Cyrus IMAP 3.0.0 and Oracle Solaris/Compiler
Jean-Christophe Delaye
Jean-Christophe.Delaye at eurecom.fr
Wed Apr 5 12:14:21 EDT 2017
On 05/04/17 15:38, Stefan Rapp wrote:
> Hello,
Hi
What is your native compiler version ?
(solarisstudio12.4 or developerstudio12.5)
>
> i tried to compile Cyrus IMAP 3.0.0 on Oracle Solaris (x86) using the
> native compiler in 64bit mode. I used the same settings as for a working
> 2.5.10 setup. During start up i got "twoskip header CRC failure" opening
> mailboxes.db.
>
> To test my setup and to experiment with code changes i compiled the
> CUnit checks, found and solved some problems:
>
> - The command_popen for /usr/bin/tr in cunit/command.testc needs
> brackets ("[" and "]") for character classes.
>
> - LITTLE_ENDIAN used in lib/crc32.c, lib/crc32c.c and lib/xsha1.c is not
> defined. The configure script generates a correct definition of
> WORDS_BIGENDIAN. So the checks for LITTLE_ENDIAN should be replaced with
> negated checks for WORDS_BIGENDIAN.
>
> - The regcomp-call in lib/glob.c needs REG_EXTENDED as the last parameter.
>
> - There is no qsort_r support in Solaris and the construct used for GNUC
> doesn't work with the Oracle Compiler. I ended up using the BSD-Source
> for qsort_r.
>
> - The function cmpstringp_mbox in lib/bsearch.c returns 0 for strings
> that are of differing lengths and identical until the end of the first
> string (error in Suite: strarray, Test: sortuniq, mbox order sort). Adding
>
> if (cmp == 0) {
> cmp = TOCOMPARE(*s1) - TOCOMPARE(*s2);
> }
>
> before "return cmp;" helped (this error is also in 2.5.10, but i didn't
> notice before because i did no CUnit checks).
>
>
> I still have 16 errors in the CUnit sieve checks, but this is identical
> to 2.5.10. As the sieve scripts (seem to) work in our setup this could
> be a problem of the checks ???
>
>
> Greetings,
>
> Stefan
>
>
>
> ----
> Cyrus Home Page: http://www.cyrusimap.org/
> List Archives/Info: http://lists.andrew.cmu.edu/pipermail/info-cyrus/
> To Unsubscribe:
> https://lists.andrew.cmu.edu/mailman/listinfo/info-cyrus
>
More information about the Info-cyrus
mailing list