Submit

Back to the channels

#primate

  1. blue Hi phaleth.
  2. phaleth hi blue
  3. phaleth clanker still sleeping I assume
  4. blue actually not
  5. blue but I just woke up
  6. blue and I have a dayjob :P
  7. phaleth that's cool, means clanker does not starve you out of all of your money while it's digging deep rabbit holes
  8. blue true
  9. phaleth https://blog.qualys.com/vulnerabilities-threat-research/2026/11/17/how-to-find-mitigate-and-patch-the-critical-java4shell-flaw-cve-2026-114938-now
  10. phaleth looks like a lot of server are affected
  11. phaleth $ grep -m1 '^flags' /proc/cpuinfo | tr ' ' '\n' | grep -E '^(avx512f|avx512bw|avx512vbmi|bmi2)$'
  12. phaleth bmi2
  13. phaleth unless of course those who stay true to old Java :)
  14. phaleth wtf is 8u513-perf
  15. blue wow
  16. blue I just forwarded this at work, becuase we use java
  17. blue dreamreal: ^ this is big
  18. blue this has been a terrible week in IT security...
  19. phaleth yeah, at work we run a bunch of Java servers too, but we will have to wait for the software vendor to actually confirm, they are prolly testing those Java updates right now
  20. blue terrble
  21. phaleth right, lot of annoying crap to update
  22. phaleth all these stupid Java services still running on VMs
  23. blue jesus
  24. blue phaleth, I got your invite to provision-research, thanks
  25. phaleth yw
  26. phaleth looks like I cannot create note against primate project knowing the issue number without having it cloned down
  27. phaleth unknown task command "--project"
  28. phaleth wasn't that a thing?
  29. phaleth so anyway, just merge and don't worry about palemoon, adding the --project flag is prolly too much of a hassle
  30. phaleth https://i.imageupload.app/567936d00accb5f362ec.png
  31. phaleth https://repopack.com/project/repopack/tasks/925
  32. blue I'll fix 925, phaleth
  33. blue https://repopack.com/project/primate/tasks/427
  34. blue phaleth, was that also meant to be backported onto 0.45?
  35. blue because that is what runs primate.run
  36. phaleth dunno, prolly too much of a hassle
  37. phaleth 0.46 is being worked on anyway right?
  38. phaleth or is 0.45 the last primate
  39. blue phaleth: 925 is done, run `rp update` and you get that flag on task notes
  40. blue (assuming you're on the unstable channel)
  41. blue 0.45 is the current stable, 0.46 is the next version
  42. dreamreal morning
  43. dreamreal and yes, I'm aware of the base64 exploits, it's hit php, java, ssh, will hit others
  44. blue This week is WILD in IT, dreamreal
  45. blue we're seeing full-scale meltdown
  46. dreamreal It's the dependency thing. Base64 processing had a hole, it got found, it got propagated like MAAAAD.
  47. blue and everyone is using the same nonsense?
  48. dreamreal it's a basic OS-level lib, yes
  49. dreamreal if you optimize base64 decoding in the recommended manner on amd64 you're vulnerable
  50. blue holy cow
  51. blue apparently my NIH-approach pays massive dividends, dreamreal
  52. * dreamreal shrugs. It's one of the nice things about rust's ecosystem
  53. blue what, NIH?
  54. phaleth reinventing it all
  55. blue lol
  56. dreamreal not necessarily NIH, but rust's policy of not depending on shared libs as a default
  57. dreamreal like, if you want image support, using image-rs doesn't use libjpeg.so or libjpeg.a
  58. dreamreal so vulnerabilities there don't matter
  59. blue they matter, just inner-rust
  60. dreamreal Right, but that means the common attack surfaces are harder to use
  61. dreamreal so this is a base64 common lib on amd64 arch, and rust... doesn't care, nor should it
  62. dreamreal rust++ here
  63. nevet rust now has karma of -1.
  64. dreamreal java cares because java sucks at binary manipulation (the design was always "use a library for your platform to decode base64" and thus Java has ... er... limited binary types) and *bang* hello
  65. blue that was a LONG dissertation to say that java sucks, dreamreal
  66. blue java--
  67. nevet java now has karma of -1.
  68. dreamreal but it doesn't
  69. blue nevet thinks otherwise
  70. dreamreal Nah, nevet has no opinion about it (and runs on java anyway)
  71. dreamreal it's people who use karma who do that, and people are stupid
  72. blue what is the karma of java?
