On Jul 25, 2006, at 21:05, Richard Conroy wrote: > MSIE does not close down instantaneously. Don't immediately > try to relaunch it.
Great, that's some light :-). I need to understand well what that means. Which is the relationship between closing ie (either via the close method or by hand closing a window) and launching another crawling? By now, in development, we just launch one or two (but pointing to different sites anyway), observe what it does, and close windows by hand or leave them there after done. We are not paying attention to the way we close windows, what should we be careful with? Now that we are on it, if there are several crawlers scrapping different sites, do they use the same object or each window represents a separate object? I am starting the crawling with ie.goto and works fine in the machine I develop, should I use ie.start for the starting URL instead? If so, could that be causing what we are seeing? > Another solution is to never close, or keep an IE window open > somewhere (so that browser closes aren't actually closing down > MSIE). Yeah, I guess with the previous questions I'll understand better the implications of closing ie. > We needed to flush login sessions properly with a non-deal > app, so we just delayed. I don't understand this, you mean you programmed logouts? -- fxn _______________________________________________ Wtr-general mailing list [email protected] http://rubyforge.org/mailman/listinfo/wtr-general
