Aleksandra Fedorova :fedora:

CI/Devops/Infra Engineer, worked on RHEL & CentOS Stream
Manager of the Anaconda and Cockpit teams
Fedora Ambassador, former FESCo and Fedora Council member
PhD in Geometry and Topology
she/her

Matrix: @bookwar:fedora.im

I'll leave this account for FOSS debates.

For more personal stuff I am going to use @bookwar

Feel free to follow one or the other, both or none.

--
I am not a native speaker, and I appreciate private notes correcting my English or German.

Public Key

npub12u3h3dchshl60nl3qwqylucf323l87d5yxhd403gsgq8jfjafqcsscq0uk

NIP-05 Address

[email protected]

Profile Code

nprofile1qqs9wgmckutctla8elcs8qz07vyc4glnlx6zrtk6hc5gyqreyew5svgpz4mhxue69uhhyetvv9ujumt0wd68ytnsw43qhvvetq

Publishing to

relay.mostr.pub

Author Public Key

npub12u3h3dchshl60nl3qwqylucf323l87d5yxhd403gsgq8jfjafqcsscq0uk

Show more details

Last Notes

2025-04-23 23:11:09 GMT

- reply

@nprofile…9k7h Right, I meant people nominated for the seat. Willing to take the place. And the turn out for the vote too.

2025-04-23 15:06:33 GMT

- reply

"But I am not running myself, what can I do?"
1) Think about who can represent your voice in #Fedora
2) *Tell them* that you think that they might be good at representing your voice in Fedora.
This is a critically important step, because often people who are good for these roles are not aware that they are good at them.
3) Learn about people nominated for elections. Talk to them, ask their opinion on various topics.
4) Vote.

2025-04-23 15:06:04 GMT

We tend to not care about elections, until it becomes too late to care about elections. And FOSS projects are no different.
For years we have very low attendance in Fedora Council, and only slightly better attendance at FESCo.
Thus, #Fedora people, please pay attention. We are entering the new Fedora election cycle and you should take a look at it.
https://discussion.fedoraproject.org/t/fedora-elections-nominations-open-april-24/150255

2025-04-23 11:33:39 GMT

- reply

How should it be addressed?
Are there insurance schemas which are applicable here?

2025-04-23 11:32:16 GMT

One of the big non-technical blockers for the "Indy IT" is handling risks and liabilities.
The argument goes like this: "we can not buy IT services from a small vendor, because if the vendor makes a mistake, we get nothing".
The assumption here is that if Google leaks your sensitive information, the settlement between you and Google is going to cover the costs for your entire life and then some.
While with a small vendor their entire budget will not be able to cover your liabilities.

2025-04-22 16:06:42 GMT

- reply

@nprofile…nvxe @nprofile…f43w
Yes indeed. Treat is "I fully support you!" message.
I do feel like we have outgrown the "open source" concept. Not that it became redundant, quite the opposite, it is necessary. But it became not enough.
This is why, for example, I like the Four Opens from OpenStack (https://www.openstack.org/four-opens/).
But we need more than a license, more than the OSI. Some new definitions are necessary.

2025-04-22 15:43:44 GMT

- reply

@nprofile…nvxe
Some time ago, in the middle of the pandemic we were sitting at some virtual conference hallway track (a midnight video call meeting) with @nprofile…f43w discussing this exact issue.
And as a result we created a repo https://pagure.io/sos-sig/
where we wanted to award badges and track projects which have simple deployment and configuration options so that they can be hosted without five layers of Docker containerization and cloud tools.
We haven't made any progress on that thing..

2025-04-17 19:32:43 GMT

- reply

@nprofile…6m2n
Thanks for the post.
The only time I use Ctrl-S in the terminal, is when I forget that I am in vim and not in Emacs and try to save the file with Ctrl-X Ctrl-S shortcut.
You can guess the outcome, and amount of cursing happening while I am searching the Internet for the ways to undo it :)

2025-04-15 23:00:53 GMT

If the only thing Phoronix commenters are complaining about is a wallpaper, it means the release is good, isn't it? :)
#Fedora #Fedora42

2025-04-13 11:46:43 GMT

- reply

@nprofile…vnh5 and @nprofile…s2e4 need to do a joint project and write Ada Book 2: "Ada's grandmother and Copyrighted Stitches".

