From: Douglas K. <kl...@he...> - 2004-03-25 21:51:35
|
I've got some more experimental results on the errors I've reported. Increasing the number of "v"'s in the rundig from 10 to 16 didn't produce any more diagnostics. Is 10 the maximum number of "v"'s which can have effect? Running tests on various search strings with the databases created with rundig executions which reported the errors I pass on on Tuesday, I find that I can get at least three kinds of erroneous results. With one string, a string found in the top-level page in the URL given as the starting point for the htdig and in some other pages, I get a list of pages in the search database of which most are described with an error message similar to the one returned by htdig and two pages without the error message which include the search string. With the string "http" which occurs in URL's, of course, but also in other html lines and is in the top-level page, I get a report of pages with the search string which really have the search string without the error message and other pages which have the search string are not reported by htsearch but I think that these pages may all be accessed through Javascript and not be referenced by LINK tags and therefore not be scanned by htdig. With other search strings which are not present in the top-level page, I get a response that the string wasn't found, even though it is in some pages. So the htsearch report for the string "http" seems to be accurate, the reports for other strings which aren't in the top-level page miss the pages which have the strings altogether, and the report on the string which is in the top-level page and many others seems to correctly report the matching files but also reports the error message for most of them. Does this provide enough diagnostic information to explain the errors I've reported? TIA. Douglas ======== Douglas Kline kl...@he... |