurd you sound when you make absolute claims like these based on nothing. Literally nothing has changed on GitHub... lol

[stujones11]

inb4 google acquires gitlab :P

[ghost]

Literally nothing has changed on GitHub... lol

Do as you wish, I don't want to keep my hand in the grinder once it's plugged, no matter the promises that it won't be turned on... by someone who has ground more than one hand.

[benrob0329]

I really don't understand this thread. People are wary of Microsoft because of the pain they have caused the FOSS community, and people are wary of a quick reaction because of the pain it could cause the project. But neither side wants to see the other's point.

[Fixer-007]

Here is real consequence of MS acquisition of github, people like @pgimeno who had pull requests that went into the trash (I imagine, there is "full" room of people who care to adopt that PR), and no way to communicate with original PR author in reasonable way.

[numberZero]

I’m leaving GitHub. I’ll probably stay here till 0.5, but not any longer.

@Fixer-007 take a look at Skype. It wasn’t even open source. But it was a nice program nevertheless. Nowadays, it’s pretty unusable: slow, full of context ads, and lacks features (or at least non-tricky ways to set them up). Well, just like many MS apps, especially nowadays.

Such changes are not instant, that would be a PR disaster, as you said. But gradually, it will become less and less comfortable (it requires tones of proprietary JS already, but that’s not the limit), and probably less usable for libre software development as well (there are various ways for that, of various severity). That’s inevitable.

[rubenwardy]

There's also 4 versions of skype, or something

[Fixer-007]

Another opinion from LinuxFoundation blog:
https://www.linuxfoundation.org/blog/microsoft-buys-github-the-linux-foundations-reaction/

[ghost]

they are backers of The Linux Foundation

'nuff said. Not the most impartial opinion.

[tukkek]

@rubenwardy said yesterday on Discord that Minetest is leaving GitHub. I'm sad to see the project leave (over nothing) and glad I at least got a commit in before you guys left :) best of luck and I wonder why they're keeping this thread open and still discussing here when the issue's been resolved.

[rubenwardy]

Lol, I was trolling. I also said we're adding IAPs and social media support

[4w-zz]

I’m leaving GitHub. I’ll probably stay here till 0.5, but not any longer.

Yep, same here. I already moved to GitLab and archived my most relevant repositories and deleted the others. The archived repositories will be deleted within the next 1-2 weeks and then the account will be left unattended.

It would be awesome having Minetest on GitLab (or – even better – MT maintainers should host an own GitLab instance on own servers, 100% control and everything MT related is nicely consolidated). If not, well, then we have to wait until Microsoft screws up GitHub (based on their history we all know that this will happen).

[paramat]

@numberZero please consider keeping your MT repo on Github until we process your currently open PRs, there are many good ones and some are complex, it will help us a lot. I will try to make these high priority for review.

Obviously i can understand contributors moving all their other repos immediately but it will help MT if contributors keep just their MT repos on Github until their existing PRs are processed or there are actual changes that force them to leave (which won't be for many months hopefully).
Otherwise contributors are harming MT for no actual practical reason.
The deal will not be finalised for months, then i expect changes will take a while to start.

[numberZero]

@paramat, as I said,

I’ll probably stay here till 0.5

I don’t plan to drop my repo with all the PRs, that would be too disruptive even for me };->

[tukkek]

I was trolling

Come on man, don't mess with my frail heart like that X( we love Minetest over here, please stay!

Also wasn't going to acknowledge it but what the hell: love it when people who are on the losing side of a reasonable argument start posting meme images implying vague stuff instead of actually reasoning and explaining their motivation. Shitposting > defending your side because you're outta arguments. Gotta love it!

[nerzhul]

@4w we have a mirror there, but no possibility to make PR merged at this point. For self host it's a noop for me, we don't have time/resources to make it possible. It's either github.com or gitlab.com, selfhosting a tiny project like MT (YEAH mt is tiny compared to gnome or freedesktop.org infrastructure) will just block it from having new contributors/visiblity

[davisonio]

Honestly, this is a terrible idea. GitHub is where most contributors are at and it makes it easy to find new contributors for fixes.

