ByteCode.News
RSS
SubmitLog in

Community

Chat Logs

Monday, October 5, 2026

  1. deebo"teaching sand to think was a mistake"
  2. deebohmm, does the multilaunch work across projects, could be useful then, otherwise meh
  3. deebonow if idea could realize i have all my datasources in datagrip and use those without copypasting them into each project separately, that would be great
  4. BombeYeah, that would be pretty neat.
  5. BombeHeh, turn out it’s actually possible and we’re just idiots: https://stackoverflow.com/questions/41934918/sharing-databases-between-projects-within-intellij
  6. BombeHmm, no, I misread, you actually have to copy & paste the sources around: https://www.jetbrains.com/help/datagrip/managing-data-sources.html#copy_data_sources_to_clipboard :(
  7. javabotBombe's title: "Data sources | DataGrip Documentation"
  8. BombeDisappointing.
  9. NeXeNwhat were you trying to solve?
  10. ParaI wonder if datasource configs live under .idea nowadays.
  11. ParaJetbrains afaik is trying to get people to use those + commit them, so the copypaste at least would be formal "this file right 'ere"
  12. NeXeNgo look. i don't have intellij installed
  13. NeXeNbut yeah they could make it a tad bit easier
  14. dreamrealNeXeN: you're a !idea person!?!? What do you use?
  15. NeXeNvsclone*
  16. NeXeNsometimes vim too
  17. dreamreal*nod*
  18. NeXeNi don't use enough features of intellij to make it worth my while. it's a little sluggish compared to some lighter vsclones out there
  19. NeXeNi have my uma vram jacked up to max, so i only have 16g of system memory and that might be part of it
  20. dreamrealI'm still in the "you'll have to rip IDEA from my cold, dead, rotted hands" phase
  21. NeXeNwhen they did the hamburger menu thing, that was the last straw i fell off
  22. NeXeNthe main feature i ever cared about was the language server. the redhat one on vsclones sucks bonky dolls but it works. so i don't really have anything to compel me to stay
  23. dreamrealI still code too much personally to leave
  24. dreamrealI know it, it's comfy and warm, it helps me do the thing
  25. NeXeNyeah if you can use the features it's got some awesome ones
  26. dreamrealand bloody hell, I was using it when beg was first releasing it on efnet #java as early access, to leave now feels weird
  27. dreamrealI was using it back when it was "oh look, it's 11:30, ANOTHER release, this last one survived almost an hour. Oh shoot, it's 11:35: another version."
  28. NeXeNi think i have this thing against IDEs.......when i was a guppy i took a course "fuzzy math in c++". and this was like 97 or 98. and they used vc++ but i couldn't run it as i used linux exclusively. so i did my homework in vi (or maybe pico) and went to school early to try and get it to compile
  29. dreamrealWe kept on asking him why he didn't just call it together/j++ but he didn't appreciate that much :D
  30. NeXeNbut the course was mostly "teach you how to use the ide" not "teach you how to build things"
  31. * NeXeN shrugs
  32. dreamrealyeah
  33. dreamrealI'm surprised anyone gets out of college appreciating Java, because of the way it's taught
  34. NeXeNi never got taught java so i really appreciate java
  35. NeXeNback in those days it was drink the koolaid
  36. dreamrealMy son took a java course as part of his degree, I was appalled. I helped him with a problem he was having, THEN showed him how I would have done it, and he was... rather surprised
  37. dreamreal"why don't we do it like that in school?"
  38. NeXeNyou know the way it was taught years ago was bad for performance because usually they tell you to stay away from static members and the like. use too many objects when few would do. gc thrashing......
