Failed CUnit test in 2.5.12 and 3.0.8

Дилян Палаузов dilyan.palauzov at aegee.org
Mon Feb 11 11:51:48 EST 2019


Hello Max,

a possible explanation is the lack of libplain.so or libdigestmd5.so in $libdir/sasl2 , coming from cyrus-sasl	.

Until recently, the place where cyrus-sasl/configure && make install put those files, when compiled, was different from
the place, where the files were searched at run time.

Regards
  Дилян

On Fri, 2019-02-08 at 17:23 +0300, Max Kosmach wrote:
> Hi, cyrus imap devs!
> 
> I'm trying to build 2.5.12 and 3.0.8 cyrus imapd and in both versions I have same problems with backend suite:
> 
> Suite: backend
>   Test: badhost ...passed
>   Test: badservice ...passed
>   Test: sasl_plain ...Server failed to find requested SASL mechanism "PLAIN"
> FAILED
>     1. ./cunit/backend.testc:198  - CU_ASSERT_PTR_NOT_NULL_FATAL(be)
>   Test: sasl_digestmd5 ...Server failed to find requested SASL mechanism "DIGEST-MD5"
> FAILED
>     1. ./cunit/backend.testc:265  - CU_ASSERT_PTR_NOT_NULL_FATAL(be)
>   Test: multiline_caps ...Server failed to find requested SASL mechanism "PLAIN"
> FAILED
>     1. ./cunit/backend.testc:314  - CU_ASSERT_PTR_NOT_NULL_FATAL(be)
>   Test: oneline_caps ...Server failed to find requested SASL mechanism "PLAIN"
> FAILED
>     1. ./cunit/backend.testc:314  - CU_ASSERT_PTR_NOT_NULL_FATAL(be)
>   Test: starttls ...FAILED
>     1. ./cunit/backend.testc:408  - CU_ASSERT_PTR_NOT_NULL_FATAL(be)
> 
> Is this correct behavior and I can ignore this?
> 
> PS. strace show thak this test search nonexistent testserver.conf file -
> [pid 21654] openat(AT_FDCWD, "/etc/sasl2/testserver.conf", O_RDONLY <unfinished ...>
> [pid 21654] openat(AT_FDCWD, "/etc/sasl/testserver.conf", O_RDONLY) = -1 ENOENT (No such file or directory)
> [pid 21654] openat(AT_FDCWD, "/usr/lib/x86_64-linux-gnu/sasl2/testserver.conf", O_RDONLY) = -1 ENOENT (No such file or directory)
> [pid 21654] openat(AT_FDCWD, "/usr/lib/sasl2/testserver.conf", O_RDONLY) = -1 ENOENT (No such file or directory)
> 



More information about the Cyrus-devel mailing list