State of this fork's delta against upstream/master.
Proposed upstream
Held, with reasons
Module rename to github.com/0magnet/afero, and dropping HttpFs — the rename is fork-only by definition. Dropping HttpFs is not: it exists because HttpFs pulls in net/http, which does not currently build under TinyGo js/wasm (roundtrip_js.go). That is a TinyGo problem, not an afero problem, so deleting the type upstream would be the wrong fix — the right one is fixing TinyGo's net/http, tracked in 0magnet/net#1.
Notes
The upstreamed version is not the fork's patch as written. The fork gated on //go:build tinygo, which is too broad: only TinyGo's js/wasm target is missing these functions. Checked each target rather than assuming —
| TinyGo target |
os.Chmod / Chown / Link |
linux (native) |
present |
wasip1 |
present |
wasm (js) |
absent |
so the PR narrows the constraints to tinygo && js. Worth folding that correction back into this fork, since as it stands it stubs out working functions on TinyGo native and wasip1 builds.
State of this fork's delta against
upstream/master.Proposed upstream
Chmod/Chown/Linkwrappers with ENOSYS stubs, andBADFDfalling back toEBADFHeld, with reasons
Module rename to
github.com/0magnet/afero, and droppingHttpFs— the rename is fork-only by definition. DroppingHttpFsis not: it exists becauseHttpFspulls innet/http, which does not currently build under TinyGo js/wasm (roundtrip_js.go). That is a TinyGo problem, not an afero problem, so deleting the type upstream would be the wrong fix — the right one is fixing TinyGo'snet/http, tracked in 0magnet/net#1.Notes
The upstreamed version is not the fork's patch as written. The fork gated on
//go:build tinygo, which is too broad: only TinyGo's js/wasm target is missing these functions. Checked each target rather than assuming —os.Chmod/Chown/Linklinux(native)wasip1wasm(js)so the PR narrows the constraints to
tinygo && js. Worth folding that correction back into this fork, since as it stands it stubs out working functions on TinyGo native and wasip1 builds.