Wednesday, 2026-04-01

opendevreviewMerged openstack/openstack-manuals master: [www] Set 2026.1 Gazpacho as released  https://review.opendev.org/c/openstack/openstack-manuals/+/98175811:48
cardoeSo I know we use uWSGI in devstack and all the projects have guides about it. But I think we're using it wrong.13:25
cardoeWould a good place to start the conversation about it be in a patch to devstack you think?13:25
cardoeHoping to engage folks that have some experience with it and running loads on it.13:27
mnasiadkacardoe: either patch to devstack or the #openstack-qa channel14:13
-opendevstatus- NOTICE: The opendev.org site is currently experiencing overwhelming load adversely impacting git operations and repository browsing since 12:20 UTC today, mitigation work is in progress14:42
gouthamr^ :/ 14:43
-opendevstatus- NOTICE: Load on the opendev.org Gitea backends is under control again for now, if any Zuul jobs failed with SSL errors or disconnects reaching the service they can be safely rechecked16:13
fricklerdoes anyone know whether we have a docs page with content similar to https://wiki.openstack.org/wiki/GitCommitMessages ? if not, where would be a good location to add it?18:49
gouthamrContributor Guide?18:49
clarkbfrickler: probably in the contributor guide (if it exists and if not where to put it)18:49
gouthamrwe reference that wiki heavily today in onboarding stuff and elsewhere18:51
gouthamrhttps://codesearch.opendev.org/?q=https%3A%2F%2Fwiki.openstack.org%2Fwiki%2FGitCommitMessages&i=nope&literal=nope&files=&excludeFiles=&repos=18:51
gouthamran excerpt of it, in a not-so-opinionated manner is already here: https://docs.openstack.org/contributors/en_GB/common/git.html#commit-messages 18:52
gouthamrbut, to someone starting out new, i still link to that wiki page - lots of examples with nuance broken down18:52
clarkbI didn't realize we published translated versions of that doc18:54
gouthamrhaha, don't know why google gave me that link18:54
clarkbhttps://docs.openstack.org/contributors/common/git.html is the source version I think. I suspect they are largely the same due to en_GB not being all the different to en_US18:54
gouthamryes18:55
gouthamrthe translated versions are linked right on top, helpfully18:55
clarkboh cool TIL18:56
sean-k-mooneyfrickler: i jsut use that wiki as a link when i need to provide it to someone18:57
sean-k-mooneybut that wiki is one of the data souce i capture for my review bot too https://github.com/SeanMooney/openstack-ai-style-guide/blob/master/references/commit-message.md18:57
sean-k-mooneythree are some aprt of it that are now defunt18:58
sean-k-mooneylike APIImpact and DocImpact18:58
sean-k-mooneywell *Impact really since we do nto use any of those any more18:59
gouthamryeah :( i loved the doc bug filing feature with the DocImpact tag18:59
sean-k-mooneywell you shoudl not need a doc bug anymore becasue docs shoudl be included in teh commit19:00
gouthamrAPIImpact was mostly for humans in the API SIG19:00
fricklerah, thx, https://docs.openstack.org/contributors/common/git.html#summary-line is just what I was looking for in this special case19:00
gouthamrsean-k-mooney: true, most of our docs are in-tree.. some aren't, and they've slowly suffered decay19:01
sean-k-mooney" The character limit is 50 characters" is one of the thigns i edxplcitly ignore for the sumary lien19:01
gouthamrhaha, it's a creative exercise man19:01
sean-k-mooneyi shoudl bookmakr that link as it actully got much fo the relevent info 19:02
sean-k-mooneyit might be a better refence for new folks since it a bit shorter19:02
gouthamrwoah, i see anubis in action21:45
fungion the lists?21:46
gouthamryes21:46
fungihttps://review.opendev.org/c/opendev/system-config/+/983061 is a change to add a similar layer to our gitea backends21:47
gouthamri think i asked claude to summarize something and it reached out to lists.openstack.org and came back empty; and i opened the archive just now and saw a brief redirection page21:47
gouthamr++ for the better, i hope we can find a way to allow legit access if its tunable enough21:48
fungiyes, i think the intent was to allow well-behaved crawlers through user agent matching21:49
fungiit's supposed to block requests from crawlers masquerading as browsers21:49
clarkbfwiw I think we are allowing legit access :)21:50
