To summarize your update, one can says that it's a filesystem accessing method, coupled with a HTML links following algorithm ?
It is good for me, and maybe I'll use that. But maybe someone could say that you loose every virtual directory information set up in the HTTP server. That was not the aim of this modification, though...
De: Ronald Aarts [SMTP:R.G.K.M.Aarts@wb.utwente.nl]
Date: mercredi 1 septembre 1999 14:44
À: Nicolas Huillard
Cc: Multiple recipients of list
Objet: RE: [SWISH-E] Suggestion for recursive file search and bug fix for command line parser
Nicolas Huillard asked:
> What do you mean when you say that your modifications allow to "recurse"
> the files, using the fs method : I use the IndexDir parameter in the conf file,
> that allow to scan files in a whole directory tree.
> Does your method allow to follow HTTP links, or FS symlinks, or define types
> of files to include ?
With IndexDir you process a directory and ALL files in its subtree, of course
limited by the FileRules. With my modifications you start with e.g. the top-level
index.html file and subsequently swish-e will follow local HTTP-links, i.e. any
occurrence of <A HREF="something"> is processed. The FileRules (including
"pathname contains" !) are obeyed to exclude files. In case a user has a
directory /Secret which is not referred to by another document, it is not
indexed. Same is true is some old file like index_19990813.html is lingering
Hope this clarifies things, Ronald Aarts.
Received on Wed Sep 1 05:42:52 1999