#!watchflakes
default <- pkg == "crypto/ecdsa" && test == "TestKeyGeneration"

Issue created automatically to collect these failures.

Example (log):

--- FAIL: TestKeyGeneration (0.00s)
    --- FAIL: TestKeyGeneration/P256 (0.00s)
        ecdsa_test.go:68: public key invalid: %!s(<nil>)

— watchflakes

Comment From: gopherbot

Found new dashboard test flakes for:

#!watchflakes
default <- pkg == "crypto/ecdsa" && test == "TestKeyGeneration"
2024-03-29 23:09 linux-ppc64le-power9osu go@89f7805c crypto/ecdsa.TestKeyGeneration (log) --- FAIL: TestKeyGeneration (0.00s) --- FAIL: TestKeyGeneration/P256 (0.00s) ecdsa_test.go:68: public key invalid: %!s()

— watchflakes

Comment From: gopherbot

Found new dashboard test flakes for:

#!watchflakes
default <- pkg == "crypto/ecdsa" && test == "TestKeyGeneration"
2024-03-29 23:09 linux-ppc64le-power10osu go@89f7805c crypto/ecdsa.TestKeyGeneration (log) --- FAIL: TestKeyGeneration (0.00s) --- FAIL: TestKeyGeneration/P256 (0.00s) ecdsa_test.go:68: public key invalid: %!s()

— watchflakes

Comment From: thanm

This failure is not a flake, it was caused by my mergelocals CL 553055 which was subsequently reverted. I am looking at the PPC64 issue now.

Comment From: thanm

Closing this one out, the stack slot merging is enabled by default now and we're not seeing any more of these failures. The powerpc problems in particular were addressed in CL 576735.