Subject: Re: pkg/32867 (pkgsrc/wip/jdk14 patch fails)
To: None <>
From: Chuck Zmudzinski <>
List: netbsd-java
Date: 02/20/2006 22:45:13
>Subject: Re: pkg/32867 (pkgsrc/wip/jdk14 patch fails)
>Date: Sat, 18 Feb 2006 22:49:40 +0000 (UTC)
>Synopsis: pkgsrc/wip/jdk14 patch fails
>State-Changed-From-To: open->closed
>State-Changed-When: Sat, 18 Feb 2006 22:49:39 +0000
>The patch stage was broken on pkgsrc-HEAD in the same way a couple of weeks
>ago.  This was caused by ^M (Carriage Return) characters in the Eyesbeyond
>patchset.  Since the wip/jdk14 package itself didn't change, it must have 
>some change in the pkgsrc infrastructure which caused this failure (I still
>haven't figured out what exactly broke it).  I then committed a simple
>workaround for this: SUBST the ^M characters out of the patchset before
>applying the patches.

I noticed that in the Makefile

>Apparantly this broke it on pkgsrc-2005Q4 systems (which were ok before) in
>turn.  Your patch (add -a to unzip's arguments) fixes it when combined with 
>earlier fix, so now it works on both pkgsrc-2005Q4 and pkgsrc-HEAD.  Thanks 
>the patch!

Glad to be of help to others.

The only other problem I had was during the build - I got a seg fault from 
gcc3 when compiling
work/hotspot/src/share/vm/opto/parse2.cpp until I maxed out the limits such 
as datasize, stackzize, memoryuse, and memorylocked. I tried first just 
increasing the datasize and stacksize, and after tripling datasize and 
doubling stacksize it still didn't work, so I just decided to max out all 
four of those limits and finally it worked. Perhaps a hint about this 
problem that can occur during the build could be displayed at the beginning 
of the build. Increasing the datasize to 262144 according to the 
instructions on the sun-java linux emulation packages was not good enough 
for me. Anyway, thanks for you good work.

Chuck Zmudzinski