Re: [Gcjwebplugin-user] building the CVS code: compilation puzzles

[Joseph E. Sacco, Ph.D.]

Using: 
- gcc version 4.1.0 20050303 (experimental)
- gcjwebplugin -0.3.2.cvs [updated this morning, 4Mar05]
        

* Set up environment
- verify that gcjwebplugin is not installed *anywhere*

- set environment variables:

export gccpath=/usr/local/gcc/4.1.0
export PATH=$gccpath/bin:$PATH
export LD_LIBRARY_PATH=$gccpath/lib
export PKG_CONFIG_PATH=$gccpath/lib/pkgconfig:/usr/lib/pkgconfig

* configure
./configure --prefix=/usr --mandir=/usr/share/man --sysconfdir=/etc \
--with-java=$gccpath/bin/gij \
--with-javac="$gccpath/bin/gcj -C" \
--with-javah=$gccpath/bin/gcjh \
--with-jar=$gccpath/bin/jar

* build and test

        ## ------------- ##
        ## Test results. ##
        ## ------------- ##
         
        All 27 tests were successful.
        PASS: ./testsuite
        ==================
        All 1 tests passed
        ==================
        

* CLASSPATH issue

There is still a classpath issue when running "make check" before
performing "make install" [see attached testsuite.log]. Adding a
classpath entry to "./test" will solve the problem.


-Joseph

========================================================================================

On Thu, 2005-03-03 at 11:08, Joseph E. Sacco, Ph.D. wrote:
> OK...  I am updating gcc via CVS as we speak.  I will report back later.
> 
> -Joseph
> 
> =============================================================================
> On Thu, 2005-03-03 at 10:59, Thomas Fitzsimmons wrote:
> > On Thu, 2005-03-03 at 10:52 -0500, Joseph E. Sacco, Ph.D. wrote:
> > >Will do...
> > >
> > >What about the two failures???
> > >
> > 
> > Yes, they look like actual failures.  We'll see if they happen with GCC
> > HEAD.
> > 
> > Tom
-- 
joseph_sacco[at]comcast[dot]net

testsuite.log
Description: Text Data