On Oct 28, 2005, kai wrote:
You seem to have overlooked the fact that my suggestion checks for *two* changes of name - not just one. You're right, I had not noticed! Thanks for pointing it out.
I haven't been able to replicate that, but I'll take your word for it.
You may well be right about readyState, although repeated tests here, loading numerous different web pages, suggest that changes of status occur virtually simultaneously. The following summary, extracted from the event log of the test script (from which delays were removed for finer granularity) is fairly typical of the results:
I have run those results many times and over long periods. I guess it is true that special problems occur with my own database sites, which run slow at times. All problems seem to derive from slow responsiveness: on my machine or on the target website or on the internet. Sometimes the script runs for several hours accessing hundreds of pages, and I lot of screwy things can happen that are not easy to replicate. My code decently, not perfectly. Hence the reason for this posting: a plea for perfect reliability for such a simple operation as getting html from a website!
Agreed - although your code seems to have altered somewhat since your original post... ;-) Ugh. I made some effort to re-arrange my code to show a nice apples-to-apples comparison. I realized it was a little hard to follow because of several handlers, and passing an id reference to my hidden capture-window. In other words, I was making effort in appreciation of the work you were already doing on my behalf - which I really do appreciate!
|