Thanks Sato-san. The fix supports a larger number of sites.
As you pointed out dynamically generated URLs are at the heart of the
problem. Unfortunately, increasingly many sites are moving towards this
direction. Take for instance the popular Google search engine.
For what I need it does the job although I don't really need to have all
the links rewritten. All I need in my context is to have every url
activation to carry a constant prefix, e.g., http://delegate..../
Again thank you very much for your help and regards, -Ng
From: Yutaka Sato [mailto:firstname.lastname@example.org]
Sent: Wednesday, June 18, 2003 5:09 AM
Subject: Re: [DeleGate-En] HTTP mediation
On 05/27/03(23:46) you "Ng S. T. Chong" <email@example.com> wrote in
|Further to my colleague's message, I found the following error message
|in the log while visiting www.microsft.com: | |... [27/May/2003:22:57:56
+0900] "GET |http://library/flyoutmenu/default.htc HTTP/1.1" 500 0
0*0.000+0.000:C:2+ | |I started up DELEGATE with the following
configuration: | |./delegated SERVER=http -P80 PERMIT="*:*:*" MOUNT="/*
http://*" -v -f | |It appears that DELEGATE cannot handle relative links.
I would |appreciate it if you could shed some light on the problem we have
|Similar problems can be found with the following two sites: |www.sgi.com
Thanks for showing the testable examples. To realize the mapping of URL,
DeleGate rewrites URLs appear in fields in HTTP header or tags in HTML body.
Other URLs appear in other contexts are not cared, and will break the
mapping if included.
Since it is difficult or impossible to completely detect where URLs are
cared. But from my technical interest, I made a experiment to apply MOUNT
for these sites.
As long as I noticed, these sites includes URLs in
- style sheet in text/css response
or in STYLE tag as <STYLE>...</STYLE>
or in STYLE attribute as <tag STYLE=...>
or in SCRIPT tag as <SCRIPT>...</SCRIPT>
or in attribute value
- value attribute in OPTION tag
I feel rewriting URLs in style sheets might be supported, but in other
context should not or should be optional.
Anyway I uploaded the experiment version to
@ @ Yutaka Sato <firstname.lastname@example.org> http://www.delegate.org/y.sato/ ( -
) National Institute of Advanced Industrial Science and Technology (AIST)
_< >_ 1-1-4 Umezono, Tsukuba, Ibaraki, 305-8568 Japan
Do the more with the less -- B. Fuller