[rfc][icedtea-web] fix to ManifestedJar1Test cases

Andrew Azores aazores at redhat.com
Fri Nov 22 13:27:57 PST 2013

On 11/22/2013 09:03 AM, Jiri Vanek wrote:
> Hi!
> This patch is fixing two long-term failing reproducers in 
> ManifestedJar1Test testcase
> The patch to netx itself is discussable  - do we wont to survive two 
> jars, both with specified main class in manifest, and no other clue to 
> main?
> I chose no - the reason is that it can be programmers error, and order 
> can be *to* easily changed. However it may be to strict, and to chose 
> the first one is correct....
> Then the fix to manifestedJar1main2mainNoAppDesc will be different.
> The fix to manifestedJar1nothing2nothingAppDesc is obvious, and I'm 
> surprised it was ever passing before. Now it looks better.
> J.

Allowing only one main JAR seems more correct to me. It shouldn't be 
valid to have more than one main defined, it just doesn't make sense. 
Allowing for this and hoping for the best when we run the application is 
simply asking for trouble IMO. Best to make this an explicit failure (as 
you've done).

For the patch in ResourcesDesc, please indent the "markedMain = jar" 
line. Also,

More then one main jar specified. Thas fatal

isn't quite right ;). "Multiple main JARs specified, refusing to 
continue" perhaps? If you want to stick to the same phrasing, then "More 
than one main JAR specified. This is fatal."

Otherwise I think this is okay.


Andrew A

More information about the distro-pkg-dev mailing list