2025-04-13 11:45:54 GMT

The more I get into crafts like crochet, sewing or knitting, the more I think that is is our hugest mistake that we let FOSS movement to be focused on software and become so disconnected from everything else we do.
If you are a FOSS geek, your grandmother knitting a pair of socks is supposed to be your best ally: All the same issues, all the same challenges and all the same joy.

2025-04-09 21:47:39 GMT

Local universities are some of the best places for building FOSS communities.
We need more of that, we would love to help you to work on that.
Get inspired and build your community, too
https://fedoramagazine.org/were-in-kenya-fedora-at-muranga-university/
#fedora

2025-04-09 14:02:06 GMT

- reply

@nprofile…3sex
Bureaucracy captcha for the win.
Fill the fields of a Bugzilla ticket according to Fedora guidelines, to prove you are not a robot.

2025-04-07 21:02:52 GMT

- reply

@nprofile…ac0z @nprofile…nvxe
Ah, Inception!
I absolutely love that moment when they let the girl to dream and she immediately starts "messing with the physics".
Like, yeah, you could dream yourself walking in Paris, but where is fun in that. Turning the world inside out, though.. this is interesting.
https://www.youtube.com/watch?v=dG22TcpjRnY

2025-04-07 20:30:12 GMT

- reply

@nprofile…nvxe
1. Alien 2
2. 2001 A Space Odyssey
3. Enemy Mine https://www.imdb.com/title/tt0089092/
4. Planetes https://www.imdb.com/title/tt0816398/

2025-04-06 17:24:35 GMT

How do you handle Nextcloud and local files sync?
For example I have ~/Music/ ~/Books ~/Notes ~/Documents
And then ~/Nextcloud synced to the server.
Should I move ~/Music to ~/Nextcloud/Music? Or make a symlink from ~/Nextcloud/Music to ~/Music ?
And there is also an external USB drive at /run/media/bookwar/Storage with its own Music/ folder, which should also somehow fit into the scheme..

2025-04-05 00:10:07 GMT

- reply

@nprofile…ltr0 Yeah, it might be. I definitely experimented with it at some point.

2025-04-04 22:21:02 GMT

Just look at that.
I think this was Gnome in 2013.
Vertical panel never actually worked correctly, there were always some rendering issues with the window list, but I liked that setup a lot.
And that background is one of my favorites. I have it on my current system and I was looking into ordering a printed fabric with it just couple of months ago.
#Fedora
https://cdn.fosstodon.org/media_attachments/files/114/282/002/484/257/585/original/c11d1cff4f5a2505.png

2025-04-04 20:15:06 GMT

- reply

Same with the stack of protocols.
I hate the OSI model so much because so many people start to explain the concept of protocols by introducing 7-layer model, which no one ever needed.
While you could start explaining the concept just by using an example of an actual mail written on paper, which you then put into an envelope.

2025-04-04 20:08:33 GMT

- reply

Network routes was my first and main barrier of using Linux back then, because we had a local university network and a PPTP connection to the outer world. There was no easy way to configure it, without some `ip route` magic.
All available resources, like Wikipedia, were, and still are, written in a completely inaccessible way. And only my father was able to explain the simple math behind the routing table on a napkin.

2025-04-04 20:08:01 GMT

Doing final takeout of Google Photos today, an, oh my, the memories it brings..
For example this picture, uploaded Nov 15, 2013, is one of the first diagrams I made in Inkscape explaining to myself what is IP, MAC Address, and subnet mask.
https://cdn.fosstodon.org/media_attachments/files/114/281/450/237/571/099/original/248f85d9a175946b.png

2025-04-02 21:58:16 GMT

- reply

@nprofile…nc7a
Not only they disrespect the work of the people storing and providing the information. The damage they do is so high that the open projects now have to artificially create "paywalls" and hide the open non-copyrighted content behind them.
This is the damage done to the core of our existence - the ability to share the non-monetized knowledge, code, or art, without a middleman trying to assign cost to it.

