-
Notifications
You must be signed in to change notification settings - Fork 2.5k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Crash on SQLite insert query #12079
Comments
Are you able to reproduce this outside of alpine linux? Musl LibC is not supported right now in Bun (#918), and using glibc compat can cause strange issues like crashes when loading e.g. Prisma Edit: I think it's unlikely that alpine is the cause - this stacktraace is entirely coming from JSC. |
Yes, I be able reproduce crash outside MUSL environment. Before Docker+Alpine+MUSL I run application in single-binary variant on Debian GNU/Linux 11 (bullseye). To make single binary, I run this command:
OR
Run build on macOS Sonoma 14.5 or on Debian 11 - not important. Same result. |
I haven't been able to reproduce it from the example on either linux x64 or macOS arm64 or on a debug build of macOS arm64, but one idea: if you move the |
This two variants produces same error:
|
I miss important thing in the description of the bug, but this thing in Makefile:
Only intel/amd64 binary have this bug. On arm64 processors all OK. |
Make some refactoring, middle Promise now absent, but error not gone.
1.07GB - it's not normal, if this metric shows us resident memory consumption, seems like memory leak. UPD: If I change line 73 and replace UPD: If I comment out line 47 UPD: If I comment out line 22 UPD: If I comment return UPD: Tryed to turn on UPD: Tryed to check theory about conflicting concurrent SQLite queries. After |
@kozharskyad in about an hour (once CI finishes), can you try running bunx bun-pr 12079 This script downloads the build artifact of Bun from a PR. Can you try running in that version of Bun and see if the issue still happens? I haven't been able to verify this, but here are two guesses I want to test:
|
Probably no success with pipeline - https://github.com/oven-sh/bun/actions/runs/9642668629
|
Download artifact from #12111 success, but:
It's on macOS, force linux/amd64 platform. Probably artifact from pipeline is not baseline version. UPD: Ok, found UPD: Got the same error:
|
@kozharskyad can you run I'm wondering if it's a JavaScriptCore bug. Specifically, I'm wondering if this PR WebKit/WebKit#30298 fixes it (which is included in that branch above) |
Sorry, but no success :(
Unbelievable, but crush address always the same: 0x80 |
How can we reproduce the crash?
OR
Difference of this machines probably means machine architecture is not that important.
On
oven/bun:1.1.13-alpine
crash not reproducible.You can also open Makefile and see "crash" task with key commands:
Relevant log output
Stack Trace (bun.report)
Bun v1.1.16 (
bf7b327
) on linux x86_64_baseline [AutoCommand]Segmentation fault at address 0x00000080
ld-temp.o:0
:WTF::jscSignalHandler
??
ld-temp.o:0
:llint_op_call_varargs
ld-temp.o:0
:llint_op_call_ignore_result
ld-temp.o:0
:llint_op_call
ld-temp.o:0
:llint_call_javascript
ld-temp.o:0
:JSC::Interpreter::executeCall
The text was updated successfully, but these errors were encountered: