tomcat-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From André Warnier>
Subject Re: AJP13 Connector and JKOptions
Date Thu, 19 Feb 2009 08:56:26 GMT
Pete Helgren wrote:
> What do the URL's look like?  Here is an example.  If I right click and 
> get the properties on the "missing" image, I see this (well I'd add a 
> more "real" looking URL but the this mailing list has rejected my last 9 
> attempts as spam...):
Hi again.
I haven't looked into the question of the image URLs very closely, but 
one thing you have to keep in mind is :

- the server, unless you explicitly do something for that, is not going 
to peek /into/ the pages that it serves and go modify links inside of 
them.  It will serve them as they are, period.
In other words, as the pages are on the disk (or are generated by a 
webapp), so will the webserver send them to the browser.
Mull over this until you are really convinced, then go to the next step.

- the actor who is going to interpret a relative link like 
"images/mypic.gif" found inside a page is /the browser/.
The browser knows from which URL the page that it is currently 
rendering, comes from.  E.g. it knows (or thinks it knows) that the 
current page came from, say, "http://thathost:thatport/somedir/abc.html".
Mull again..

- when the browser, in this page, finds a link like "images/mypic.gif", 
it is going to "fill in the missing part", on the base of the current 
page's origin.  In other words, it will take the current page's URL
then remove the trailing "abc.html", and replace that by the relative 
link "images/mypic.gif", thus giving
And then it is going to issue a brand-new request to the server for

- Now if it so happens that this link that the browser is building for 
the image, when issued to the server, does not return what you expect, 
then tough ! It means that your way of putting links in pages, is not 
compatible with your setup.

What I mean is that, in terms of these relative URLs, it is not the 
server that does something special.  The server is not involved.  Only 
the content of your pages, and the browser, are involved.
If you want the server to be involved, and go modify those links in the 
pages /before/ returning them to the browser, then you have to add a 
special filter at the server level, which will read all outgoing pages, 
find the links in them, and change them to something else.
That exists, but it is something you have to /add/, the server won't do 
that on its own.

(1) for the purists : really, the browser is going to connect to 
"thathost:thatport", and then send over that connection a request for 
"/somedir/images/mypic.gif".  But it's easier to explain it as above.

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message