[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: commit: ldap/libraries/libldap result.c

Hallvard B Furuseth wrote:
Howard Chu writes:
Pierangelo Masarati wrote:
Not strictly related, but does it make sense to have a malloc for a
struct timeval in global/per handler data?  Wouldn't it be any better to
use tv_sec = -1 to indicate no timeout?  We should be able to change
that, since the struct is opaque...

Hm. What about changing the get_option call? Currently it mallocs a timeval
and returns a pointer to that. It ought to just copy the contents of the
timeval. They've only been documented in HEAD; can we change it or would that
break something?

Do you mean ldap_get_option()?  It generally returns malloced data,
IIRC.  "undocumented" just means the code has been the documentation, so
I presume a change will break any old code which uses it.

Yes to all of the above. But since LDAP_OPT_TIMEOUT was completely unimplemented before, I don't think it really makes any difference. I.e., the option was not in the LDAPv2 API or the LDAPv3 draft API, and it never actually did anything. Why would it have ever gotten used?

  -- Howard Chu
  Chief Architect, Symas Corp.  http://www.symas.com
  Director, Highland Sun        http://highlandsun.com/hyc
  Chief Architect, OpenLDAP     http://www.openldap.org/project/