- From: David W. Morris <dwm@xpasc.com>
- Date: Sat, 19 Apr 1997 10:35:19 -0700 (PDT)
- To: Jeffrey Mogul <mogul@pa.dec.com>
- Cc: http-wg@cuckoo.hpl.hp.com
On Fri, 18 Apr 1997, Jeffrey Mogul wrote: > But this is probably just "advice to client implementors", not > something we need to put into the spec. (And I've seldom seen > any content-generator realize when it has generated bogus data, > although I have certainly seen enough cases where the output was > clearly broken.) Probably because if the server/application cares enough to know the data might be bogus, it has made provisions to purge the partial output and start over. If you've been busy generating HTML and forwarding it to a client and discover some kind of error condition, it is really tough in the general case to generate additional HTML which one can be sure will be seen by the user in a meaningful way. (For example, if one is in the middle of generating a table but the error routine doesn't close the table, the browser may not display the error message, etc.) My latest project takes the effort to queue all output until the response is complete. If an intermediate failure is detected. the queued output is purged and an error response generated. Dave Morris
Received on Saturday, 19 April 1997 10:40:31 UTC