Hi John,

You're right! I had thought about doing it this way but discarded it because
it was just an extra step. Btu you're completely right, it makes the build
more predictable and this is a good thing. Thanks for convincing me! :-)

I was trying to think about other uses case but I was not able to come up
with any good one ;-) Let's wait till there's one...

-Vincent

> -----Original Message-----
> From: John Casey [mailto:[EMAIL PROTECTED]
> Sent: mardi 12 avril 2005 17:16
> To: Maven Developers List
> Subject: Re: [M2] Need for 2 test resource dirs?
> 
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> This is just my 2 cents, but I've always found it pretty easy to do
> something similar to:
> 
> ClassLoader cloader = Thread.currentThread().getContextClassLoader();
> URL resource = cloader.getResource("/classpath/relative/path");
> 
> File theFileIWantToUse = new File(resource.getPath());
> 
> Since you know that path to be in the test resources, it shouldn't have
> a problem resolving it to a file when you need it...
> 
> Is there some reason for explicitly excluding some resources from the
> classpath (are they getting in the way of legitimate test/class
> resources)?
> 
> As for adding a convenience method to retrieve a file from
> external-resources, I'd be a little worried about tethering my test
> classes to maven in this case. I've often seen the case where a test
> expects to be run from a particular ${user.dir} and will break if it
> isn't, due to relative pathing problems. Additionally, I've seen the
> case where running a test in an IDE results in a different ${user.dir}
> being used than what maven might do. So there is a question of
> consistency if you're going to offer that type of method or whatever to
> help tests resolve those files.
> 
> I'd also be a little hesitant to offer this type of facility through
> surefire, because then you have to approximate it in your IDE if you
> want to run the test from there...or any other test harness which would
> otherwise be capable of running that test, for that matter.
> 
> Again, this is just some food for thought. If you have a use case where
> the aforementioned 3 lines of code wouldn't suffice, can you explain it
> a little more? I'm interested in covering the bases here, but don't want
> to cross the line into code that's unpredictable if we can avoid that.
> 
> Thanks,
> 
> john
> 
> Vincent Massol wrote:
> > Hi,
> >
> > I'm trying to convert the Cargo build to m2. I have some junit tests
> that
> > require external resources like WARs, EARs. I have initially put them in
> > src/test/resources but I'm wondering if this is the right place.
> >
> > Indeed, resources put in src/test/resources will get copied to the test
> > execution classpath which is not required for my junit tests. They need
> a
> > reference to the File objects.
> >
> > Thus I'm wondering if there isn't the need for 2 standard test resource
> > dirs:
> >
> > 1/ src/test/resources
> > 2/ src/test/resources-external (or some other name)
> >
> > 1/ would get copied to the test CP while 2/ would not.
> >
> > Also, we could add a default system property in the surefire plugin so
> that
> > tests can easily get a reference object pointing to
> > src/test/resources-external.
> >
> > What's m2 developer's view on this?
> >
> > Thanks
> > -Vincent
> >
> >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: [EMAIL PROTECTED]
> > For additional commands, e-mail: [EMAIL PROTECTED]
> >
> >
> >
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.2.6 (GNU/Linux)
> 
> iD8DBQFCW+YtK3h2CZwO/4URAkAKAKCcax9u7Z92vO81I8hK1lm1oF0sBQCgpE/1
> orhZ7bF/6EvDBRsJ1AjZAVw=
> =0rvF
> -----END PGP SIGNATURE-----
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]





---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to