  39. NeXeNbut now we have value objects and records and other things decades late :P
  40. dreamrealOh, his course used eclipse as a build tool and avoided the classpath
  41. dreamreal"copy this stuff into your source" and then they'd have to constantly have remediation for plagiarism
  42. NeXeNeclipse is aiight. i think i was big netbeans in the eclipse days
  43. dreamreal"You copied these four classes!" ... yes, just like your idiot syllabus demanded
  44. dreamrealeclipse is fine for what it is: if you think like beck, it's awesome and THE WAY. But as a build tool... nah
  45. NeXeNyeah i stick with maven
  46. dreamrealAny build tool > any IDE as a build tool
  47. NeXeNyears ago we didn't have it so the IDE thing i guess develped parallel to real build tools
  48. dreamrealWell, people used the IDE as a build tool, it's a leftover from early IDEs
  49. dreamrealbut then people realized an IDE was a tool and not a standard
  50. dreamrealyou don't say "we use a craftsman hammer on this house," you use a hammer, the specific one doesn't matter as long as it drives a nail
  51. dreamrealarticle Java taught vs Java practiced
  52. nevetIdea session started: "Java taught vs Java practiced". Use 'content <text>' to add body paragraphs, 'includeai' to enable AI summary/tags, 'done' to save, or 'cancel' to discard.
  53. dreamreallogs 6h
  54. nevetAdded 43 log messages (last 1h) as content block #1.
  55. dreamrealincludeai
  56. nevetAI summary enabled for "Java taught vs Java practiced". On 'done', a generated summary and suggested tags will be appended to the draft.
  57. dreamrealdone
  58. nevetIdea saved as draft: "Java taught vs Java practiced" (1 content block). AI summary appended for admin review.
  59. dreamrealHuh, that might actually be useful
  60. dreamrealnevet++
  61. dreamrealooo and a bug
  62. dreamrealthree bugs, actually
  63. dreamreal(So what just happened, if you're interested: I had nevet capture an article idea from the channel, with a title just on a lark, then I told it to use 6 hours of logs as part of the content. You can contribute blocks specifically AND/OR the logs; I didn't bother with the blocks, just the logs. The "!includeai" tells it to use AI to scan the content for themes and content; it actually wrote a pretty
  64. dreamrealdecent "article" but the idea is to capture concepts for human editors.)
  65. dreamrealIt doesn't try to be especially original, just identify actual *conversation* from the logs (so other people might have their own conversation, but it'll try to summarize the main one.) In this case, it actually looks a lot like an observer writing about what nexen and I were saying.
  66. NeXeNyeah i have built a similar thing for a customer who needed to have some chat bots that didn't look like chat bots to get some prelim info out of people contacting their support
  67. NeXeNit would take the rolling log of what they had discussed and summarize for the agent so they can get quick, with clickable links to the various things said. most people don't figure out it's a bot for a good while
  68. dreamrealnevet's going to part and rejoin in a few minutes, the backend takes some time to build/test/deploy (mostly the test cycle)
  69. dreamrealI keep thinking nevet would be good for #java unmuted but then it'd compete with javabot
  70. NeXeNuse different triggers
  71. dreamrealit has a different signal character (! vs ~) and I VASTLY prefer ~ but... again, javabot
  72. NeXeNyou can do @, which is quite conventional for some people
  73. dreamrealwell, there are also some automatic responses in there, so muted is better, AND I really don't want them to compete per se
  74. dreamrealfor IRC they look like very similar applications (an infobot) but nevet's not really an infobot, it HAS an infobot
  75. dreamrealMost of the changes over the past few weeks have had very little to do with the infobot facet (although today's deployment does!)
  76. dreamrealmostly information management in the irc information generation, because some boundaries changed recently and the tests didn't exercise the boundaries well
  77. dreamreal(!ideas threw an exception in the orchestration, and this exposed another issue in the logs that was trivially easy to fix in place)
  78. dreamrealhttps://bytecode.news/posts/2026/10/why-dont-more-developers-use-the-platform
  79. dreamrealhttps://bytecode.news/posts/2026/10/four-horsemen-or-four-decisions
  80. dreamrealbuilding nevet takes a while now because I removed parallelism from the build - it has a lot of modules and docker didn't really do well with 20 modules all creating postgres testcontainers :/
  81. dreamreal... all at the same time. The nonparallel build is pretty solid.
  82. dreamreal(nevet uses postgres-specific search features.)