In fact, Microsoft contributes quite a lot to open source and this acquisition is having no effect at the moment apart from drawing away people who are still stuck at hating Microsoft from the 90s...times have changed. Let's wait this one out.

[benrob0329]

still stuck at hating Microsoft from the 90s

First, they've done things far more recently than the 90s. Second, people who have been burned tend to avoid the stove, even after it's been turned off.

Frankly, I find it appalling that with Microsoft's history people are treating those weary as 'knee jerk haters who don't know any better' (my own quote from the attitude of some people). While I don't believe Microsoft hasn't changed, I also don't believe they have. Contributing shows an interest, what scares me is the motive for said interest.

[tacotexmex]

Regardless of Microsoft's interests I think people tend to overlook the massive disrupting impact moving may have on contributions for Minetest. If the strategy of Microsoft would be to rid the platform of FOSS projects it's working, but only because of their own self-mutilation in becoming invisible to the majority of developers on the planet by moving.

Let's adopt a pragmatic position in the interest of Minetest specifically, then we may see how well it can adapt to the more idealistic view long-term.

[tukkek]

people who have been burned tend to avoid the stove, even after it's been turned off

Do you realize how stupid it is to be afraid that a stove will burn you when it's turned off? xD Why would you even say that?

'knee jerk haters who don't know any better'

The thing no one here has been to reply up to this point, here in this thread or elsewhere, that makes the entire point is: on a worst case scenario, what harm can Microsoft possibly do to Minetest through GitHub?

It's amazing this issue has survived for so long when no one has answered this essential question in any way that approaches a reasonable explanation for moving away from GitHub - which is why, all of this is a big knee-jerk, panic-driven reaction to a news story that affects absolutely no one on GitHub. It's absurd.

[ghost]

My personal position is certainly not one of panic. More like JWZ's: https://www.jwz.org/xscreensaver/xscreensaver-windows.html

I didn't panic out, I stormed out out of anger against GH for letting this happen. MS are monopolistic, and I don't want to help them make their monopoly even bigger. Making their monopoly bigger is what harms all free software including Minetest. There's your answer, happy now?

[tukkek]

There's your answer, happy now?

My question very clearly was: on a worst case scenario, what harm can Microsoft possibly do to Minetest through GitHub? So no, I'm not at all satisfied since you didn't even begin to address my question.

[Fixer-007]

Moving away from github will harm the project, it is no brainer, you loose drive-by bug reporters and contributors.

[lhofhansl]

In case it was not clear from my earlier comment I am _against_ moving MT away from github.

[numberZero]

Summarizing the comments, the only GitHub advantage is its popularity, but that is critical for us.

[tukkek]

Summarizing the comments:

1. GitHub is the best code hosting and collaboration platform and it's been widely agreed here that all other options are subpar in one way or another;
2. Nobody in this thread has been able to provide a single direct, objective explanation of why Minetest should migrate from GitHub or what harm Microsoft can inflict to the project in short, mid or long terms;
3. One of the biggest GitHub advantages is its popularity, which is critical for us;

[shimmark]

If anybody has evidence that the increased contributions projects get from being on GitHub are not negligible, I'm listening. If they are negligible, which is my impression, then GitHub's popularity is not critical for Minetest and the increased openness and superior features of GitLab, along with the misalignment of soon-to-be Microsoft-owned GitHub's incentives with Minetest's goals, are a reason to migrate.

[ghost]

The summary of the comments made by the people that matter are:

Does seem to me now a case of 'when' instead of 'if', however there's no rush, we will likely have a few months to consider a move.
(paramat)

Using Github or gitlab is fine for me, i have a very high preference for gitlab because they are more competitive and their tool is more powerful than github
For self host it's a noop for me, we don't have time/resources to make it possible. It's either github.com or gitlab.com
(nerzhul)

I think we should either use gitlab.com or a self hosted gogs instance.
(rubenwardy)

I've been inactive by some time but I feel same way github owned by Microsoft is not a place I want to be any longer.
(sapier)

together with the general idea of waiting a few months.