2025-04-02 21:52:16 GMT

- reply

@nprofile…nc7a In this case, the problem is not the copyright.
Indeed the information is open and is _designed_ to be shared. Almost everyone, take Wikipedia, or Linux Kernel or OpenStreetMaps, or Fedora intentionally provide ways for anyone interested to fetch a full archive of the data.
But instead of using those effective ways of fetching the public information, AI people use completely unnecessary technics of scraping the content, which causes significant costs to projects who publish it.

2025-04-02 15:39:41 GMT

This aggressive AI-scraping of FOSS infra is quite radicalizing.
Instead of taxing the rich, we tax the people who do work for free and publish for everyone to use.
I am considering starting a campaign to add
"Any engineer working for a company who was using such a practices is banned for life from participating in any FOSS community" into the default Code of Conduct.
Because if there is one universal evil in FOSS, than that's the one.

2025-03-24 12:32:08 GMT

- reply

@nprofile…t8rs
I am guilty of the first one. I think middle-click is not a useful button, since it became a click on a scroll wheel. It somehow became too much effort to really click it, without scrolling a little bit.
And Ctrl-click requires a hand on a keyboard. Which is not that typical when you are heavily browsing stuff.

2025-03-23 21:28:58 GMT

Got my ticket to https://why2025.org/
I know no one there. But I do have a tent. And it is an outdoor camp event in Netherlands in August - it must be good, right? :)
#why2025

2025-03-19 09:58:24 GMT

- reply

@nprofile…ak7a
We are working on this in the open here:
https://discussion.fedoraproject.org/t/ai-policy-in-fedora-wip/144297
Everyone is welcome to participate in a constructive way.
@nprofile…l86s

2025-03-19 09:58:08 GMT

- reply

@nprofile…ak7a
And what about it?
Yes, Red Hat invests in development of (really open) AI technologies.
And Fedora is the place where you showcase the latest and greatest FOSS technologies, and is an obvious goal for Red Hat development teams to present their features. There is no secret about it.
Of course Fedora community also is concerned about state and use of GenAI and LLM tools. That's why AI Policy is an important topic for the Fedora Council to define.
@nprofile…l86s

2025-03-18 20:33:32 GMT

Fedora is a community-driven project sponsored by Red Hat.
By writing news like "Red Hat presented Fedora KDE Edition" you do not just misrepresent the Fedora Project, you also undermine the effort of Fedora community members who did all the work to make KDE Spin into an Edition.
Please do recognize and give a credit to the community.
And if you have questions about Fedora governance, we are here - at discussion.fedoraproject.org, at release parties, on social networks, just ask
#Fedora #KDE

2025-03-18 13:28:12 GMT

I have been iterating on this idea for some time.
And I think I have found an easy way to test the GitOps approach in Fedora without interrupting any of the existing work, while still bringing some value in the process.
https://discussion.fedoraproject.org/t/rfc-new-community-initiative-gitops-for-fedora-packaging/146990
Your comments and opinions are welcome.
But volunteering to work with me on this initiative is welcome twice more :)
#Fedora

2025-03-18 13:20:31 GMT

- reply

@nprofile…f43w
I wish I were more capable in the legal things. I feel like Open Infra Europe is needed. It actually has been needed for a long time now...
@nprofile…nvxe

2025-03-18 13:17:02 GMT

- reply

@nprofile…nvxe
Time for OpenStack and all the other FOSS projects, which we knew for years are needed and critically important to shine.

2025-03-18 11:56:31 GMT

- reply

Generally, writing a new project is the easiest step in FOSS. Volunteers can do it 100 times per day and they will be successful in it.
Much more successful than the government-funded academics by the way.
Government funding and government help is need in the *next* steps: unlocking the ecosystem, building the adoption, educating users, setting up governance models, setting up security and sustainability..

2025-03-18 11:53:00 GMT

- reply

Yes, existing projects are not ideal, they have issues and legacy, and history.
But
1) they are real.
2) Any new project you fund will get the same issues over time. By launching new project, you are not bypassing the complexity of dealing with FOSS projects and communities. You set yourself up for repeating the path people already went through.

