Decompression of LZMA streams may fail if the input stream is too slow #1009
Labels
No labels
Affects/CppNix
Affects/Nightly
Affects/Only nightly
Affects/Stable
Area/build-packaging
Area/cli
Area/evaluator
Area/fetching
Area/flakes
Area/language
Area/lix ci
Area/nix-eval-jobs
Area/profiles
Area/protocol
Area/releng
Area/remote-builds
Area/repl
Area/repl/debugger
Area/store
bug
Context
contributors
Context
drive-by
Context
maintainers
Context
RFD
crash 💥
Cross Compilation
devx
docs
Downstream Dependents
E/easy
E/hard
E/help wanted
E/reproducible
E/requires rearchitecture
Feature/S3
imported
Language/Bash
Language/C++
Language/NixLang
Language/Python
Language/Rust
Needs Langver
OS/Linux
OS/macOS
performance
regression
release-blocker
stability
Status
blocked
Status
invalid
Status
postponed
Status
wontfix
testing
testing/flakey
Topic/Large Scale Installations
ux
No milestone
No project
No assignees
3 participants
Notifications
Due date
No due date set.
Dependencies
No dependencies set.
Reference
lix-project/lix#1009
Loading…
Add table
Add a link
Reference in a new issue
No description provided.
Delete branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
If the input LZMA stream is split / arriving too slowly, the libarchive processing machinery might receive from the LZMA code an
LZMA_BUF_ERROR:libarchive code does not recover from this error and seems to be considering it fatal, bubbling into Lix as "failed to read compression data: No progress can be made".
Solutions includes:
(Credits where it's due: all the investigative work has been done by @pennae)
Work done by @yu-re-ka seems to indicate this is caused by a bump in our dependencies:
This was confirmed to be the curl bump from 8.14.1 to 8.16.0.
dealing with the test case for this currently
will release it when this exists as we may not know how to actually fix it (though we have some ideas)
More messages as per @pennae demand:
Discovered:
raito put together a reproducer that was kind of unreliable, running it in rr chaos mode made it reliable enough to debug and bisect. https://github.com/curl/curl/pull/18428 introduces the failure by subtly changing how paused streams are handled, with the effect that pausing a stream immediately trashes it and aborts that transfer with an error.