The summary of the comments made by people that don't matter are:

MS :+1: :tada:
MS :-1: :zap: :anger: :boom:
GH :+1: :heart:
GH :-1: :angry: :imp:

[ghost]

One of MS specialities is changing APIs to make them more convoluted, in order to harm competition.

One of the possible first moves by MS could be to change the export APIs in order to make it difficult or impossible to migrate the data to their competition, and lock people down on GH without being able to export.

Food for thought.

[tukkek]

changing APIs to make them more convoluted

Great, I'll be the first to support a migration on the day APIs become convoluted.

change the export APIs

Great, I'll be the first to support a migration on the day data exporting becomes convoluted. Until then, make weekly backups just to be sure and we'll lose nothing.

Both of those arguments are very weak: say we move to GitLab, for example, what is preventing GitLab from making APIs or data exports more convoluted over there two days from now? Nothing. Conjecture scenarios are not actual arguments, especially when they don't harm the project even if they do happen.

comments made by the people that matter

This is way beyond stupid. This is a technical issue, no one cares about the people behind the ideas.

[tukkek]

evidence that the increased contributions projects get from being on GitHub

Here you go, easy. Do you think Minetest would have over 300 contributors if it wasn't on GitHub? If you think so, find us similar projects with as many contributors on other code hosting/collaboration platforms to prove it.

I for one, am only here because it's GitHub. I'm assuming 95% of you are on the same boat too lol

[rubenwardy]

I'm here because Minetest is here - I found this site and signed up here in 2012 due to Minetest being here. If Minetest were on some other site I would have gone there. Contributing a patch is much harder than signing up for a site, the bigger thing to worry about are issue reporters rather than (patch) contributors

[tukkek]

Regarding the number of contributors I linked before, I forgot to add an important metric: the fifth most active contributor in this repository has ~280 commits under his name. This means that, with 300+ contributors for the project as a whole, even if each only had a single commit, "the GitHub community" would already be the fifth most active contributor to the entire project. If that doesn't convince someone that the size of the GitHub community is a massive factor here, then they're choosing not to see it, simple as that.

I'm here because Minetest is here

That's fair Ruben but I just opened 3 people at random from the comment list here and they've all had other projects under their GitHub profiles so I'm afraid you're the exception here, not the norm. Of the 3 of them, 2 had over 30 repositories on GitHub.

I know it's a small group I've clicked through but I think it's only common sense that most people are here because being on GitHub made it easy to contribute to Minetest, not the other way around (even if that's the case for a few of us here). If someone wants to do the experiment with everyone on the list, by all means, be my guest - I'm sure it'll only prove this point even further.

[shimmark]

Do you think Minetest would have over 300 contributors if it wasn't on GitHub? If you think so, find us similar projects with as many contributors on other code hosting/collaboration platforms to prove it.

Sure. The first project I looked at, 0 A.D., uses Subversion on a self-hosted Trac instance. Certainly this is as unpopular and hard to use as it can get. The 0 A.D. community is also at most a third the size of Minetest's, and probably less.

And yet. The version control system for 0 A.D. has 21,696 revisions while this GitHub repository has only 7,457 commits. Minetest has 339 contributors according to the repository's page while the revision log for 0 A.D. has 114 different author names. After glancing, I found that only three looked like duplicates and two like bots. So 109 contributors, in a community that is smaller than Minetest's and that, I am not kidding you, uses Subversion and self-hosted Trac. And the revision count, compared to Minetest's commit count, would lead you to believe that they contribute much more than the average Minetest contributor, too. Though we should account for the fact that 0 A.D. has existed for longer than Minetest.

Maybe Minetest should switch to Subversion.

[tukkek]

So you were able to look at the entire Internet and all you found was a project that is backed by a commercial company and it's still a third of the size of Minetest at GitHub? I guess that proves my point then.