  83. NeXeNoh yeah?
  84. NeXeNi love tsvector and tsquery, i think it's marvellous. i've even gone past gin and gist indexes and tried out tin, which is nice
  85. NeXeNeasily highlight words in the surrounding text and configure how much you get and how you can easily keep track of docs
  86. NeXeNi implemented a system over more than a decade that finally we were able to get all of it indexed and such. the users went nutz, like dang i can just search for it and there it is......they had been using google to look at the public facing docs but often missed contextual policies that aren't on the public site
  87. dreamrealYeah, the alternative was lucene or solr and, um, no
  88. dreamrealThey'd work, but be much slower, plus the indexing scheme would MANDATE solr (lucene wants to store on a filesystem, not a database, with all the synchronization issues that implies; solr has a db-index which is, um, much smarter) and solr ends up duplicating what postgres gives me for no real benefit
  89. ParaLucene/Solr/Compass/ElasticSearch are great as long as they're configured correctly. They're absolutely useless for e.g. Finnish if you run them on default settings, which a lot of places did, and then decided to proclaim that they outright suck.
  90. dreamrealnobody cares about the finns anyway though
  91. ParaHaven't use Postgres for non-English text so can't speak for its capabilities, but with correct lexer/stemmer Lucene is just unfairly good.
  92. dreamreallucene's access mechanism is gross, though
  93. dreamrealyou'd have to use solr for multiuser
  94. ParaYeah, it's quite raw without anything on top of it.
  95. ParaThe aforementioned Compass, basically elasticsearch-before-elasticsearch was the first time I felt that this is actually alright to use.
  96. dreamrealI used to work with Shay Banon and Uri Cohen :D
  97. ParaCompass is/was effectively single-node ElasticSearch. Same creators, even, and I suppose they learned from that enough to completely abandon it in favor of ES.
  98. dreamrealAlso: one of my modules in OpenSymphony was called WebCompass. :D
  99. ParaI was more clever and named my Compass wrapper thing Magellan. I don't even remember why and what the wrapper did at this point.
  100. NeXeNhrmm maybe i should rebuild a system i built years ago and leverage some newer stuff
  101. ParaLearn...from your...mistakes?
  102. dreamreal~osgi
  103. javabotOSGi is a module system for Java. It's used in Eclipse, Glassfish, and a host of other environments. See also ~equinox, ~felix, ~knopflerfish, and ~karaf.
  104. dreamreal~equinox
  105. javabotdreamreal, equinox is Eclipse' implementation of an OSGI container. See http://eclipse.org/equinox/
  106. nevetEquinox Documentation
  107. dreamreal~felix
  108. javabotApache Felix is a community effort to implement the OSGi Service Platform and other interesting OSGi-related technologies under the Apache license. http://felix.apache.org/
  109. dreamreal~knopflerfish
  110. javabotdreamreal, knopflerfish is an OSGI container. See http://www.knopflerfish.org/
  111. nevetKnopflerfish OSGi - open source OSGi service platform. OSGi Release 6
  112. dreamreal~karaf
  113. javabotApache Karaf is a small OSGi based runtime which provides a lightweight container onto which various components and applications can be deployed. see http://karaf.apache.org
  114. cheeserRIP OSGi
  115. dreamrealhttps://bytecode.news/posts/2026/10/updates-for-bytecode-news
  116. javabotdreamreal's title: "Updates for ByteCode.News | ByteCode.News"
  117. dreamrealOSGi is alive and well, just not so much as a developer-facing platform
  118. dreamrealand it can only blame itself for it
  119. cheeserzed's dead, baby.
  120. dreamrealhttps://bytecode.news/posts/2026/10/updates-for-bytecode-news
  121. javabotdreamreal's title: "Updates for ByteCode.News | ByteCode.News"
  122. dreamrealhttps://news.ycombinator.com/item?id=49965413 - I would deeply appreciate upvotes here, I want to see how the new front end handles being hammered
  123. nevetHacker News
  124. dreamrealGah, HN had a problem with it thanks to the canonical link being wrong: fixed
  125. dreamreallots of cool new features for nevet, should be a lot more stable again now, been updating things like maaaad