05:38:51 repro builds wouldn't prevent any of that 05:39:04 people downloading don't check hashes, and that's not going to change 05:39:16 the auto-updater is the closest thing we can do to checking it for them 09:10:53 ^ also true 09:12:38 all that repro builds do is prevent the person finalising the build from compromising it 09:12:55 it's not like when binaryFate uploads a binary any of us go and disassemble it 09:13:02 ain't nobody got time fo date 09:13:03 *dat 09:13:11 fluffypony: btw do you have any idea how we could move the mac archive from tar.bz2 to .dmg while not breaking updater? 09:13:22 a redirect maybe? not sure if the software would follow it 09:13:29 hmmmm 09:13:35 we could test a redirect 09:13:52 the problem is that you can automatically unpack a tar.bz2 and overwrite the existing one 09:14:02 I don't know if there are automation tools for handling a DMG 09:14:05 I mean, there must be 09:14:10 but the existing code won't know what to do with it 09:14:24 don’t think we have this code 09:14:28 or not aware of it 09:14:35 the updater only downloads the new file 09:14:36 do the notary tools insist on producing a .dmg? 09:14:39 yes 09:14:40 I thought they produce .apps 09:14:46 and the .app is signe 09:14:48 *signed 09:14:48 .dmg or .zip 09:15:05 we can unpack the .zip and repack it as .tar.bz2 without losing the sig 09:15:12 I'm 25% certain 09:15:13 lol 09:15:20 yea does not work lol 09:15:35 blah 09:15:36 the .dmg gets signed and the .app 09:15:40 apparently 09:15:46 but the .zip doesn't get signed 09:15:50 or is it a zip of a dmg 09:16:02 no, but there were issues with the .zip 09:16:09 that I don’t remember currently 09:16:15 .dmg was better 09:17:05 are these signed per-machine or a single universal sig? 09:17:40 (app store app installs are signed for that machine only, afaik) 09:20:34 I think universal 09:20:50 I remember the issue with .zip now, you can’t staple the result 09:21:04 meaning that apple’s server get pinged 09:21:25 with dmg you can staple the "signature" and it also works offline without pinging apple 09:21:46 we had the idea of doing it in two steps 09:22:10 add the code to download .dmg instead of .tar.bz2 in the next release 09:22:21 and then the release afterwards we bundle it as dmg 09:22:33 but this will only work for those who updated to the latest version 10:11:39 I think we can expect teething issues for now 10:11:40 so don't expect perfect forward compatibility 10:11:53 we can add the code in, and do the next one as a .tar.bz2 10:11:56 and then .dmg thereafter 10:12:11 we just need to automatically mount the dmg -> copy the .app to where the current one is 10:12:15 (which may not be /Applications) 10:20:17 fluffypony: the auto updater only verifies the file and opens the file explorer, no auto extracting currently 10:20:30 so should be ok with .dmg 10:21:14 ah ok 10:21:20 is the plan to go full auto-update eventually? 10:21:33 maybe, probably not trivial to do cross platform 10:22:02 would require a second binary that spins up -> deletes existing app -> replaces it with the new one -> restarts app 10:24:33 would be a nice to have in the future 10:29:11 possibly make this a seperate project that is able to manage both monero-gui and monero binaries (starting, stopping, updating) 10:30:19 we need to create Steam/Battle.net, for crypto 10:31:17 half joking, half not 13:34:32 dsc_: https://github.com/moneromooo-monero/monero-update 13:34:50 needs to be picked up, but moneromooo already started it 13:35:20 if someone wants to run with it I'm happy to re-home it to the main GitHub 13:36:05 I have a patch for it to work with the GUI. I'm just waiting for repro builds to push it. 17:42:02 i was thinking of only checking the binary hashes from two or more locations and thats it. with immutable git repo to prevent what pony was talking about backdating commits and repo builds when that happens. ill stop bringing it up as the updater is already overkill.a 17:42:09 * i was thinking of only checking the binary hashes from two or more locations and thats it. with immutable git repo to prevent what pony was talking about backdating commits and repo builds when that happens. ill stop bringing it up as the updater is already overkill. 20:31:21 .merges 20:31:21 -xmr-pr- #2933 #2937 #2947 #2950 #2959 #2961 #2962 #2963 21:13:05 2933 conflicts 21:13:29 xiphon: ^ 21:17:24 2963 also 21:32:09 we might have to revert 2947, it seemed like a simple patch and I also tested it a bit but it doesn’t fully work as expected. 21:54:44 luigi1111w: resolved merge conflicts