2025-03-18 11:51:31 GMT

French government does exactly what I expected and exactly what I wish we could avoid.
We should not launch new government-funded FOSS project, we should find existing projects and invest government funding in them.
But the way funds work, it is much easier to ask for money for a yet another greenfield project, which will be years behind the available alternatives than to convince people that existing projects need maintenance and investment in the ecosystem.

2025-03-17 10:01:13 GMT

- reply

@nprofile…f8d6
At some point I wanted to use this particular thing as a Software Engineer job interview question.
But then decided against it, as it is a bit unfair to play mind games like this in the interview environment.

2025-03-17 09:31:55 GMT

- reply

And yes, nowadays it becomes a story about anything vs AI :)

2025-03-17 09:30:59 GMT

- reply

When I grew up, it became a joke about Pure Maths and Applied Maths.
Professor at the Uni gave us a task of calculating the curvature of the torus. And a friend went that long path of calculating the curvature via metric functions.
But torus have sections with two different circles, and you can get the curvature with zero calculation from that fact.
That day I understood that real Maths is all about "cheating" your way out of calculations.

2025-03-17 09:25:04 GMT

There was a joke once on what's the difference between a mathematician and a programmer:
You ask to multiply all numbers from -10 to 57.
Programmer will write a script which will multiply numbers in a loop. Mathematician will say: it's 0.

2025-03-14 13:05:55 GMT

- reply

@nprofile…qvhx
Good try :)
It is really unfortunate that the word 'maintain' is such a mouthful - it doesn't stick well to anything.

2025-03-14 12:37:25 GMT

- reply

@nprofile…qvhx
No, it was already a stretch but you are taking it too far.
I mean, yes, compose is a lock file and technically anyone can create a compose and doesn't have to be a distribution for that.
But the distributions role is to produce composes, which work = meet a certain set of criteria.

2025-03-13 17:55:46 GMT

Hot take or not hot take
The main function of the distribution is not to create packages, but to create composes.
---
P.S. In Fedora-speak, compose = "a compound artifact usually contains netinst.iso, Live iso, qcow image,.. and several package repositories"

2025-03-13 16:11:22 GMT

- reply

@nprofile…qvhx
Of course I also have to be honest, and say - no, I do not know a recipe to win in the development advocacy game.
I bet I advocate for Zuul CI longer than you advocate for NixOS and we know where I am with it :)
The Ideal above is ideal not because it allows you to win. But because it enriches both parties in the conversation.
And because building a shared understanding of the problem can cause that other ecosystem to drift slightly in the direction to narrow the chasm.

2025-03-13 15:59:28 GMT

- reply

@nprofile…qvhx
> One can't sell Linux to Windows users that way.
I'd say this is exactly _how_ you sell Linux to Windows users. By asking them and identifying that their main obstacle is gaming and vendor lock on Microsoft Office, enforced by a third-party, and addressing those.
If you check the German news, right now a certain German government moves its infrastructure to use LibreOffice _before_ moving to Linux. And I think that's a right thing to do.

2025-03-13 11:43:12 GMT

FOSS Dev Advocacy:
Worst:
Mere mortals! You will never comprehend the real power of The Tool.
Bad:
Our tool does everything so much better than yours.
Better:
Our tool can do this and that.
The Ideal:
In the implementation you have chosen, you will have this and that problem. In our tool we dealt with that differently by using this and that approach. You can try to implement our solution partially, but it will take some time. Which obstacles prevent you from using our tool in this way?

2025-03-11 21:28:57 GMT

- reply

@nprofile…g800
@nprofile…e0sg happens every year in Bonn in August.
Half of the talks are in German, but the environment and the audience is English-friendly

2025-03-11 21:18:41 GMT

From @nprofile…jwmj on Discussion forum:
https://discussion.fedoraproject.org/t/containerfiles-considered-harmful/146812
I haven't used mkosi tool, but I do share a sentiment on Containerfiles being unmanageable.
#DevOps
https://cdn.fosstodon.org/media_attachments/files/114/145/869/778/479/001/original/f9b3419b5e71f713.png

2025-03-05 14:23:45 GMT

- reply

@nprofile…3udr
I had a manager once who said that "improve" is a forbidden word in tasks, PR descriptions and status reports.
Whenever you feel like writing "improved" you should stop and write what exactly has happened or is expected to happen instead.