  73. blue FactoidUpdatedEvent(selector=what)
  74. nevet ok, blue: updated what.
  75. blue ...
  76. phaleth not sure you
  77. phaleth oops
  78. dreamreal karma java
  79. nevet java has karma of -1.
  80. dreamreal Want me to check the logs to see how that was arrived at? :D
  81. phaleth not sure you've noticed this one https://www.phoronix.com/news/Redox-OS-Golang-GIMP from lasts week
  82. blue oooo
  83. dreamreal I did not
  84. phaleth people kept thinking that FreeBSD will be available as desktop OS, but it seems like Rust devs are gonna get something of their own done sooner
  85. blue still waiting for the rust rewrite of gimp, 'rimp'
  86. phaleth they ported gimp to that OS already
  87. blue "Redox OS does not accept contributions generated by LLMs (Large Language Models), sometimes also referred to as “AI”."
  88. dreamreal The way I figure it, some languages suck and truly; there're multiple reasons that this can be true. Scala, for example, as *a language* is pretty damn nice. Scala as a languae a person would use, though... sucks and I refuse to bother, because of the ecosystem.
  89. blue braindead...
  90. phaleth they have some sort of rust libc or whatever nonsense
  91. dreamreal PHP, on the other hand, as a language BLOWS GOATS FROM HERE TO SUSSEX. But the ecosystem is actually pretty nice. Still sucks.
  92. dreamreal sentiment
  93. blue I find that the term "whatever nonsense" serves well to describe rust in general
  94. nevet Sentiment +1/10 | Intensity mod | Themes: base64 exploit, Rust vs Java/PHP, karma bot, Redox OS | Summary: Friendly banter with playful language-bashing; mild scorn at Redox LLM policy
  95. blue how do I make "whatever nonsense" point to "rust"?
  96. blue hm
  97. dreamreal whatever nonsense is <see>rust
  98. dreamreal FactoidUpdatedEvent(selector=whatever nonsense)
  99. nevet ok, dreamreal: updated whatever nonsense.
  100. dreamreal whatever nonsense
  101. nevet whatever nonsense is <see>rust.
  102. dreamreal Hrmm
  103. blue O_O
  104. dreamreal whatever nonsense.forget
  105. nevet ok, forgot whatever nonsense.
  106. dreamreal sorry, still thinking javabot
  107. blue no worries
  108. dreamreal whatever nonsense.see = rust
  109. dreamreal FactoidUpdatedEvent(selector=whatever nonsense.see)
  110. nevet ok, dreamreal: updated whatever nonsense.see.
  111. dreamreal whatever nonsense.see
  112. dreamreal whatever nonsense
  113. blue I think you broke nevet :(
  114. phaleth well the thing is if they succeeded and showcase they can do just about whatever then things will change, which is annoying of course, but who can be bothered patching all the 50 years old shit
  115. dreamreal whatever nonsense.see.forget
  116. nevet ok, forgot whatever nonsense.see.
  117. dreamreal nah, I just need to remember the factoid ontology
  118. phaleth I mean it's not all 50 years old, but it's based on that
  119. blue even if it were 500 years old, rust isn't the answer
  120. blue in fact, rust isn't the answer to anything
  121. phaleth like the biggest problems there are inherently in all the infrastructure, for example null pointer terminated strings
  122. dreamreal Ah!
  123. dreamreal rust
  124. nevet Rust is a systems programming language with an emphasis on safety without a garbage collector, using a compile-time ownership model to prevent resource leaks. Now being used in the Linux Kernel and many other such projects. URL: https://rust-lang.org/
  125. dreamreal whatever nonsense.alias=rust
  126. dreamreal FactoidUpdatedEvent(selector=whatever nonsense.alias)
  127. nevet ok, dreamreal: updated whatever nonsense.alias.
  128. dreamreal damn it
  129. blue lol
  130. dreamreal whatever nonsense.alias.forget
  131. nevet ok, forgot whatever nonsense.alias.