[shimmark]

Wildfire Games gets its income from donations through Software in the Public Interest. They don't pay people to develop 0 A.D., unless you count the less than $500 they spend on travel for meetups as development. That 0 A.D. is “backed by a commercial company” is misleading in this context and in any case does not explain the number of contributors and their number of contributions.

You also assumed that 95% of Minetest contributors were only here because of GitHub, which makes a 95:5 ratio or 19:1. If that were the case and the projects were otherwise comparable, you should have expected Minetest to have not three times but nineteen times more contributors than 0 A.D.

What's more, looking at how many contributors very different projects have cannot prove (or disprove, which I never claimed it did) your point about the software and website used to develop the project. The purpose of this is to know how important the popularity of the website used to host the repository and manage issues and code reviews is in relation to other factors that influence the number of contributors and contributions. If you assume that the differences between two projects are due to whether they're hosted on GitHub, then you are assuming the very thing it is you claimed you were providing evidence for.

[ghost]

If it matters, which it doesn't as Mark has said, GIMP has 359 contributors to date (including me) and it was not even hosted in a social code platform until this year, in which they have installed a self-hosted GitLab instance.

[Fixer-007]

I was here not because of github too, people were already there and urged me to stop posting bugs on forum and move to github instead, so I did.

[appgurueu]

I would only start moving if GitHub was starting to show bad signs, eg ads appear or so. Else GitHub is better, you'd break links if you attempted to move, plus it would be an effort. Maybe we should make a online clone repository on GitLab that is updated regularly by some script ?

But note : MOVING ANYWHERE ELSE WOULD REQUIRE THE ENTIRE COMMUNITY TO MOVE, TOO !

So I think it's not an option for the closer future.

(That hooray emoji below is meant sarcastic, thinking of our current situation)

[numberZero]

@appgurueu The clone is there already. What we may need is to advertise it, and maybe (not sooner than 5.0) to consider it as the primary repo and GH’s one as a clone.

[paramat]

I agree that self hosting is not realistic. It may be ideal, but we are already critically low on core dev time and we don't have any money to spare.
What will happen is somewhat decided by c55 already, and core devs are practical and in no rush to leave, so it seems we will move when or if there is reason to do so.

[4w-zz]

Basically "So let's ignore the shit piling up around us. Just try to sneak away if the shit collapses."

We'll see.

[appgurueu]

4w : You summarized it pretty well.

[TacoTheDank]

tbh he probably summarized it the best

[paramat]

It's the usual negative, cynical and and rude comments from 4w.

There is no 'sh*t' yet, the deal hasn't been finalised. They say they hope to finalise it by the end of the year, it's just an agreement to a deal. So MS does not yet own GH, no bad changes have been made to GH yet, the GH staff is probably mostly unchanged.

I can understand individuals leaving GH to send a strong protest message that they disapprove of future deal. But moving MT is different, it's is a big disruptive task that can't be rushed, and is best left until after 5.0.0. We are already critically low on core dev time and can't add more disruption at the moment.
We are being responsible, practical, calm and rational about this, and are doing what is best for MT.

My personal opinion is: MS 'seem' (i am no expert) to be an unpleasant company. I am very unhappy they bought GH and it makes me uncomfortable. I expect they will slowly make GH worse, maybe after an initial period of trying to appear to be 'hands-off' (as with Mojang). I expect MT will leave, but it will probably be at least 6 months before that happens or has to happen.

[tukkek]

it will probably be at least 6 months before that happens or has to happen

Completely void, wild guess with no basis in any sort of concrete reality except Microsoft's ancient history. The fact people keep saying absurd stuff like that and expect to be taken seriously is completely crazy to me.

Why 6 months, what could MS possibly do to hurt MT or GH (in a way in which we can't just pick our stuff an leave if/when we become displeased)? I've asked this question n times here and no one can answer it still.

I hope most people here understand that there is zero rationale behind leaving, only fear, uncertainty and doubt over a non-issue - and a issue that, if, in the future presents itself, still allows us to migrate at that date.

