02:01:07 ek: same result 02:02:14 voy4g3r2: And if you look in /usr/local/poudriere/ports/HEAD in the jail, the ports tree is up-to-date? 02:02:34 I suppose it might help if you include the Poudriere error(s) and build log. 02:02:46 I'm completely in the dark as to what's going on. 02:07:01 me too 02:07:09 thanks for the help, i may just have to wait for updates to the ports 02:07:19 i got a version of llama.cpp i can use.. so it is not mission critical 02:08:17 haha the .failed file 02:08:29 lang/python311 python311-3.11.15_2 checksum checksum 0 02:09:20 i'm laughing at this point because it is like.. wtf and where to go from here, basically a nervous reaction 03:08:25 voy4g3r2: I don't suppose you're going to provide these logs? 03:08:52 i could do that 03:09:01 got distracted 03:09:11 all the logs or just the package? 03:09:53 Well, the Poudriere build log for the package, for sure. 03:09:58 That would be a start. 03:10:21 I'm just trying to figure out where this error is happening. 03:10:40 Perhaps it's just a Poudriere configuration issue? 03:16:48 that may be possible, i have built packages "before" 03:16:55 in early may without this issue and python has been updated since then 03:17:01 found the log and getting it pasted somewhere, one more minute 03:17:33 https://pastebin.com/iZVgJie5 03:17:42 this was the ONLY file in the error log directory 03:18:24 the directory was /usr/local/poudriere/data/logs/bulk/150Ramd64-HEAD/2026-05-28_19h00m15s/logs/errors 03:18:34 which just does a sym link.. one directory up from errors 03:43:33 voy4g3r2: ===> Giving up on fetching files: python/ceac1efc66516ac387eef2c9a0ce671895b44f03.patch 03:44:31 This is concerning. Can you past the distinfo file for lang/python311? 03:45:17 s/past/paste/ 03:46:36 Odd that it's only the patches. 03:47:14 heston76: Yeah. 03:47:28 I'm curious to see what their distinfo contains. 03:52:18 sure, i can do that 03:54:27 https://pastebin.com/84HHaNYP 03:54:37 i did check against the github links, in the freshports.org entry 03:54:42 i could NOT see any difference.. 03:55:23 https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=295595 03:55:36 this is why i have been hyper focused on the checksum and the files not downloading.. but was going in circles 04:01:30 which then brings me back to the point.. i assume python311 is a pretty common thing to get rebuilt by smarter people than me but yet not seeing anything 04:02:13 its late, i will tackle more of this tomorrow.. it is eluding me 04:06:00 voy4g3r2: Have you tried setting the fetch reget to higher than 1 just to see what happens? 04:06:17 I haven't seen this issue, but it appears to have helped the user in the PR you posted. 04:07:09 voy4g3r2: maybe a "FETCH_REGET=4" or something in /etc/make.conf? 04:14:19 i can give that a try 04:16:42 running 04:16:50 and this time i am bouncing.. i shall see what happens by morning 04:17:03 thank you for the patience and guidance, it is greatly appreciated 04:59:12 FETCH_REGRET amirite 05:28:51 hi 07:21:56 < Ltning> Hecate: What application is this? // It's a Haskell web app, https://github.com/flora-pm/flora-server 07:22:12 362u/9 07:22:15 (woops) 11:39:48 haha 11:39:51 ek: That was it 11:47:13 thank you! 12:21:30 Is there a way to run the third stage of "freebsd-update install" at a later stage, despite /var/db/freebsd-update being mangled by freebsd-update cron in the meantime? 13:39:36 Hello, I'm building jail on the fly (CI/CD pipeline) and I was wondering about IP assignation: I'm looking for a solution for a DHCP-like solution for non-VNET jails, any idea? Basically I'm looking for a tool (Consul ?) in which a can assign a pool of IPs and having way to atomically fetch/release an ip from this pool 15:13:39 mage: may i ask why it has to be a non-vnet jail? it is my understanding you can not do DHCP work (from my experience) in a non-vnet jail. 15:14:23 host: setup dynamic pair interfaces in a bridge jail: have the jail get an ip, attached to the dynamically created pair interface. Then run your tests, close down the jail.. host releases the pair interface 15:58:55 voy4g3r2: so you didn't move the old files away as I had suggested yesterday and make fetch didn't validate that the already downloaded files are indeed the wanted files, before it skipped to download the files 17:01:01 Hey if anyone is interested in XMPP dev, there will be a dev event in London in Nov, if you are interested let me know: https://wiki.xmpp.org/web/Sprints/2026-11_London 17:01:29 as for how this relates to BSD, we really need more BSD adoption, so if someone is interested in helping with running BSD based XMPP services, please also consider coming. 18:23:56 mage: and why would dhcp not work? 20:54:18 nimaje: it was not that. I even created a brand new tree. Only when the FETCH_REGET=4 was added to the make.conf that it worked. 20:55:05 nimaje: so, made new tree, tried again, nothing. re-updated 2 trees nothing. Added FETCH_REGET=4 to make.conf and then it worked. 20:56:26 Could be something to do with the way your ISP does routing or something. 20:57:48 that is out of my element and trying to prove that 20:57:59 because it is onlY 1 package.. if multiple one, i would push more on ISP 20:58:19 i ended up building 115 packages in total :) 20:59:59 which all starts because my CPU does not support AVX, AVX2, FMA and F16C instructions that llama.cpp is built for by default. 21:00:13 and thanks again ek, kind of weird little nuance buried in all those notes 21:01:21 Yeah. Definitely strange. Especially considering it's only an issue with the patch files. 21:01:44 And it's pulling from github so it's not like some random hosting site. 21:01:54 You're welcome! 21:51:02 Oh, on the topic of XMPP: https://ekollof.github.io/xepher/ <-- Just fixed that today for FreeBSD and OpenBSD. Using it right now in weechat. 21:52:05 There is a FreeBSD MUC at freebsd⊙mxf for people that are interested