  132. blue whatever nonsense
  133. blue kaputt :(
  134. dreamreal give me a minute, I don't remember the ontology and the docs were written later
  135. blue help
  136. blue help ontology
  137. blue docs
  138. blue hm
  139. dreamreal Give me a cotton-pickin minute
  140. dreamreal online help for streampack would bury you
  141. blue I'm grasping for straws here
  142. dreamreal It's not a difficult problem to solve, the diataxis was written LONG after the factoid stuff was done
  143. dreamreal I just need to document the ontology
  144. blue *nod*
  145. dreamreal And you're picking a great day for it anyway, PLUS that particular factoid set is likely to get elided by the admin, since he doesn't care for actively negative/spurious factoids :D
  146. dreamreal those get left in javabot
  147. dreamreal nevet is actively constructionist, negativity is fine when it's absolutely deserved but in general humor is mildly beneficial and the same goes for content
  148. blue I've never seen anyone justify longer his "I'm the admin" situation :P
  149. blue you're the tool's owner, you can do whatever you want
  150. blue I do not need a reason for it, nor do I deserve one
  151. blue I just don't like rust
  152. dreamreal I don't care if you don't like rust, dislike it all you like
  153. dreamreal but I think in ontologies, so "I don't like X" is rarely justification for me
  154. blue you don't like scala
  155. blue I don't see the diff
  156. blue also what is diataxis?
  157. dreamreal whatever nonsense
  158. nevet Rust is a systems programming language with an emphasis on safety without a garbage collector, using a compile-time ownership model to prevent resource leaks. Now being used in the Linux Kernel and many other such projects. URL: https://rust-lang.org/
  159. blue very good
  160. blue you're done a good deed today, dreamreal
  161. dreamreal I found bugs, yes, I'm trying to figure out why that failed before
  162. blue it failed because you have the best tester on the planet, that's why
  163. dreamreal no, it failed because the dumbass coder didn't anticipate whitespace there
  164. blue yes, that too
  165. dreamreal that was the bug, is being fixed along with the docs
  166. blue but mostly the tester being superb
  167. dreamreal yes, he's the best, the biggest, the brightest, the humblest
  168. blue yes
  169. dreamreal Nothing like a good wednesday morning with an additional five spinning plates. But this also lets me deploy a different JVM image, so there's that. :D
  170. blue YES, all thanks to the tester! this nameless tester!
  171. dreamreal nameless tester, probably uses a color as his IRC nick, etc., so nobody knows who he is
  172. dreamreal what a total drip
  173. dreamreal :D
  174. dreamreal I'm moderately pleased that I DID remember the ontology after a bit, defintely annoyed at the diataxis gap, and the bug is even MORE annoying
  175. dreamreal easy to fix, though
  176. blue I think I'll call my company Ontology Diataxis Ltd.
  177. dreamreal Diataxis Ontology makes more sense
  178. blue ok, so Diataxis Ontology Ltd.
  179. dreamreal Although I guess Ontology Diataxis might work
  180. blue ok, so back to one
  181. dreamreal (diataxis for ontologies...)
  182. blue it really doesn't matter, because I don't know what either means
  183. blue I could also call it Limited Ontology Diataxis
  184. dreamreal define ontology
  185. nevet define ontology: ontology (noun): (uncountable, philosophy) The branch of metaphysics that addresses the nature or essential characteristics of being and of things that exist; the study of being qua being. (Wiktionary via FreeDictionaryAPI.com, CC BY-SA 4.0: https://en.wiktionary.org/wiki/ontology)
  186. dreamreal define diataxis
  187. blue crickets. diataxis means crickets
  188. * blue learnt
  189. dreamreal diataxis
  190. nevet Diataxis is a way of thinking about and doing documentation. It's a quadrant of tutorials/explanation, how-to vs reference. What to write, how, and how to organize it. Should guide all documentation. Doesn't. URL: https://diataxis.fr/ Tag: documentation
  191. dreamreal It's a word the free dictionary may not know about
  192. blue I read that ontology definition, and I still don't know what it means
  193. blue as for diataxis, something like meta-documentation?
  194. dreamreal an ontology is a schema for knowledge
  195. blue so ontology is to knowledge what json schema is to... json?
  196. dreamreal so a factoid has an ontology: a text that is human-readable, has associated urls, maven coordinates, types, references, etc, etc
  197. dreamreal yes
  198. blue why not just call it... data structure?
  199. dreamreal a json schema is an ontology for json documents to which it refers
  200. dreamreal because that's not what it is
  201. dreamreal but you could
  202. dreamreal ask what's the difference between an ontology and a schema?
