IRC channel logs

2025-10-15.log

back to list of logs

<Darelelve>Hello everyone! While researching filesystems in the GNU Hurd, I came across Ognyan Kulev's article (https://web.archive.org/web/20060625204302/http://debian.fmi.uni-sofia.bg/~ogi/hurd/ext3fs/ext3fs-msc-thesis/) about the ext3fs translator and the difficulties he encountered. Are these same issues still present in the GNU Hurd, and is this why
<Darelelve>implementing ext3fs/ext4fs is impossible, or is there some other reason? My apologies if this is a dumb question.
<sneek>Welcome back Darelelve, you have 3 messages!
<sneek>Darelelve, gnucode says: if this is your first time messing with the Hurd...or your first time trying to help on open source project....I wouldn't recommend that you try building the Hurd from source. That's just really really hard. It's much easier just to run the debian hurd image in qemu.
<sneek>Darelelve, gnucode says: I personally run the hurd in real hardware. My T43 cost me $25 - $50. The T60 and T400 should work as well and should be fairly cheap.
<sneek>Darelelve, p4r4D0xum says: Yes, you need to build/have a cross-compiler in order to build kernel on another system with different kernel
<youpi>Darelelve: I can't read cyrillic so can't really grok anything without vague automatic translation
<youpi>some limits I see there is the 2G limit, which does not exist any more, see https://www.gnu.org/software/hurd/faq/2_gib_partition_limit.html
<youpi>it also talks about reading ext2fs code being difficult, well that's inherent to any filesystem
<youpi>try to read the same in Linux, you'll get the same
<youpi>the text seems to emphasize on having to modify the hurd considerably
<youpi>that has to be moderated a lot
<youpi>only libdiskfs would need to
<Darelelve>youpi: So there are currently no problems with the implementation of ext3fs/ext4fs?
<youpi>there is no fundamental problems
<youpi>it "just" needs some people to work on it
<youpi>while it'd be probably be more beneficial to just plug rumpfs
<youpi>which has its own challenges, but no fundamental problem either
<youpi>in the end the posix interface is essentially the same, the Hurd RPC interface is not really much different from that
<nikolar>lol what's funny is that i am not bulgarian and i can understand a considerable part of that article