Time Nick Message 17:27 Krock Hm. It's been a while since the last necessary bump, thus I might be misremembering an earlier discussion on the protocol bump policy. I thought it would only be done for the release candidate build - but turns out there's no such restrictions in the wiki. Sorry for the fuss. 17:36 sfan5 we should write down that "During development if any feature that requires a protocol bump is added, the should be done at the same time." then 18:08 sfan5 rubenwardy: please build from here https://github.com/luanti-org/luanti/tree/ci 18:22 rubenwardy build failure https://github.com/luanti-org/luanti/actions/runs/24415141557/job/71322171475#step:5:395 18:29 sfan5 oops, fixed 19:09 sfan5 windows builds verified to work 19:22 Krock sfan5: I might be rather late to the party - do you need help with something? Changelog? 19:23 sfan5 hmm i don't think so. I have a commit preprared for the changelog locally. 19:23 sfan5 prepared* 19:24 Krock okay 19:24 sfan5 you could verify that the bugs are actually fixed by the backported patch, because I haven't 19:25 Krock sure. running a test mod... 19:31 Krock https://github.com/luanti-org/luanti/commit/5fe84437 I wish this had a better commit description what made this necessary. uh anyway 19:32 Krock the ItemVisuals and Sanitize commits work 19:33 sfan5 it was a prerequisite of the other commit but if you want to test then core.get_dir_list(core.get_worldpath() .. "/worldmods") should work, and wouldn't work before that 19:33 sfan5 (and of course worldmods should not be writable) 19:34 sfan5 rubenwardy: eta on android? or should I not wait? 19:34 rubenwardy I use CI to produce builds 19:35 sfan5 I know that but it finished a while ago :) 19:36 rubenwardy I usually use the tag but I could use the commit hash 19:36 sfan5 yeah should work 19:36 Krock sfan5: sorry to disappoint but worldmods is writable by worldedit_commands 19:36 Krock > local x, err = io.open(core.get_w 19:36 Krock orldpath() .. "/worldmods/test.a", "w"); 19:37 Krock test.a was created in the directory 19:37 sfan5 let me check 19:38 Krock tbf I'm using symlinks and Wine, which might cause funny false-positibes 19:39 Krock okay. It's blocked in current master with no symlinks. Now testing what makes the difference 19:39 sfan5 https://x0.at/HHxK.png wfm on master for me at least 19:39 sfan5 if symlinks somehow break it that sounds like its own bug 19:41 Krock does still work on Wine with no symlinks in the world path 19:43 sfan5 https://x0.at/Qn8V.png huh yeah this is on windows 19:44 Krock probably a / vs \ thing 19:45 sfan5 ⭢ io.open(core.get_worldpath() .. "\\worldmods\\something2", "w") 19:45 sfan5 ⭠ file (0x01c6aa51f490) 19:45 Krock actually no... PathStartsWith contains code to handle slashes 19:45 sfan5 could still be 19:48 sfan5 time for debugging 19:54 sfan5 ok so this isn't present in master 19:57 sfan5 I bet the check is working correctly and it sets *write_allowed but still returns true 19:57 sfan5 in master there's https://github.com/luanti-org/luanti/commit/45083abf5eb5280fdc76eed80c78cff1a2a07f79 which fixed this loophole 20:03 Krock yes. that makes sense. 20:04 rubenwardy ok built 20:05 sfan5 possible fix pushed to 'ci' branch again, will check after build is done 20:05 sfan5 (sorry ruben) 20:05 sfan5 this situation is my fault for mixing two changes in one commit 20:05 * Krock sneaks into bed almost silently. 20:15 sfan5 success https://x0.at/Ukvs.png 20:15 sfan5 rubenwardy: you can do it again 20:55 rubenwardy done 20:58 sfan5 I guess we're read to go then? 21:01 sfan5 ready* 21:17 sfan5 I take that as a yes 21:27 sfan5 can you also upload the android stuff to the gh release? 21:43 sfan5 forum post done, advisories done 21:44 sfan5 will wait with updating the website until we have at least the android builds (tomorrow?) 22:39 rubenwardy done