RFR: JDK-8058797: Building with sjavac broken after JDK-8058118

Discussion:

Erik Joelsson

2014-09-19 08:45:34 UTC

In JDK-8058118, a small java tool is compiled very early in the make
process. This compilation cannot use sjavac since it's done with the
boot javac. The fix is simple. Just add DISABLE_SJAVAC := true to the
macro call. I also added a comment documenting this already existing
feature.

Bug: https://bugs.openjdk.java.net/browse/JDK-8058797
Webrev: http://cr.openjdk.java.net/~erikj/8058797/webrev.root.01/

/Erik

David Holmes

2014-09-19 08:56:34 UTC

Permalink

Hi Erik,

Post by Erik Joelsson
In JDK-8058118, a small java tool is compiled very early in the make
process. This compilation cannot use sjavac since it's done with the
boot javac. The fix is simple. Just add DISABLE_SJAVAC := true to the
macro call. I also added a comment documenting this already existing
feature.
Bug: https://bugs.openjdk.java.net/browse/JDK-8058797
Webrev: http://cr.openjdk.java.net/~erikj/8058797/webrev.root.01/

I'm a bit confused about the comments and code in SetupJavaCompilation.
We seem to have a maximum of 16 args; but we've documented 18
possibilities. ???

David

Post by Erik Joelsson
/Erik

Erik Joelsson

2014-09-19 08:59:34 UTC

Permalink

Post by David Holmes
Hi Erik,

I'm a bit confused about the comments and code in
SetupJavaCompilation. We seem to have a maximum of 16 args; but we've
documented 18 possibilities. ???

That's a bit peculiar I admit. The reason is that we never use all the
arguments at once. The maximum number of arguments used in one call is
simply increased as needed. It's just a technicality in the
implementation of the named args.

/Erik

Post by David Holmes
David

Post by Erik Joelsson
/Erik

Magnus Ihse Bursie

2014-09-19 09:41:14 UTC

Permalink

Looks good to me.

/Magnus