Hello Przemek
Przemyslaw Czerpak-2 wrote: > > I guess that you are talking about client application not server > and it stops inside connect(). > You are correct. > Usually low level TCP/IP code in the > OS should interrupt this operation after some timeout if it cannot > be realized. Sometimes this timeout is really huge. It depends on many > things, i.e. average time of packet confirmation which in TCP IP > is dynamically updated following the transport layer quality. > Anyhow it's out of Harbour or any other application controll. > If connection freeze the process forever then I can only guess > it's side effect of used firewall or some other similar protection > software. Otherwise it's serious bug in TCP/IP implementation which > should be reported to MS but I rather do not believe that Windows > has such serious problem with basic functionality. > If you can exclude some other trivial problem like mistakes in > routing tables or sth like that then for me it has to be problem > with firewall which blocks connection only partially and may > send some sync packages waiting for user interaction (i.e. to > confirm if connection is legal) what effectively blocks timeout > on client side. > If you want to you can pass as 3-parameter of netio_connect() timeout > in milliseconds. Default is -1 what means not user space timeout and > operation should be interrupted by OS inside connect() function. > If you use 10000 then after 10 sec. netio_connect() will return .F. > signaling the problem. > Of course it's only workaround with protection against freezing > the application not real solution but because the problem is outside > Harbour I cannot suggest you anything more as programmer solution > (I do not want to introduce some low level TCP/IP hacks because in > most of cases they increase the problem instead of help ;-)) > Look at your firewall configuration. It's possible that the problem > is trivial, i.e. firewall needs valid DNS entry for client and/or > server station and it's enough to disable such verification. > Thanks for the explanation. I will make tests with some real-time environments and will report back. Regards Pritpal Bedi -- View this message in context: http://www.nabble.com/NETIO---Real-Web-IP---Application-Behavior-tp25440212p25455343.html Sent from the Harbour - Dev mailing list archive at Nabble.com. _______________________________________________ Harbour mailing list Harbour@harbour-project.org http://lists.harbour-project.org/mailman/listinfo/harbour