clarkbif you claim to be a browser and can't do browser things then I would consider that not legitimate21:50
gouthamrawesome, let me try something and report back :P 21:50
clarkbfungi: and yes I believe our ruleset should allow common bots through as well as clients like git and curl21:51
fungifor example, our testinfra exercises rely on curl and we didn't change anything to allow them through21:51
fungieven though it has no javascript capability at all21:52
gouthamrokay, sorry false alarm. Claude uses its WebFetch API to hit lists.openstack.org, and it's apparently not allowed to construct URLs: https://platform.claude.com/docs/en/agents-and-tools/tool-use/web-fetch-tool21:57
gouthamr(and it didn't, it actually was able to get multiple threads and do its thing21:58
fungiah good, so we didn't break your workflow21:58
clarkbbasically there are two sets of bots roaming the internet right now. Those that say "I'm a bot hello nice to meet you" and those that says "I'm a human behind a browser honest". For the most part the latter are the problem not the former21:59
clarkbthe bots that properly identify themselves are operated by household names and they seem to try and behave21:59
clarkbthe crawlers spoofing user agents to fetch all the data off the internet do no care21:59
funginow if people would just realize that feeding a 3-sentence prompt into an llm to generate a huge message for someone else to point their llm at and get a 3-sentence summary could save everyone time by just sending the 3 sentence prompt to the ml instead, we'd all be better for it22:00
gouthamrfull disclosure about this "workflow", i started doing a thing with the last TC meeting here because i felt like we were missing some cross project threads on the mailing list, so i have claude run through the list for the past week and summarize things that the TC needs to look at, revisit in addition to the planned agenda.. 22:00
fungihopefully it reminded you about the sudden increase in pink elephants the community is struggling to tame for fun and profit22:01
clarkbI know this isn't the way the modern internet works, but why wouldn'y you run that against your local mbox folder ?22:02
fungimaybe that's only a valid assuption for !=gmail users22:02
clarkbI feel like if people did more of ^ and considered the impact of "just set the bots loose" we'd be in a much happier staet and I wound't spend every working hour figuring out how to sort the bots from the humans22:02
gouthamri could, honestly.. the ML is just one source, meetings.opendev.org is another22:02
clarkbthis idea that you're more efficient is creating a direct consequence on my ability to get anything done22:03
* fungi also has local irc channel logs, fwiw ;)22:03
clarkbwhich is fine when its one person but when everyone decides to do it then I'm wondering why the hell I even bother anymore22:03
fungii don't point an llm at them, but they're very useful for my advanced grep-based workflows22:03
clarkband I'm particularly cranky about it today because I woke up to the openstack release being stuck behind this stuff22:04
gouthamrtrue, and i'm being a bad citizen despite being adjacent to this whole mess22:04
fungiat least in my case, searching local files is far faster than trying to fetch them all from the web every time i need to rerun a query22:05
gouthamryeah agreed, i should turn on the local logging.. i'm parked in a bunch of channels where the relevant chatter happens already22:06
clarkbfungi: one thing that bugs me about weechat but I've never dug into it because it works well enough is that my scrollback in weechat is minimal after a restart despite having longer logs on disk that it could oad22:06
clarkbfungi: so instead I just grep teh files directly and good enough22:06
fungiyeah, i don't know how it goes about persisting buffer contents across restarts, never looked into it but is presumably configurable (the thing has more knobs than nasa mission control)22:08
fungii also restart my weechat process on average once a year, so hasn't bothered me much22:08
clarkbfungi: re gmail you can just ask gemini to summarize it for you there22:15
clarkbagainst the local inbox I mean22:15
clarkbfor a value some value of local to gemini22:15
fungithe only gemini i've ever used was an atari 2600 clone made by coleco22:17

Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!