[rfc][icedtea-web] reprducers for main class specified in manifest instead in application descriptor
Pavel Tisnovsky
ptisnovs at redhat.com
Wed Mar 7 08:48:18 PST 2012
Hi Jiri,
it looks good. I have two questions:
1) in both files MANIFEST.MF it seems that DOS EOLs are used (I can see ^M on the end of the lines). Did you use DOS
files deliberately?
2) probably another test case should be included: two jars, both with manifests, both set as main (it should fail AFAIK)
Cheers,
Pavel
Jiri Vanek wrote:
> 2012-03-05 Jiri Vanek <jvanek at redhat.com>
>
> Added test for main-class in manifest for jnlp
> * Makefile.am: (prepare-reproducers.stamp) fixed manifest handling.
> Till
> now was manifest copied as any other non java file, and so was
> rewritten
> by jar tool
> * tests/jnlp_tests/simple/ManifestedJar2/srcs: secondary jar file which
> should have manifest and so should help ManifestedJar1 with testing
> *
> tests/jnlp_tests/simple/ManifestedJar2/resources/META-INF/MANIFEST.MF:
> manifest for ManifestedJar2.jar
> * tests/jnlp_tests/simple/ManifestedJar1/srcs: main testing jar
> *
> tests/jnlp_tests/simple/ManifestedJar2/resources/META-INF/MANIFEST.MF:
> manifest for ManifestedJar1.jar
> *
> tests/jnlp_tests/simple/ManifestedJar1/testcases/ManifestedJar1Test.java:
> testing class for this reproducers
> * tests/jnlp_tests/simple/ManifestedJar1/resources/: six reproducers
> jnlps
>
> Currently I'm testing:
> * if two jars with manifest specified, none is main and no main
> class, then first one is loaded
> * if one jar with manifest, is not main, and no main class then is
> lunched
> * if one jar with manifest, but not marked as main and no main
> class then is lunched
> * if one jar with manifest, marked as main and no main class then
> is lunched
> * Two jars, both with manifest, First is main, but specified
> mainclass belongs to second one, then second one should be lunched
> * Two jars, both with manifest, seconds is main, no mainclass, then
> the one marked as main is lunched
>
>
> Any idea for more (meaningfull :) ) combinations? (even as another
> patch...)
>
> Best regards, J
>
More information about the distro-pkg-dev
mailing list