Closed jbglaw closed 1 year ago
@jbglaw Thank you for the bug report, and the fix! The master branch has been updated accordingly.
Also, I'm glad to see someone making use of retro-fuse. I hope it's working well for you.
Sure! This is a nice project, especially due to its direction of using original sources.
Just out of curiousity: Did you ever find a real bug in former upstream filesystem drivers?
Just out of curiousity: Did you ever find a real bug in former upstream filesystem drivers?
I did! I discovered a bug in v7 that leaked an inode whenever the fs ran out of space at just the right moment (fixed here). Interestingly, the bug also appeared in v6 and in 2.9bsd. It was finally fixed (in the bsd lineage at least) by the time 2.11bsd shipped.
Just another thought: Maybe add an install
target with the usual $prefix
/$DESTDIR
stuff? For the CI builds (http://toolchain.lug-owl.de/laminar/jobs/retrofuse), I keep the build artifacts. (Ie. the SIMH pdp11
used for testing within retro-fuse
is an extracted tarball of my OpenSIMH CI builds.) Right now, I'm guesstimating the four FUSE binary names. The Makefile
though has them all (with $(ALL_PROGS)
), so that would be a no-brainer to implement?
Hi!
As I found the retro-fuse binaries useful, I wanted to add them to my CI builds setup (see http://toolchain.lug-owl.de/). Building was just a breeze, but I see unexpected errors during
make test
. In my setup, everything is built within a Docker container, but even runningmake test
as root gives me some issues:With this patch, things go further:
...and tests pass with current OpenSIMH: