This avoids disabling the artifact until cache is reopened. The same has to be implemented for Cancel in a future change.
Signed-off-by: Ophestra <cat@gensokyo.uk>
These are memory management and caching primitives. Having them as part of Cache is cumbersome and requires a temporary directory that is never used. This change isolates them from Cache to enable independent use.
Signed-off-by: Ophestra <cat@gensokyo.uk>
This is cleaner than checking cpu count during cure, it is impossible to avoid impurity in both situations but this is configurable.
Signed-off-by: Ophestra <cat@gensokyo.uk>
XZ happens to be the only widely-used format that is awful to deal with, everything else is natively supported.
Signed-off-by: Ophestra <cat@gensokyo.uk>
This is quite error-prone and causes cache inconsistency similar to the store inconsistency seen on nix when a similar condition happens. Keep this behind a flag in case it is ever beneficial.
Signed-off-by: Ophestra <cat@gensokyo.uk>
This makes package check portable, and removes nonportable behaviour from package pkg, pipewire, and system. All other packages remain nonportable due to their nature. No latency increase was observed due to this change on amd64 and arm64 linux.
Signed-off-by: Ophestra <cat@gensokyo.uk>
These are too general to belong in the container package. This targets the v0.4 release to reduce the wrapper maintenance burden.
Signed-off-by: Ophestra <cat@gensokyo.uk>
The more explicit API removes the arbitrary limit preventing use of SCHED_OTHER (referred to as SCHED_NORMAL in the kernel). This change also exposes priority value to set.
Signed-off-by: Ophestra <cat@gensokyo.uk>
This side is the read end of a pipe and buffering reads from it ended up performing better than buffering one half of the TeeReader (which already goes through the kernel page cache anyway).
Signed-off-by: Ophestra <cat@gensokyo.uk>
This is not as necessary as it was for nix, since internal/pkg only unblocks exclusive artifacts one at a time. Still, this is useful when running alongside an unprivileged music player which cannot set itself to a higher priority.
Signed-off-by: Ophestra <cat@gensokyo.uk>
This got brought up earlier today as a potential optimisation. This change documents why it is not viable, and hopefully clears up some performance implications of using IRDecoder, namely that its decoding costs do not amortise.
Signed-off-by: Ophestra <cat@gensokyo.uk>
This significantly improves performance and is a good assumption since the primary use case of FileArtifact is over the network.
Signed-off-by: Ophestra <cat@gensokyo.uk>
TContext no longer validates FileArtifact ahead of time, validation outcome is instead determined after consuming the reader to EOF. All data must therefore be treated as untrusted input until the reader is closed.
Signed-off-by: Ophestra <cat@gensokyo.uk>
This should hopefully provide good separation between the artifact curing backend implementation and the (still work in progress) language. Making the IR parseable also guarantees uniqueness of the representation.
Signed-off-by: Ophestra <cat@gensokyo.uk>
cat
