IRC channel logs
2025-07-28.log
back to list of logs
<damo22>why do you need filesystems to be compatible with each other? <damo22>you cant mount a filesystem of a different type with a different translator <youpi>sneek: later tell azert it's essentially the libdiskfs API that represents the vfs <sneek>azert, youpi says: it's essentially the libdiskfs API that represents the vfs <azert>youpi: the question is if the journaling shouldn’t record operations at the inodes level. Milos doesn’t think that necessary, but I’m thinking what could be the drawback of his approach? <youpi>I don't know journaling enough to have an opinion <youpi>my general opinion would however be to be humble and look at how things are done elsewhere, and really check why I'd want to do things differently :) <azert>indeed, there should be some damn good reason <azert>but my understanding is that elsewhere kernel modules can directly access the node api, while that is not currently the case in the Hurd <youpi>that's not a problem that cannot be fixed :) <azert>I don’t even think he needs to fix specifically that for his use case <azert>youpi: if a part of the Linux kernel is licensed gplv2+, does this make it more suitable to be redistributed as part of the Hurd? <youpi>sneek: later tell azert yes that's easier to integrate, licence-wise