strace 5.13 in core-updates-frozen appears to build fine but fails its
"readlinkat" test for me on AArch64 (real hardware; a ROCK64). This is
the only test that fails.
>From the build directory, strace-5.13/tests/readlinkat.dir/exp contains
readlinkat(AT_FDCWD,
"\x74\x65\x73\x74\x2e\x72\x65\x61\x64\x6c\x69\x6e\x6b\x61\x74\x2e\x6c\x69\x6e\x6b",
0xfffff7e2cfea, 22) = -1 ENOENT (No such file or directory)
readlinkat(AT_FDCWD,
"\x74\x65\x73\x74\x2e\x72\x65\x61\x64\x6c\x69\x6e\x6b\x61\x74\x2e\x6c\x69\x6e\x6b",
"\x74\x65\x73\x74\x2e\x72\x65\x61\x64\x6c\x69\x6e\x6b\x61\x74\x2e\x74\x61\x72\x67\x65\x74",
22) = 22
+++ exited with 0 +++
but strace-5.13/tests/readlinkat.dir/out shows
readlinkat(AT_FDCWD,
"\x2f\x70\x72\x6f\x63\x2f\x73\x65\x6c\x66\x2f\x65\x78\x65",
"\x2f\x74\x6d\x70\x2f\x67\x75\x69\x78\x2d\x62\x75\x69\x6c\x64\x2d\x73\x74\x72\x61\x63\x65\x2d\x35\x2e\x31\x33\x2e\x64\x72\x76\x2d"...,
4096) = 62
readlinkat(AT_FDCWD,
"\x74\x65\x73\x74\x2e\x72\x65\x61\x64\x6c\x69\x6e\x6b\x61\x74\x2e\x6c\x69\x6e\x6b",
0xfffff7e2cfea, 22) = -1 ENOENT (No such file or directory)
readlinkat(AT_FDCWD,
"\x74\x65\x73\x74\x2e\x72\x65\x61\x64\x6c\x69\x6e\x6b\x61\x74\x2e\x6c\x69\x6e\x6b",
"\x74\x65\x73\x74\x2e\x72\x65\x61\x64\x6c\x69\x6e\x6b\x61\x74\x2e\x74\x61\x72\x67\x65\x74",
22) = 22
+++ exited with 0 +++
The only difference is an additional line at the start of the generated
output. I see this in the strace package definition
(gnu/packages/linux.scm:2256):
;; XXX: This test fails because an extra readlink call is made
;; by the glibc when using the ld.so cache.
(("readlink.gen.test[^:]") " ")
Perhaps the same is true for readlinkat on AArch64?
--
Simon South
si...@simonsouth.net
