Ok, I've dug some more on this one and would appreciate some comments on this -- where to look further ...
... as www.cnn.com is a bit large for a test case, I added debug printfs to convert_all_links() to count and print all instances where the url contained the string "1.gif" -- the file whose URL doesn't always get converted to a local/relative link.
there were 54 instances found; albeit split between 2 servers, so that 2 files were actually downloaded -- 1.gif and 1.gif.1.
after tidy-ing and grep-ing the output index.html, there were 58 instances of 1.gif being mentioned. 4 failed to be converted to mention the local file.
[ I could cut the test case down considerably if I could figure out how to submit a html file to wget for processing, rather than have it wade thru cnn.com ... ]
any insights as to how to continue to debug this one? -- I'd appreciate more comments than already rec'd [none :-O]. is this a known bug? is there some limit I'm hitting ...?
thanks,
Mark