2025-02-09 20:43:00 GMT

- reply

@nprofile…g800 Fun fact: In my early Linux days I knew nothing about Linux but somehow I knew your name in it. Thank you for being there.
Not very fun fact I learned on the job: anxiety doesn't go away with experience.
You just learn to accept and accommodate for it:
Drink camomile tea not coffee, disconnect and do not talk to people an hour before the talk, do the "power walk" or a comforting speech to yourself..
Whatever trick helps, do not be afraid to use it, however "silly" it might look.

2025-02-08 13:33:32 GMT

Let me use an opportunity for PSA:
Everyone interested in discussing AI on #Fedora - you are welcome at discussion.fedoraproject.org.
I am currently heading home from a very packed Fedora Council meetup with the notes and action items on that topic.
We have some ideas and we will need your feedback.
Spoiler: "there will be no AI assistant in Fedora by default" is one of the items in the draft.

2025-02-08 13:00:46 GMT

The Düsseldorf(Weeze) low-cost airport is not only 1,5 hours away from the actual Düsseldorf city. They also built an airport in Germany without direct train connection. You can only get out of there by bus.
I guess they got so many complaints from people with wrong expectations that they renamed the airport and now it is called Niederrhein - "lower Rhein" - accepting the fact that this is an airport in the middle of nowhere somewhere in the region.

2025-02-07 07:33:30 GMT

I need a postcard which says:
I went to FOSDEM and all I got is yet another list of To Do items.

2025-02-06 20:31:09 GMT

- reply

@nprofile…3udr it definitely isn't.
https://cdn.fosstodon.org/media_attachments/files/113/958/827/438/628/980/original/2594b305719f70ba.jpg

2025-02-06 20:26:00 GMT

#Fedora Council meetup is going great :)
https://cdn.fosstodon.org/media_attachments/files/113/958/817/428/098/289/original/34d15da3c89fa5e8.jpg

2025-02-01 12:56:53 GMT

#Codeberg / #Forgejo booth is well hidden in the UA building behind the MySQL and GitLab banners.
Go find them and say hi.
#FOSDEM #fosdem2025

2025-02-01 12:51:38 GMT

#FOSDEM Mission accomplished:
Fosdem t-shirt, Mastodon t-shirt and bag of waffles.
https://cdn.fosstodon.org/media_attachments/files/113/928/715/298/642/838/original/543014d68db021e8.jpg

2025-01-30 10:06:18 GMT

It is going to be dry and sunny, they said.
But we know that FOSDEM simply doesn't work this way.
Hello FOSDEM crowd, I will be at #CentOS Connect in an hour.

2025-01-28 21:26:19 GMT

- reply

@nprofile…vvfe
> AI haters build tarpits to trap and trick AI scrapers...
> AI haters
The words you looked for are "AI-survivors".

2025-01-28 14:40:08 GMT

- reply

@nprofile…ne93
This is what social networks are for, right?
...
Right?

2025-01-27 19:02:20 GMT

What would be a good topic for #Fedora Flock conference?
I would suggest anything about a "day 2" contributor challenges - things you do *after* you learn how to write a spec-file.
---
- How to survive maintaining your package over multiple upstream releases and one big rename
- Things I do daily, which are not described in any of the Packager guides
- Landing a Change and lessons learned from it.
---
Btw, CfP is open
https://fedoramagazine.org/flock-to-fedora-2025-prague-june-5th-8th/
Note: I am not on the program committee.

2025-01-24 16:59:03 GMT

- reply

@nprofile…9k7h
Additionally to that, I think the point of having themes is more or less to have a teaser for when people do not have better ideas.
Seasoned contributors do not need to be teased. And from my perspective it was a given that if you have a topic, which you think is important to be discussed with a Fedora contributor community, you just do it.
Maybe it is worth to highlight it more explicitly there.
@nprofile…t2vs

2025-01-24 16:52:34 GMT

- reply

@nprofile…f43w @nprofile…9k7h @nprofile…t2vs
Can you provide it as a feedback directly to Fedora Council along with your suggestion how to change it?
Ticket or a forum post will do. It is quite an open process, and it is not too hard to patch it

