Re: Bootstrap failure in 4.4.1

[David Ronis <ronis at ronispc dot chem dot mcgill dot ca>]

gjar is in /usr/bin, and gives:

gjar --version
jar (GNU Classpath) 0.98

Isn't this built as part of gcc?  It has a similar timestamp as my
installed 4.4.0 gcc.  Finally, shouldn't a bootstrap build use its own
gjar?  One seems to have been built:

find -name gjar -ls
3503030    4 -rwxr-xr-x   1 ronis    ronis        2048 Jul 22
23:57 ./i686-pc-linux-gnu/libjava/classpath/tools/gjar

On the other hand, trying to run this version gives:

 ./gjar --version
./gjar: line 48: /usr/share/classpath/tools.zip: No such file or
directory
./gjar: line 48: exec: /usr/share/classpath/tools.zip: cannot execute:
No such file or directory

(In fact, there is no /usr/share/classpath directory).

David

On Thu, 2009-07-23 at 18:20 +0200, Andrew Haley wrote:
> On 07/23/2009 06:17 PM, David Ronis wrote:
> > I was building gcc-4.4.1 on an i686-linux-gnu slackware-12.2 box with
> > the default optimization flags.  I'd last built/installed 4.4.0.  The
> > current build dies at:
> 
> > here=`pwd`; cd ../../../gcc/libjava/classpath/lib; \
> > 	find gnu java javax org sun -name .svn -prune -o -name '*.class' -print
> > | \
> > 	gjar -cfM@ $here/libgcj-4.4.1.jar
> > jar: internal error:
> > java.lang.NullPointerException
> >    at
> > gnu.classpath.tools.jar.Creator.writeCommandLineEntries(libgcj-tools.so.10)
> >    at gnu.classpath.tools.jar.Creator.run(libgcj-tools.so.10)
> >    at gnu.classpath.tools.jar.Main.run(libgcj-tools.so.10)
> >    at gnu.classpath.tools.jar.Main.main(libgcj-tools.so.10)
> 
> OK, so the gjar in your path is hosed.  Find it, and get back to use,
> and we can try to debug it.
> 
> Andrew.
>