At 5:23 PM +0200 4/9/02, Michael Zimmermann wrote: > As I saw, you preferred to set 'hangup', when a write to the client > fails (instead of Clifton Royston's solution to catch SIGHUP and SIGPIPE > signals). May I ask you - out of curiosity and the wish to learn - > for a rationale about this? Or is this a matter of taste and coding style?
This was simply a case of trying to do a minimal patch for the immediate problem. I like Clifton's solution and want to include it very soon.