Re: FYI: v9fs/test rework
Eric Van Hensbergen <[email protected]> Fri, 1 May 2026 09:29:08 -0500
| Newsgroups | dev.linux.lists.v9fs |
|---|---|
| Message-ID | <CAFkjPT=sF=cxhLuun9UaX6dT6soqZze5G1cXZPJRAXfCW5QXbw@mail.gmail.com> |
Awesome, thanks for the pointers, I'll have a look. On Fri, May 1, 2026 at 9:11 AM Pierre Barre <[email protected]> wrote: > > Hi Eric, > > Glad to hear you're picking v9fs back up. In case any of it is useful as a reference point, I've been maintaining a fair amount of 9p CI for ZeroFS, exercising the Linux v9fs client against the ZeroFS server across a few transports and mount option combinations: > > https://github.com/Barre/ZeroFS/tree/main/.github/workflows > > Not a generic test harness like what you're building, but there may be some scenarios or pitfalls in there worth borrowing, including xfstests: > > https://github.com/Barre/ZeroFS/blob/main/.github/workflows/xfstests-9p.yml > > Happy to answer questions if anything looks relevant. > > Best, > Pierre > > On Sat, Apr 25, 2026, at 22:58, Eric Van Hensbergen wrote: > > I've started picking up v9fs again for something work related, but set > > out to first fix my test infrastructure. Being lazy, I'm having > > Cursor do the test infrastructure rework for me -- it is still a work > > in progress, but I think it's already handling the old tests and > > starting to add new ones. I need to suck in more of the diod testing > > framework as well so that I have full coverage there, and then I'm > > going to look at adding hugelgupf/9p and go9p. > > > > As part of this, I should be able to fix it for nightly builds and/or > > at least tracking upstream with regression tests across the multiple > > servers. > > > > Currently running in the main branch of github.com/v9fs/test > > > > Fortunately for everyone I'm not tilting at the caching stuff quite > > yet, I'm more interested in the synthetic file system angle. So don't > > worry about me breaking everything (yet). > > > > -eric