gdb: bypass unit test on non-x86 #8672
Conversation
test/scylla-gdb/run
Outdated
| @@ -1,5 +1,12 @@ | |||
| #! /bin/bash -eu | |||
|
|
|||
| if [ "$(name -m)" != x86_64 ]; then | |||
There was a problem hiding this comment.
maybe you meant uname, not name?
I'm not familiar with a tool called "name"...
There was a problem hiding this comment.
Fixed. How did that happen?!
There was a problem hiding this comment.
Simple - even if you set -e to has bash exit on error, a failure inside a command interpolation $(name -m) does not cause bash to exit. It just interpolates to nothing and prints a "command not found" warning which you probably didn't see, and the equality just never succeeds. I guess you didn't check the patch on x86_64?
The gdb self-tests fail on aarch64 due to a failure to use thread-local variables. I filed [1] so it can get fixed. Meanwhile, disable the test so the build passes. It is sad, but the aarch64 build is not impacted by these failures. [1] https://sourceware.org/bugzilla/show_bug.cgi?id=27886
avikivity
force-pushed
the
gdb-aarch64-sad
branch
from
6e74164
to
72e3a77
Compare
Unfortunately, it's up to the gdb maintainers to fix thread-local storage on aarch64. |
The gdb self-tests fail on aarch64 due to a failure to use thread-local variables. I filed [1] so it can get fixed. Meanwhile, disable the test so the build passes. It is sad, but the aarch64 build is not impacted by these failures. [1] https://sourceware.org/bugzilla/show_bug.cgi?id=27886 Closes #8672
The gdb self-tests fail on aarch64 due to a failure to use thread-local
variables. I filed [1] so it can get fixed.
Meanwhile, disable the test so the build passes. It is sad, but the aarch64
build is not impacted by these failures.
[1] https://sourceware.org/bugzilla/show_bug.cgi?id=27886