Checking for internet connection.
in [General]
|
Prev: Class "COM" not found
Next: SQL Queries
From: Kim Madsen on 21 Dec 2009 12:42 Bob McConnell wrote on 21/12/2009 15:05: > Both at home and at work there are caching DNS on the LAN. So a DNS > request may come back with a valid IP address when the WAN connection is > down. I still won't be able to connect to the remote site. Then use fopen() to read a page you know exists? -- Kind regards Kim Emax - masterminds.dk
From: Andy Shellam on 22 Dec 2009 15:07 > > I'm confused... what's the problem with just trying to hit the update server? If you can then you check for updates, if not then you, erm, don't. Simples, no? True, I think I said this same thing in a previous post - I suggested the DNS option if all the OP wanted to do was check if an Internet connection was there.
From: "Bob McConnell" on 22 Dec 2009 15:16 From: Andy Shellam >> >> I'm confused... what's the problem with just trying to hit >> the update server? If you can then you check for updates, if >> not then you, erm, don't. Simples, no? > > True, I think I said this same thing in a previous post - I > suggested the DNS option if all the OP wanted to do was check > if an Internet connection was there. And I was pointing out that this would not be a valid test when there is a caching DNS on the LAN. Too much of the conversation and most of the attribution was stripped too early for this to be coherent. Bob McConnell
From: Andy Shellam on 22 Dec 2009 17:03 > > And I was pointing out that this would not be a valid test when there is > a caching DNS on the LAN. I also pointed out how to avoid caching issues - the comment was aimed at the author of the message before mine. > > Too much of the conversation and most of the attribution was stripped > too early for this to be coherent. Why the negativity? A question was asked and several possible solutions were provided based on that original question. All the "conversation" was relevant IMO.
From: "Bob McConnell" on 23 Dec 2009 08:35
From: Andy Shellam >> And I was pointing out that this would not be a valid >> test when there is a caching DNS on the LAN. > I also pointed out how to avoid caching issues - the > comment was aimed at the author of the message before mine. > >> Too much of the conversation and most of the attribution >> was stripped too early for this to be coherent. > > Why the negativity? A question was asked and several > possible solutions were provided based on that original > question. All the "conversation" was relevant IMO. But long before it was done it was impossible to tell who had asked which questions, who had provided which answers and who had countered those answers. In several instances, replies appeared to be directed to the wrong individuals. Some people here tend to go way too far when trimming context from replies. Yes, I know it gets difficult to read when there are more than ten or twelve levels of attribution, but stripping all but the last layer is even worse. Removing the participants names from the top should be a hanging offense. I don't keep copies of every message in any of the dozens of mailing lists and news groups I follow, so there is no simple way to go back through the conversation to figure out where it all came from. Bob McConnell |