Option to *not* recurse directories
"David Brown" <[email protected]> Fri, 22 Sep 2006 19:10:25 -0700 (PDT)
| Newsgroups | gmane.comp.web.swish-e |
|---|---|
| Message-ID | <008901c6deb5$5ce90620$03010b0a@DHB> |
Swish-e gurus, Am I missing (in swish-e 2.4.3) a command line or configuration option to index only files in the specified directory without recursing into subdirectories? I am thinking I will approximate this by adding the following to the configuration file: FileRules directory contains regex /./ I suppose since I'm building this configuration file dynamically and will know the starting path, I should be able to take advantage of "Directory names do *not* have a trailing slash" and use FileRules dirname contains "$path/" ..where $path is replaced by the value of IndexDir. So if the dirname has an extra slash beyond the starting path, it ought to be a subdirectory, right? I think this will be faster, but it's less general-purpose. And there might be issues with \ vs. / as delimiter under Windows? Am I on the right track? Other suggestions? I guess I'll discover what works soon enough. Okay, I left that in "drafts" while finishing up the actual code. I went with the "FileRules dirname" method -- using a forward slash under Windows -- and it did what I want. Still seems like there ought to be an easier way to turn off recursion... maybe something to consider for 3.0? Thanks, Dave (Background for the curious...) If the whole file structure -- thousands of PDFs -- is indexed at once, our shared host runs out of resources (even with -e), so we want to index each folder recursively and merge up back to the root. Merging partial indexes has worked well for me before. For this application, the directory structure is arbitrarily deep and wide; indexable files may exist in any directory. The routine I'm writing will also check the mtimes on the indexes vs. files to be indexed and so will be doing its own version of incremental indexing (yup; I see that coming in 2.5/3.0). -- David Brown [email protected]