This is definitely a change that should be made. I've been wanting to
look into it myself for quite a while, but it has never made it far
enough up the list of priorities for me.
James Taylor <[EMAIL PROTECTED]> writes:
> Sounds interesting, however note that 'getRealPath("")' is used to
> determine the application root, which is then used in lots of places
> (including everywhere that calls Turbine.getRealPath(...)) so getting
> turbine to work from a WAR may require more than just that change.
>
> However I'm amicable to making the change, does anybody have a good
> reason not to? Obviously this would mean certain defaults (logging to
> the application root) would not work when using a WAR.
>
> -- jt
>
> On Tue, 2002-05-28 at 05:08, Brian Topping wrote:
> > Hiyas,
>>
>> I'm just getting started using Turbine on JBoss 3.0 with an eye toward deploying
>Jetspeed there.
>>
>> The first problem I have run into is that Turbine always tries to open static files
>from the webapp directly from an exploded webapp directory. This works okay under
>standalone Tomcat since it uses an instantiation of FileDirContext, but when embedded
>in JBoss, Tomcat uses a WARDirContext. If I have my head screwed on straight, this
>means reliance on ServletContext.getRealPath() is incorrect except for portability
>with Servlet specs prior to 2.1.
>>
>> Is there any reason that these calls shouldn't be updated to Servlet.getResource()?
> I presume the calls to getRealPath are in there from the early days of Turbine.
>I'll go ahead and generate the diffs if this is reasonable and someone wants to merge
>them when I am done. If not, maybe someone could clue me in on another path around
>the problem of loading resources out of the WAR file...
--
To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
