On 10/10/2012 01:41 PM, Adam D. Barratt wrote: > On Mon, 2012-08-27 at 21:48 -0700, tony mancill wrote: >> On 08/25/2012 06:15 PM, Cyril Brulebois wrote: >>> Looking at the diff (please bear in mind I know little to nothing in >>> java/jni, so those are naïve questions): >>> - is tcn_socket_t for internal use only or is it exposed outside the >>> library? >>> - same question for tcn_pollset_t. >>> - is it ok to get rid of some functions, like update_ttl? >>> >>> If answers are 2*internal+yes, we should be able to unblock it. > [...] >> Here is the change you're referring to: >> >>> -static void update_ttl(tcn_pollset_t *p, const apr_pollfd_t *fd, apr_time_t t) >>> +static void update_last_active(tcn_pollset_t *p, const apr_pollfd_t *fd, apr_time_t t) >> >> update_ttl() (and it's replacement, update_last_active()) are only used >> in poll.c and is declared as static, so I don't see how that would >> change users of the library. > > Apologies for the delay in getting back to you again. > > Looking through the earlier conversation, I couldn't see a final answer > regarding whether tcn_socket_t is exported by the library. The reason > for the query being: > > tcn_nlayer_t *net; > + apr_time_t last_active; > + apr_interval_time_t timeout; > } tcn_socket_t; Hi Adam, Okay, I understand the reason for the question better. Thanks for the follow-up. Sorry I'm not able to answer right off; I'm not much with C. If tcn_socket_t were exported by the library wouldn it show up in a symbol dump? The following doesn't return any anything. $ nm -a libtcnative-1.a | grep -i tcn_socket Also, there the source package contains a file tcn_api.h that carries this comment: /** * @file tcn_api.h * @brief * * Tomcat Native Public API */ The tcn_tocket_t doesn't appear in any of the prototypes in that file. I realize the comment isn't definitive. Is there something else I should check in the resulting library? Thank you, tony
Attachment:
signature.asc
Description: OpenPGP digital signature