24 Jun 2022 |
uniq | back when that code was written check updates wasn't running on gitlab ci. | 21:30:41 |
uniq | A scheduled task like check updates just for fetching new reproducible signatures would be pretty great actually | 21:31:40 |
uniq | an idea along those lines was to only fetch signatures from 3rd party fdroid repos | 21:32:16 |
Sylvia | I wonder what's the best place to make a GitLab issue to track this. Given GitLab CI might fix it fdroiddata seems logical but it feels like it'll get lost there easily | 21:41:14 |
uniq | I'd track it in fdroidserver, since that feature requires code changes in that project | 21:43:57 |
Sylvia | Hmm, yeah I guess that'd be cleaner than having fdroiddata CI create metadata entries | 21:50:29 |
uniq | I think would make sense to work roughly similar to check updates | 21:57:48 |
uniq | code in fdroidserver, ci job in fdroiddata | 21:58:09 |
Sylvia | https://gitlab.com/fdroid/fdroidserver/-/issues/1013 | 22:05:35 |
grote | In reply to @SylvieLorxu:matrix.org Well, specifically, you can apparently use fdroid signatures to get signatures that somehow F-Droid can use to check reproducibility. And if it matches it will publish 2 APKs. But when using Binaries: it will only publish the upstream APK. Why can't the signature be extracted when the Binaries: statement is there and also publish 2 APKs (for easier upgrading)? In the case of Briar for example, it always started with its own sig and never wants builds with other sigs circulating. Publishing 2 APKs is a bridge technology for apps that F-Droid had once signed itself. | 22:48:22 |
Sylvia | I wasn't aware of reproducible builds existed when I submitted my apps to F-Droid (which is probably part of the reason the number is so low) | 22:49:11 |
Sylvia | Probably many people are in this situation :P | 22:50:05 |
grote | The sad truth is that reproducible builds cause even more issues to F-Droid publishing than there's already. | 22:50:15 |
grote | Apart from getting an app to build at all, it needs to be reproducible as well and in the past this broke frequently either due to fdroidserver changes or upstream android changes. | 22:51:42 |
Sylvia | Sounds painful | 22:51:57 |
grote | Yep, but like Hans I like to still dream of a future where there"s lots of independent app builders and fdroid publishes apps that enough (trusted) people verified to be reproducible. Then you could update the index several times a day :) | 22:53:56 |
Sylvia | Unrelated but for those who have Mailcow setup already: they're currently having some server issues so that's why you're getting errors. Not really affecting us yet anyway because it's still not live | 22:58:32 |
proletarius101 | In reply to @SylvieLorxu:matrix.org So 23 total I think it's also worthwhile to suggest the upstream use our docker image so they may not need to turn of r8 or anything that ppl usually use | 23:44:51 |
25 Jun 2022 |
linsui | In reply to @_oftc_Licaon_Kter[xmpp]:matrix.org linsui: flexboru ended in an odd way, wow, open source crack lol They can say that 5 months ago... | 04:49:43 |
Licaon_Kter[xmpp] | Eh, you linsui went in too excited, wanted to merge a cool app (I guess lol) but the "pay features unlocked" part is tricky. | 05:19:16 |
Licaon_Kter[xmpp] | But they are right on this needing a fork or just plain move to Archive now. | 05:21:05 |
linsui | Yes, I'm going to archive it. | 05:21:32 |
proletarius101 | I don't think pay to unlock is against any spirits of open source or free software, as long as they don't embed a proprietary library | 06:09:12 |
jochensp | I think pay to unlock is against the spirit of free software and if I would care enough for the app I would fork it and patch that out | 06:11:52 |
Licaon_Kter[xmpp] | I think the spirit still lives when you fork. I don't think it helps to host it if upstream doesn't want us. And if they don't want it's "fake foss" eg. Signal | 06:31:02 |
proletarius101 | In reply to @_oftc_jochensp:matrix.org I think pay to unlock is against the spirit of free software and if I would care enough for the app I would fork it and patch that out At least it's encouraged by GNU | 08:37:31 |
proletarius101 | And if you think it's not acceptable on fdroid, well we have many to take down | 08:37:50 |
proletarius101 | First is Fairemail | 08:38:05 |
Licaon_Kter[xmpp] | Netguard too, etc | 09:04:29 |
Licaon_Kter[xmpp] | proletarius101, jochensp: do read https://forum.f-droid.org/t/paid-features-in-opensource-apps/4099 ;) | 10:06:01 |