2025-01-24 16:37:18 GMT

- reply

@nprofile…9k7h
My internal translation engine reads it as:
* Tools practices and processes used by the project
* Onboarding contributors to the project
* New features coming in the next Fedora releases.
Which seems to be the most generic set of topics a project can come up with.
Can you give an example of a topic which would be interesting to you, and not fitting into these three?
@nprofile…t2vs

2025-01-24 14:47:05 GMT

- reply

@nprofile…t2vs
It would be nice to see you there.
This is a more contributor rather than user-oriented conference, so the topics on development practices, how to work with upstream or what future features to look forward to can be interesting.

2025-01-17 19:01:30 GMT

I might be not getting it. And you can try to change my mind.
But when American billionaire comes to Brussels to talk to European FOSS community trying to convince them that he is a good Open Source player, it does seem like a big thing, which deserves the keynote level talk at FOSDEM.
And we should congratulate each other and FOSDEM orgs, that at its 25th anniversary FOSDEM reached that level, where it gives us the opportunity to (peacefully) boo that billionaire.
Am I too optimistic?

2025-01-17 10:32:40 GMT

- reply

@nprofile…wtue
One of my first jobs was a SysAdmin on duty watching after monitoring systems in a large hosting provider. I have my share of scary stories but also proud moments on that job :)
If you think you will do it better while self-hosting, go for it. It is a personal choice and personal commitment.
I know i won't. And I hope that by paying a provider, which is a Platinum partner for Nextcloud, I also support the Nextcloud development a little bit.
@nprofile…q6xl

2025-01-16 11:14:42 GMT

- reply

@nprofile…q6xl Yes, same here. And that's exactly what that managed offering does.
You bring your own domain name, and they give you everything else with access to Nextcloud admin panel.

2025-01-16 11:06:32 GMT

- reply

@nprofile…q6xl
You can get managed Nextcloud from a provider like IONOS for example: https://www.ionos.de/office-loesungen/managed-nextcloud-hosting
It doesn't mean you can't be hacked, but it won't be your fault :)

2025-01-13 19:18:35 GMT

- reply

@nprofile…cfw2
"We are willing to work on this issue further but we are not able to, because we are waiting for the info from you"
I think I would not perceive "need_info" as a sign of urgency or priority, just a sign of - the work won't move forward without additional input from this person.
Some other variable (for example Severity) should be responsible to determine how critical is that.
So I guess my answer would be: 2)

2025-01-10 09:44:31 GMT

- reply

@nprofile…jrff @nprofile…zvft
While duplication and fragmentation of a discussion makes it harder to coordinate, it makes it more accessible.
Local communities give you much better sense of belonging and ownership. And it gives you much wider and diverse opinions.
If you discuss the same topic five times in five different environments, you learn five times more about the topic, then if you would have discussed it in a single environment even with the same five sides present.

2025-01-10 09:39:51 GMT

- reply

@nprofile…jrff @nprofile…zvft
I have been thinking about this for a while.
Especially with Matrix, IRC, Discord, Telegram. All these channels are essentially the same chats. So they don't even have that difference in speed and volume, like Matrix Vs Mailing lists. Thus it is natural to try to bridge them all into a single global chat space.
But then I came to conclusion that it is not the right goal.
1/2

2025-01-09 21:14:47 GMT

- reply

@nprofile…jrff Yes, different communication channels create different local communities on each of them.
I don't think it is possible to prevent it completely. But it is probably worth investing some time in setting up the information exchange policies, like the rule that meeting logs from the chat must be sent to mailing list/discussion or that certain mailing list discussions could have the "open hours" on chat for synchronous conversions.

2025-01-08 16:52:08 GMT

- reply

@nprofile…gnt3
Check https://torrent.fedoraproject.org/
You can get there from getfedora.org if you click on "Other Downloads" button.

2025-01-06 13:58:10 GMT

The "culture fit" question at a job interview should be:
"when do you think it is appropriate to use `channel` and `here` handles in a Slack chat?"

2025-01-06 10:55:43 GMT

I have mentioned it before, but it is worth mentioning again.
Zealdocs.org is a very cool project, which gives you offline access to all the dev docs you could possibly need. No ads, no Google AI "suggestions", no cookie banners.
I am on the train with a very much nonexistent wi-fi today, and it saves my day.

2024-12-25 13:44:37 GMT

- reply

@nprofile…t2vs @nprofile…zcq4 @nprofile…h9qy @nprofile…05jc yes, we will fix it in the article, once folks get online.
But MIT or GPL was not a critical difference in that discussion.

2024-12-24 10:25:39 GMT

- reply

@nprofile…nvxe
Right. I have a bit of a personal issue with this as I had to do a mass change in CentOS Stream once, which led me to fork about 600 projects to send a one-line change in each of them. And there is no way in GitLab UI how I can bulk delete them now.

2024-12-24 10:04:57 GMT

- reply

@nprofile…nvxe I still think that "Clone and Pull" model would be superior for drive-by contributors. "Fork" usually means overhead of long-term maintenance of your own "version" of a project in your own space, while often you just want to send a change request one time, not to become a real admin of your own full copy.
But that's an implementation detail.

2024-12-23 13:01:12 GMT

Das @nprofile…fsry
https://cdn.fosstodon.org/media_attachments/files/113/702/248/195/373/162/original/18b8b6d3b18f8ebb.jpg

2024-12-22 11:33:12 GMT

- reply

@nprofile…nvxe @nprofile…qe73
With my Fedora Council hat on, I'd say it is a win for Fedora.
It has been a long open discussion in the project whether tools to make the project should be considered part of it. This discussion is far from done, but I am glad that we are recognising the value of FOSS and collaboration on the development tools, not just the content being developed.

2024-12-13 18:02:37 GMT

- reply

@nprofile…k3w4
I am personally more on the M-x query-replace-regexp side. You know, just to be safe.

2024-12-13 17:54:09 GMT

- reply

Q: Have you ever seen a successful valid refactoring effort in real life?
A: As a matter of fact, I made one. The key to success was that I had not enough knowledge about the project therefore my refactoring scope was limited by external reasons.
The refactoring was - to remove Windows-style line breaks from a 200000-lines Java project :)
The second refactoring, to format that project to one code style, failed on 3).
Apparently auto-formatting tools in Java IDEs are not always idempotent.

2024-12-13 15:33:52 GMT

How to differentiate a valid refactoring effort from an infinite yak-shaving loop?
...
1) Refactoring effort should have a limited scope defined before you get into the weeds.
2) The fully refactored state of the project(within that scope) should be theoretically possible.
3) The refactoring action should be idempotent: attempt to refactor the refactored code should not produce any changes.
4) There should be one active refactoring effort at a time.

2024-12-12 22:32:30 GMT

>Dear Aleksandra,
>
>Thank you for registering as a supporter of the FSFE! We greatly appreciate your support and your trust in us.
Thank you for your work, FSFE!
https://fsfe.org/index.en.html

2024-12-11 11:26:23 GMT

- reply

@nprofile…3sex
Same for COPR , Koji and for ex. Testing Farm. Do each of those systems need to scale the executor pool into various clouds and different hardware pools? Yes, they do.
Is it necessary for each of them to develop a custom implementation of this functionality - I doubt that.
Shared solution like nodepool could be a better way https://zuulci.org/docs/nodepool/latest/
Of course running it still requires resources and effort from the team managing the service. And we shouldn't underestimate that.

2024-12-11 11:20:32 GMT

- reply

@nprofile…3sex
Sorry, but no, this arguments doesn't work for me.
Does every GitForge need an integration with a scalable build system? Yes.
But does every GitForge needs a custom in-house implementation of a scalable orchestration engine? No. We would be in a much better place if we didn't have build engines vendor-locked to each particular Git Forge, but rather have the ability to interoperate and reuse the existing solutions for that.

2024-12-11 09:33:06 GMT

- reply

@nprofile…3sex
I wasn't trying to undermine the value of systems like COPR.
I am arguing that being a build system is not a defining feature for them.
If you look at GitHub, it runs tons of builds per day. Do we call GitHub a build system then? No, we call it GitForge, because while building things is a necessary part of it, it is not what it _is_.
There are many scalable systems which can execute build recipes in the cloud, only one of them is COPR because it doesn't just build things on request.

