For now disable upstream's debug info stripping. #142
Conversation
This avoids trying to add a debuglink to the resulting binary twice which then fails the CRC check when reading the debuginfo. Closes Karm#138
|
We should create a quarkus issue which deals with this as well. Right now, I'm leaning towards keeping the quarkus approach until upstream is properly fixed (it should use relative paths for links). |
I am personally against merging this (i.e. working around the issue in the test suite). I think we should directly fix it upstream by keeping the quarkus approach at least for the time being (I can prepare the PR on Monday). In the long run it might be good to adopt the GraalVM approach for consistency. |
|
Could we leave it failing for now? Or is it too much noise? |
It is too much noise. We keep missing failures. But if a quarkus fix is about to land, that's good too. Otherwise, we should land it, then revert once a quarkus fix is in. |
|
I am on it :) Update: quarkusio/quarkus#31294 |
|
quarkusio/quarkus#31294 fixes this on the quarkus side. Withdrawing this one. |
This avoids trying to add a debuglink to the resulting binary twice which then fails the CRC check when reading the debuginfo.
Closes #138