16:00:19 #startmeeting tor anti-censorship meeting 16:00:19 Meeting started Thu Jun 10 16:00:19 2021 UTC. The chair is cohosh. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:00:19 Useful Commands: #action #agreed #help #info #idea #link #topic. 16:00:25 here's our meeting pad: https://pad.riseup.net/p/tor-anti-censorship-keep 16:01:30 hello 16:02:07 feel free to add items to the agenda 16:03:02 dcf1: i think that first item is yours, do you want to lead the discussion? 16:03:33 meskio, I see you're working on gettor, does your work touch the uploading part of it? 16:03:58 yes, the plan is to rewrite it so it works automatically, now is a manual process and is never up to date 16:04:04 for reference, this is the script with the ia upload: https://gitlab.torproject.org/tpo/anti-censorship/gettor-project/gettor/-/blob/main/scripts/update_files 16:04:16 oh thanks 16:04:29 ia? 16:04:44 every time I look at https://archive.org/details/@gettor it seems strange 16:04:51 ia == Internet Archive == archive.org 16:04:57 :) 16:05:13 i'm guessing in the short term, it's a matter of just updating that one line 16:05:13 (hi) 16:05:21 with one "item" per file, even separating the binaries and signatures into separate items 16:05:25 https://archive.org/details/torbrowser-install-win64-10.0.14_zh-TW.exe 16:05:29 https://archive.org/details/torbrowser-install-win64-10.0.14_zh-TW.exe.asc 16:05:29 first time I look on how the uploads are in ia 16:05:59 I see, they could be a single item with with different files in them, isn't it? 16:06:19 It's possible to have multiple files per item, so there could be one item for torbrowser-10.0.14, and then all the localizations and OSes could be inside it. 16:06:20 will it make sense to have one item per release? or per release-arch-lang? 16:06:25 https://blog.archive.org/2011/03/31/how-archive-org-items-are-structured/ 16:06:44 Or yeah, could be one item per arch, or at least put the signatures in the same item as the executable. 16:06:48 I think it makes sense, I will need to explore the ia API to see how to do that, but I think is a good point 16:07:04 I think anarcat has a lot of experience with archive.org, you might get a second opinion from them. 16:07:18 nice, I'll ask him 16:07:32 One thing is I don't know if it impacts usability, i.e. 16:07:51 At https://archive.org/details/torbrowser-install-win64-10.0.14_zh-TW.exe there is a "WINDOWS EXECUTABLE" link off to the side, 16:07:53 I guess we can't modify the old ones for now, as we might not want to break compatibility with links that are already distributed 16:08:15 and if you get that far, clicking that link will get you what you want. Whereas if it's dozens of files maybe not. 16:08:33 But if gettor emails are linking directly to a single file, it wouldn't be a problem. 16:08:41 Oh yeah, I would not modify anything already uploaded. 16:08:42 if I understand correctly what you get from gettor is the actual download link 16:08:48 you should not se the page of the item 16:09:08 A direct file link would look like e.g. https://archive.org/download/torbrowser-10.0.14/torbrowser-install-win64-10.0.14_zh-TW.exe 16:09:18 I.e. with "/download/" instead of "/details/" 16:09:47 sounds pretty straight forward, yes 16:09:48 That's all I have to say. Sorry, I know it's random, I just think about it every time I see the page. 16:09:54 thanks for noticing it 16:10:20 yeah thanks 16:10:33 I guess at some point in the future we could remove or reorganize old content, so the page looks cleaner, but we'll need to wait a bit of time 16:11:07 nah I would just leave the old stuff forever 16:11:25 :) 16:11:25 cool URLs don't change 16:12:15 :D 16:12:21 anything else for discussion today? 16:12:23 * meskio goes to write a note about that in the issue 16:13:51 okay any needs help with? 16:14:14 arlolra: sorry i didn't get around to looking at the broker work this week 16:14:22 i still have it on my list 16:14:29 and am planning to look at it this week 16:14:41 that's ok, dcf provided some valuable feedback and i reworked it a bit 16:14:54 no rush 16:15:04 nice 16:15:23 arlolra: yeah it looks like I was wrong in initially suggesting unifying "legacy" with the version numbers as I suggested to cohosh to do 16:15:35 Because the "legacy" really does only belong to the HTTP layer 16:15:52 SO the two-level decision tree really is appropriate 16:16:09 if it's going to cause problems in the future, we could change this 16:16:10 that's somewhat easier to see when things are separated as in the patch 16:16:18 yeah 16:16:28 i've followed up with that refactor 16:16:30 as you'll see 16:16:41 one layer of the tree happens at the HTTP handler, the lower layer happens in the factored-out common code 16:16:48 hey 16:17:50 https://gitlab.torproject.org/tpo/anti-censorship/pluggable-transports/snowflake/-/merge_requests/39/diffs?commit_id=0ef22502808f473bc9f368166c1c3c723fa71d09 16:19:19 good assessment, thanks 16:21:02 anything else for today? 16:21:34 How was the RightCons' session? 16:22:04 i didn't attend RightsCon this year 16:22:13 i think ggus might have 16:23:35 ah ok I think I read about Snowflake somewhere on the program. 16:24:45 I didn't hear about that. At https://blog.torproject.org/events/month I see "RightsCon Tor Panel: Building strategies to diversify funds" 16:25:10 yeah me neither, i'm looking up the program now 16:26:19 hm i don't see it 16:27:17 sorry, my bad 16:27:25 oh no worries :) 16:27:38 anyhow I read the survey about Snowflake, great work!! 16:27:45 thanks anadahz! 16:28:22 okay i think i'll end the meeting here 16:28:35 #endmeeting