2024-12-10 20:18:05 GMT

- reply

@nprofile…3sex
I don't see much difference between build system and generic "execute bash script in an isolated environment and give me all the receipts" system.
It is the management of the produced artifacts, which make a difference, not the scaling of containerized executors into the cloud.

2024-12-10 17:32:07 GMT

- reply

And one day someone will develop a Change-Request-as-a-Service or EverythingHub.
And you will have Change Requests, pipelines, labels, tags, CI validation, reviews, comments, issues, rebases, changelogs, diffs..
and you will be able to choose a backend which might be a Git, or Nextcloud Images or Google Docs folder or Ravelry Crochet patterns or the Grundgezetz.

2024-12-10 17:23:50 GMT

- reply

@nprofile…ne93
The fact that you are changing jobs not as a result of a fire but more as a result of getting to a certain level of confidence in the future of the project - is an achievement on its own. Highly appreciated!

2024-12-10 15:52:53 GMT

- reply

And often the reason why we want #GitOps is not because we want Git, but because we want Change Request.
So we have to map our non-code artifacts into Git somehow, to get the Merge Request functionality "for free".
While we should really just have more artifact storages/registries/file storages/doc collections.. to provide Merge(or Change) Request functionality and interface by default.
(Review Board https://www.reviewboard.org/ tried to do it right)

2024-12-10 15:38:26 GMT

- reply

More gates
https://cdn.fosstodon.org/media_attachments/files/113/629/255/316/675/452/original/0d3b681151573560.png

2024-12-10 14:46:46 GMT

Hot take:
---
No one needs a build system.
Everyone needs a gating system (= artifact storage with gating and promotion mechanics).
---
Git Merge Request is just an example with artifact storage being Git repository, and promotion being push from feature-branch to stable branch.
All other non-Git types of artifacts need a Change Request too.

2024-12-09 12:04:15 GMT

- reply

Also, yes, I am going to have my PTOs soon.

2024-12-09 12:03:24 GMT

- reply

I have to say, this would be a really good problem to have :)
When I woke up I tried so hard to remember what was that totally valid and important reason, why everyone should have been on Fedora Council..
And I was not able too. :(
2/2

2024-12-09 12:02:36 GMT

Having a dream about work is not a good sign. This time it was quite funny though.
The main part of it was that suddenly all Red Hat Subteams got a need to have a representative in the Fedora Council.
And I was like: wait, there is a difference between Fedora Council and Fedora Parlament!
And we were about to brainstorm the options of what we can achieve and how.
Somehow we needed to build a catwalk into the meeting room for that.
And then I woke up.
1/2

2024-12-04 17:08:08 GMT

- reply

@nprofile…5rql I think this is a generic setting for the cli library. It recognises the option by enough symbols so that it is uniquely defined.
--she will probably work too. But then everything will break if later you add a legitimate --shelve option.

2024-12-04 00:04:59 GMT

- reply

@nprofile…qvhx @nprofile…xn96 Ok, let's forget about steps.
How about:
Pipeline has Input and Output.
Conveyor with gates is not a good pipeline, because it may skip or delay the Output.
The outcome of the pipeline shouldn't(*) be influenced by external events beyond Input. The pipeline itself shouldn't(*) not have the side effects beyond the Output.
(*) For a spherical pipeline in a vacuum.

2024-12-03 23:32:11 GMT

- reply

@nprofile…xn96 The definition you give covers any i/o operation, doesn't it?
So there has to be something else which makes us reach for the word pipeline instead of "operation", "action" or "process" and other synonymous terms.
And in my mind i always think about CI pipeline as a sequence of steps. Even though the word itself suggests a certain fluidity. I guess this is caused by the concept of Jenkins pipelines imprinted in my brain by now.

2024-12-03 22:41:52 GMT

- reply

How about:
1) Pipeline is a sequence of transformations, which is isolated from sideeffects. Same input produces the same result.
Or
2) Pipeline is a process represented by a linear sequence of steps. No "forks", or cycles, every input goes through the same steps in the same order.