Test PkgEval effect of deleting slow path debuginfo lookup on 1.12 #58676
Conversation
This slow-path is hard to justify, both because it is slow (it scans 1-by-1 through every symbol in a library) and also because its results are very frequently thrown out and replaced with fast-lookup from the DWARF info right after. The latter issue could be resolved with some re-factoring, but the former is bad enough that I think we should just delete this and take the policy that stack frame lookup requires unwind or debug info in your libraries. (cherry picked from commit f1b424a)
|
@nanosoldier |
|
The package evaluation job you requested has completed - no new issues were detected. Report summary❗ Packages that crashed1 packages crashed only on the current version.
✖ Packages that failed17 packages failed on the previous version too. ✔ Packages that passed tests6 packages passed tests only on the current version.
7 packages passed tests on the previous version too. ➖ Packages that were skipped altogether1 packages were skipped only on the current version.
1 packages were skipped on the previous version too. |
|
Both OrdinaryDiffEq and GPUCompiler seems to be broken at thet time when running this. OrdinaryDiffEq should have been fixed by JuliaObjects/ConstructionBase.jl#102 and GPUCompiler needs a tweak. |
|
@nanosoldier |
|
The package evaluation job you requested has completed - possible new issues were detected. Report summary✖ Packages that failed5 packages failed only on the current version.
13 packages failed on the previous version too. ✔ Packages that passed tests7 packages passed tests only on the current version.
7 packages passed tests on the previous version too. ➖ Packages that were skipped altogether1 packages were skipped on the previous version too. |
There are quite a few packages that time out on 1.12. I want to see if it has to do with this slow debuginfo lookup that happens for deprecation warnings.