Add zip support and streaming append to Bun.Archive - #33318

Open

robobun wants to merge 19 commits into

mainfrom

farm/b16f2c14/archive-zip-streaming-append

Commits on Jul 12, 2026

Add zip support and streaming append to Bun.Archive
robobun
committed
38c3259
Stream Bun.file() entries in the object form instead of writing empty files
robobun
committed
91e3bfa
clippy: use ?, or_else, and bun_core::time::timestamp()
robobun
committed
44a4eeb
Reject NUL bytes in entry paths, and keep an explicit compress override
robobun
committed
08771a8
Don't pad an abandoned entry out to its declared size
robobun
committed
66d1444
Deliver a failed append's rejection before the cascade it triggers
robobun
committed
afdbaed
docs: note that format has no effect on an already-built Archive
robobun
committed
42e1359
Archive a sliced Blob's window, not its parent's bytes
robobun
committed
1e3f49a
Write an entry's bytes in a loop; archive_write_data clamps to INT_MAX
robobun
committed
b3aacc0
ci: retrigger
robobun
committed
e290ac3
Remove the spill-to-disk sink
robobun
committed
f613e00
Stop claiming a memory bound the spill removal took away
robobun
committed
2560cc2
Add archive.stream(), so a large archive never has to fit in memory
robobun
committed
cd5fc71
Don't leave a stream()'s consumer waiting on an append that failed
robobun
committed
4af5d64
Don't park a stream()'s append against a buffer nothing drains
robobun
committed
a9f0863
Don't claim backpressure on the Bun.serve path, and signal the stream when end() fails
robobun
committed
ef4ff5c
Cover the getReader() path of a failed append's stream error
robobun
committed
b522ae4
Cover the awaiting-loop-body shape of the getReader() consumer too
robobun
committed
80dee1b
Keep TaskContext's doc comment attached to TaskContext
robobun
committed
9c3a670

Navigation Menu

Uh oh!

Add zip support and streaming append to Bun.Archive - #33318

Add zip support and streaming append to Bun.Archive#33318
robobun wants to merge 19 commits into
mainfrom
farm/b16f2c14/archive-zip-streaming-append

Commits on Jul 12, 2026

Add zip support and streaming append to Bun.Archive

Stream Bun.file() entries in the object form instead of writing empty files

clippy: use ?, or_else, and bun_core::time::timestamp()

Reject NUL bytes in entry paths, and keep an explicit compress override

Don't pad an abandoned entry out to its declared size

Deliver a failed append's rejection before the cascade it triggers

docs: note that format has no effect on an already-built Archive

Archive a sliced Blob's window, not its parent's bytes

Write an entry's bytes in a loop; archive_write_data clamps to INT_MAX

ci: retrigger

Remove the spill-to-disk sink

Stop claiming a memory bound the spill removal took away

Add archive.stream(), so a large archive never has to fit in memory

Don't leave a stream()'s consumer waiting on an append that failed

Don't park a stream()'s append against a buffer nothing drains

Don't claim backpressure on the Bun.serve path, and signal the stream when end() fails

Cover the getReader() path of a failed append's stream error

Cover the awaiting-loop-body shape of the getReader() consumer too

Keep TaskContext's doc comment attached to TaskContext