  203. nevet A schema defines the *structure* of data (fields, types, constraints), while an ontology defines the *meaning*: concepts, their relationships, and rules for reasoning about a domain. A schema says "this field is a string called `author`"; an ontology says "an author is a person who created a work," which lets systems infer and share knowledge across datasets.
  204. blue so ontology is... definitiology?
  205. dreamreal so when a factoid has a .see attribute, the ontology says "this is what this means" and it's not just a type
  206. dreamreal streampack is updating now, factoid fix + documentation coming soon
  207. blue yay
  208. phaleth https://social.anoxinon.de/@codebergstatus/117636595327014447
  209. dreamreal Sorry, delay on nevet redeploy is to cull the CVE vulnerability, exploring options there
  210. dreamreal chose to disable the base64 AVX option, because teh hardened images are HUGE
  211. phaleth one day Java might not be such a memory hog https://openjdk.org/projects/valhalla/value-objects
  212. dreamreal It's already vastly better now with compressed headers
  213. dreamreal and as always there're tradeoffs involved
  214. phaleth cool
  215. blue phaleth: codeberg is a joke. this was identified foremost on the zig article
  216. phaleth yeah, they should have just quickly registered the domain somewhere else instead of crying about americans
  217. blue ya
  218. nevet [ByteCode.News] Geertjan Wielenga: An Introduction to the Java Visual Recognition API - https://bytecode.news/posts/2026/11/geertjan-wielenga-an-introduction-to-the-java-visual-recognition-api
  219. blue dreamreal: congrats, ANOTHER contribution!
  220. dreamreal well, I wrote that
  221. blue oh
  222. blue I thought Mr Wielenga wrote it
  223. blue I should open links...
  224. blue oh, the article's not done yet, dreamreal?
  225. blue it's three paragraphs
  226. dreamreal It's not long, no
  227. dreamreal the original source is pretty well done
  228. blue oh ok... I'm accustomed to more... as phaleth puts it, "walls of text" from you
  229. dreamreal One of the problems BCN has is the walls of text
  230. dreamreal I've actually been trying to make a point out of writing more succinctly
  231. dreamreal and one thing I DO NOT want BCN to do is "here's this thing over here, but you don't need to read it because it's all here"
  232. blue ya
  233. blue visrec is a good name. the spi provider thing is actually one of the things I love most about java
  234. dreamreal even the walls of text are written such that they add to the original content rather than just summarize or replace
  235. dreamreal at least, by intent
  236. dreamreal yeah, SPI is powerful
  237. blue I once had a problem with imageio not parsing a certain jpeg, due to malformed exif data
  238. dreamreal yeah... but imageio sucks
  239. blue I dropped into some twelvemonkeys jars, didn't change a line of code, voila, it worked
  240. dreamreal I live in imageio every day :(
  241. dreamreal twelvemonkeys sucks too
  242. blue it felt like magic. TRUE magic.
  243. dreamreal for the record :D
  244. blue oh come on you party pooper
  245. dreamreal hey you do this stuff ALL THE TIME
  246. dreamreal and in this case, I literally have to work around imageio's limitations every day
  247. blue me? we're talking the same world-class tester who found out nevet can't parse spaces in factoids?
  248. dreamreal and yes, we use twelvemonkeys and a host of others
  249. blue I never party-poop!
  250. dreamreal I have yet to see you pick up a balloon without trying to pop it :D
  251. blue hey, it's for SCIENCE!
  252. blue I'm a contrarian, but it's *usually* for a good cause
  253. dreamreal Sure. But that's why I find imageio utterly insufficient, too :D
  254. dreamreal malformed exif data in the base libs are only the very beginning of the iceberg
  255. blue yes yes, the java native impl or jdk or whatever it is is horse manure. HOWEVER. the spi architecture means that doesn't matter one iota
  256. dreamreal sure
  257. dreamreal (note: $dayjob involves a lot of image analysis)
  258. blue yes, I know (though chat mightn't)
  259. dreamreal Thus why $dayjob and not any more detail than that
  260. blue yes yes, we know that as well, you can't talk about it
  261. dreamreal Well, literally speaking, I CAN, I just won't :D
  262. blue at least it's better than fisa courts, where you can't even able to talk about the fact you can't talk about it
  263. dreamreal I could actually give out details but why
  264. blue true

Search posts