Hans-Christoph Steiner @eighthave@librem.one
Joined Aug 2019
For anyone who wants to relive the #Alphabet #DMAWorkshop the video has been released:
#DMA #compliance #competition #EC #EuropeanCommission #EU #gatekeepers #monopoly #mobile #Google #AI #GooglePlay #GooglePlayServices #GoogleFree #EDRi #FSFE #FDroid
A simple metric from #FDroid #metrics data: app downloads per week. Start with data from 1 of 2 servers for f-droid.org: http02, add hits for paths ending in ".apk". That gave about 2 million. Multiply by 18 (fronters + mirrors) and get ~36 mil app downloads a week.
import requests
hits = 0
r = requests.get(f'https://fdroid.gitlab.io/metrics/http03.fdroid.net/2025-05-26.json')
data = r.json()
for path in data['paths']:
if path.endswith('.apk'):
hits += data['paths'][path]['hits']
print('APKs', hits)
https://forum.f-droid.org/t/experiment-in-f-droid-org-metrics/32454
#Google funds this #EU think tank to put out policy papers saying #DigitalMarketsAct will break their lovely #PlayProtect scare screens, making us all less safe and "it require[s] Google to allow developers to insert links inside their Play Store apps".
https://ecipe.org/publications/eu-dma-undermine-security-mobile-operating-systems/#_ftn13
As I've always said in relation to the #DMA, let @fdroidorg compete on trustworthiness. I'd love to see this think thank include analysis malware rates of #CalyxOS with #FDroid and compare that to #GooglePlay #security
We're starting to implement support for split APKs in #FDroid. #Google wants to gather as much data about its users as possible, so trying to hide info about language, country, device specs was not a design concern for them. It is central for us. We want the official client to leak as little data as possible to any server, be it ours, mirrors, or custom repos. We welcome input:
https://gitlab.com/fdroid/fdroidclient/-/issues/2963
#Android #APK #privacy #DataEfficiency #efficiency #technology #data #metadata
The first time I went to FOSDEM, I was constantly explaining what F-Droid was to people who never heard of it. I was introducing it and promoting it.
Second FOSDEM, people said "oh yeah, I heard of it, what does it do?"
Third FOSDEM, people said "I'm a user!"
Fourth FOSDEM, I listened to people tell other people what F-Droid is, and heard groups of people talking about it as the app store that everyone in the group had on their phone.
It has been quite a ride!
On my own time, I have to read a ~50 page document produced for the #EuropeanCommission in order to effectively participate in a two hour meeting where #FDroid is pitted against #BigTech on the #DigitalMarketsAct and its requirements around installing and allowing other #AppStore options.
Its all NDA'ed so I can't ask for help.
This game is really rigged for the megacorps. Wish me luck! Here's to fighting the good fight!
Three years ago, #FDroid had a similar kind of attempt as the #xz #backdoor. A new contributor submitted a merge request to improve the search, which was oft requested but the maintainers hadn't found time to work on. There was also pressure from other random accounts to merge it. In the end, it became clear that it added a #SQLinjection #vuln. In this case, we managed to catch it before it was merged. Since similar tactics were used, I think its relevant now
The #DigitalMarketsAct and other #competition actions against #gatekeeper app stores are based on the idea that an app store companies should not "self-preference" their own apps or services. This makes sense to a certain degree, especially when thinking about business. Ethical reasons must also be considered. #FDroid preferences apps based on #FreeSoftware and Anti-Features, which we as a community define. We should always be allowed to preference apps that follow standards of #UserFreedom.
I wish the #AndroidSDK team would follow repository best practices and stop silently reissuing binary releases under the same name/version. #MavenCentral does not allow this, for example. The #FDroid transparency log shows the newest violation: two version of sources-34_r01.zip with the file name, version code, and metadata.
git fsck makes it much harder to attack a git repo, but it seems that the normal git workflow does not enable it by default. In #FDroid it is enabled for all fetches in our config:
https://f-droid.org/docs/FAQ_-_App_Developers/#how-can-i-handle-fsck-error-in-packed-object-for-my-app
But I still can't find a clear answer about what checks #Git does by default. Anyone know?
I got the opportunity to go to #FOSDEM and of course I had my #FDroid "hat" on. I wrote up some quick impressions of my trip, including what I learned about the #EU's #DigitalMarketsAct #CyberResilienceAct and #ProductLiabilityDirective
There really are a lot of important projects represented there:
https://f-droid.org/2024/02/06/at-fosdem.html
Based on @maarten 's post https://blog.nlnetlabs.nl/what-i-learned-in-brussels-the-cyber-resilience-act/ I think the only people listed in my example that would be at all regulated by the #CRA would be the last one: "contracted contributors". It sounds like they might be considered "open source software stewards" with obligations under Article 17a depending on whether the #EU considers F-Droid as "intended for commercial activities"
https://www.cyberresilienceact.eu/the-cyber-resilience-act/
My guess is #Nextcloud/#Ubuntu would be considered commercial while #FDroid/#Debian would not
@roptat thanks for your nice visualization of #l18n in apps in #FDroid
https://i18n.lepiller.eu/i18n.html it is interesting to see which languages are the most active. If you're interested in more data sources, there are a lot of public data sources:
https://f-droid.org/docs/All_our_APIs/
For example, you might enjoy looking at the most popular search queries with the included language and country data:
https://fdroid.gitlab.io/metrics/search.f-droid.org/2024-01-29.json
After #FOSDEM my current understanding of how #EU #CRA and #PLD affects #FDroid and anyone who contributes to it:
* F-Droid org makes the "product" so it would be liable
* F-Droid is currently entirely non-commercial, handles no money
* Volunteer contributors are very clearly exempt from all this
* Donation funded contributions are also exempt
* Contracted contributors are helping build the regulated product, so the legal entities of the contractors would not be liable for F-Droid's "product"
So I messing around a bit more with stats data from the f-droid.org mirror hosted by @FAU in Germany:
* About 25% of the bandwidth of the mirror is for #FDroid
* Countries by percentage total bandwidth usage since /fdroid/ was added in 2019:
47.83% Germany
9.04% United States
5.13% China
3.68% Italy
2.69% France
2.55% United Kingdom
2.07% Russia
1.93% Switzerland
1.87% Estonia
1.82% Poland
1.70% Austria
1.40% Netherlands
1.38% Czechia
1.28% Canada
1/
#FDroid is consistently growing in its bandwidth usage over the years, as shown by this stats graph from the #UniFAU mirror. Interesting to see the short downward section when we added new official mirrors in April and November.
Thanks @FAU for the mirror, the bandwidth, and the stats! https://ftp.fau.de/cgi-bin/show-ftp-stats.cgi
@jcaleitao #FDroid has been trying to use your #IPFS gateway, but it does not seem to be working. https://ipfs.joaoleitao.org/ipfs/bafybeifx7yeb55armcsxwwitkymga5xf53dxiarykms3ygqic223w5sk3m returns "502 Bad Gateway" but that CID works on other gateways. Are you still maintaining your gateway? Are there restrictions on what it will host? f-droid.org is all free open source software apps, so should be uncontroversial.
In my work with #FDroid I've discussed our work with gov regulators for South Africa, UK, EU and Japan as well as competition litigators from multiple US States and the EU. From this, I'm starting to see a picture of #Apple's and #Google's semi-related strategies of making "sideloading" (installing apps outside of their #gatekeeper control) look bad as a way to keep their monopolies in the face of #DMA and other regulatory actions. I'm still looking for data about the actual real world risks 1/
Joined Aug 2019
