> From: Robert Waldner <[EMAIL PROTECTED]>
> well, this naturally led me to think that keeping those records local > should speed up starting netscrap, so I had tcpdump running while starting > netscrap but I didn´t see a single packet (having, of course, made sure that > those records aren´t cached). the netscrap-version I´m using is 4.61...so > much for trying to smart it out :-/ I used strace on Communicator a couple of weeks ago, and found a couple of interesting things. - When Netscape (well, Communicator) hangs chewing up lots of CPU but doing nothing, it seems (if I didn't mix up the pipe file descriptor numbers) to be trying to read from a resolver process that already closed its end of the pipe. - Sometimes Netscape would freeze for a while, again using lots of CPU, but then it would come back after a while. However, if Java had been started (by visiting a page with an applet), then when Netscape froze (with high CPU usage), it would not come back. - Have a large (~3000) number of bookmarks seems to aggravate the problem. - Something in the combination of problems points to memory leaks or screwed up heap-allocation code. I sure hope Mozilla is written competently. Daniel -- Daniel Barclay [EMAIL PROTECTED] (Hmm. A little worrisome: http://www.junkbusters.com/cgi-bin/privacy http://www.anonymizer.com/snoop.cgi )
