Xu, Qiang (FXSGSC) wrote:
It just seems to me that the behavior of ldap_simple_bind() is
> almost identical to the behavior of ldap_simple_bind_s(). It
shouldn't be, right? Could you hightlight more on this issue, Anton?
i think thats kinda common misconception about async calls in libldap. they are not I/O async they are simply LDAP level async. ie sync calls send requests and wait for responses before they return and that can take arbitrary amount of time due to what happens with requests on the server side. async libldap calls allow you so send requests then poll for responses at your own pace so if the server takes too much time to process requests you have an option of doing something else or even abandoning such requests if they take way too long. async in this case however does not mean raw I/O level async. we do however have some options, as the one suggested for NSPR/NSS usage, to enforce timeouts on raw I/O level itself. _______________________________________________ dev-tech-ldap mailing list [email protected] https://lists.mozilla.org/listinfo/dev-tech-ldap
