ant-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Steve Loughran" <stev...@iseran.com>
Subject Re: compile jsp (request)
Date Wed, 21 Mar 2001 22:22:53 GMT

----- Original Message -----
From: "Di Iorio, Matthew (ISS Atlanta)" <MDiIorio@iss.net>
To: <ant-dev@jakarta.apache.org>
Sent: Wednesday, March 21, 2001 08:09
Subject: RE: compile jsp (request)


> >Now there is an interesting task for someone: a URL spider that calls all
> >JSP/ASP/whatever files it finds beneath a URL, fails if one or more of
the
> >files returns an error or is just plain missing. Of course, detecting an
> >erroneous compilation only works if the servers return an HTTP error
code,
> >or there is an error pattern to look for...
>
> That would be nice.  It could be part of the get task or inherit from the
> get task.  The question is how does the spider know which files to get?
> There is no way to get a directory listing of available files unless the
web
> server is setup to allow that.  One possibility is to give it a directory
> and say "for all the filenames in this directory, make a request from this
> root URL for those filenames".

it could use the get task, but it shouldnt be part of it because it is so
very very different.

1. You can actually do it today if you have a cgi-bin/servlet which can
takes a URL parameter and does the walk; then you have a get task to fetch
'http://search/cgi-bin/walk?url=http://server2/base

2. There is an HTTP extension to list all files in a subdirectory, even
those hidden by an index page. PROPFIND of WebDAV does it. But the directory
mapping trick would work for all servers -and even pick up on situations
where the PC dev box has got the case wrong and a file isnt findable on the
unix server

-stve

-steve




Mime
View raw message