[benrob0329]

Why 6 months, what could MS possibly do to hurt MT or GH (in a way in which we can't just pick our stuff an leave if/when we become displeased)?

See, that question is inherently flawed. We don't know what they can and cannot do. I could throw ideas out, like a paywall, advertisements, data selling, requirement of DRM utilities so that source code can't be copied off the site. But we just don't know, so being prepared is the best option.

Their "ancient history" is more recent than you think, they still develop and support Windows 10 (which multiple people have done packet sniffing to see that its doing shady things). Just a couple years ago, Skype for Linux was still cripplingly out of date without a word (causing much workplace humiliation when things don't work, even when people payed for it) and today breaks with almost every update.

Quite frankly, I don't understand how people are so quick to trust Microsoft. The dark, foggy past is enough for a second thought. While the fog is lighter it's still there, and still enough to make me keep my distance.

[tukkek]

You say "that question is inherently flawed", I read "I agree that we cannot possibly argue the most fundamental question about why should we move away from GitHub". Semantics? I think not.

You say "DRM utilities so that source code can't be copied off the site", I read "I lack the basic understanding of how git works".

It's comments like those trying to pass as actual arguments that keep me from taking the doomsayers seriously. Sorry!

[benrob0329]

You say "I read" I see "I changed what you said to better suit my argument".

I read "I agree that we cannot possibly argue the most fundamental question about why should we move away from GitHub"

No, the most fundamental reason is Microsoft's history of screwing people over, and that has been stated multiple times.

I read "I lack the basic understanding of how git works"

Oh? Because while I do understand how git works, its somewhat useless if you don't have permission to clone a repo, and your git hosting site encourages shared source, rather than open source. But trying to argue the ideas of how they could ruin Github is silly when you ignore the actual tangible evidence we have; Microsoft's history. The likes of which you seem to like to call "ancient" despite the fact that they still do crap to this day, so unless you consider your own message "ancient history" I recommend you rectify your argument for why we should ignore it :-)

[numberZero]

I don't understand how people are so quick to trust

@benrob0329 Because companies and governments do everything to ensure that.

Anyway, as devs decided not to move until the release (for a very good reason), and as we have a place (GitLab) where we may move at any time, if desired, I ask everyone here to code now and to continue this discussion only after the release.

[ghost]

Moving discussion from https://github.com/minetest/minetest/issues/8071#issuecomment-454796081 to not derail that thread.

@paramat

Note however that after 5.0.0, I'm going to delete my GitHub user, so I won't be able to comment here.

Nothing bad has happened to anyone yet as a result of the MS buyout, so please don't.
MS may well be malicious in some ways but they may never actually act maliciously through Github, and have not yet.
What risk is there in keeping a Github account simply to comment here? Keeping your code elsewhere is fine because we can import patches easily enough, but not being able to comment here will make contributing very difficult.

Isn't it important to continue to support MT development, at least as long as there is no actual harm being done to you by MS? You're a talented and much valued contributor, we don't want to lose you.
The highly ironic fact is that militant anti-MS attitudes amongst some MT contributors are significantly harming MT while MS actually hasn't harmed MT yet.

MS have already acted maliciously. They have started to provide free private repositories, which was the main competitive advantage that Bitbucket and GitLab had over GitHub. They are trying to increase their monopoly in this area by suffocating all competition to death. Monopoly means there are no options, and having no options harms consumers. That includes me.

The actual harm will come later, when people already have no other choices. That's the purest MS style.

image source

I wish MT did something, but I'm not going to push for it. However, I as an user have an option, and my option is to not support this process, by fleeing from here and supporting places that will badly need it in a short time, if not already.

I have decided to stay until 5.0.0 is released, because I want to minimize the harm to MT. But there are things more important than MT for me. I'm just sad that MT looks at this with the immediate future in mind, rather than with a wider perspective.

[ghost]

Also, implementing #7901 would be a solution.