11 Feb 2005 18:34
Re: Error from mh-index-search
Eric Jensen <ejensen1 <at> swarthmore.edu>
2005-02-11 17:34:46 GMT
2005-02-11 17:34:46 GMT
OK, thanks for your help. Below is that buffer's content (between the --- lines). Aha! I see from this output that it found a match in my "aliases" file, which is obviously not a mail message, so I should put "aliases" in my ignore list in the swish-e config... I wonder if part of the problem is that the indexing is catching a few files that aren't mail messages, and that's confusing the scan formatting later? OK, I just fixed my swish-e config file to ignore this file, re-made the index, and re-ran the indexed search, and sure enough, that fixed it! So it looks like part of the problem is how to handle more robustly the case when search results are returned that aren't actually mh messages. And correspondingly, how to modify the recommend default .swish/config in the docs to help avoid this in the first place. Maybe a line like: FileMatch filename contains ^\d+$ By the way, I'd also recommend "DefaultContents TXT*" since otherwise swish-e use HTML2, which takes a lot longer to do the indexing. Are there any cases where messages would be stored with filenames that contains characters other than digits? Thanks as always for the quick help! Eric(Continue reading)

RSS Feed