[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Classpathx-discuss] ANT and JUNIT issues (automake???)

From: David Brownell
Subject: Re: [Classpathx-discuss] ANT and JUNIT issues (automake???)
Date: Wed, 05 Dec 2001 15:02:03 -0800

The last version I tried was from CVS, it says 1109, but
all the earlier versions I used had the same failure mode.

The simplest way I know right now to reproduce this is
to get the gnujaxp 1.0beta1 and "./configure --with-gcj",
then try to build it.  Note that this compiles to class files,
not to native code.  I seem to recall the classfiles wouldn't
validate with Sun's JVM even when I used some different
compilation commands to make it behave.

It took me a while to get convinced the root problem was
the "@filelist" stuff, and then unfortunately I had to finish
up some other work so I haven't gotten back to that issue
for a while.

- Dave

p.s. Yes, I'd gladly take a new autoconf/automake
    setup that even had a "build to shared-library"
    option ... I'm relatively sure the compiler is
    generating correct enough ".o", at least for the
    SAX code if not DOM.  It's just that supporting
    all the different compiler invocation syntaxes
    is a PITA!

----- Original Message ----- 
From: "Tom Tromey" <address@hidden>
To: "David Brownell" <address@hidden>
Cc: "Nic Ferrier" <address@hidden>; <address@hidden>
Sent: Wednesday, December 05, 2001 2:18 PM
Subject: Re: [Classpathx-discuss] ANT and JUNIT issues (automake???)

> >>>>> "David" == David Brownell <address@hidden> writes:
> David> Well, GCJ had been segfaulting with the @filelist syntax and
> David> not when I give the file names on the command line, so ...
> David> Maybe a more recent snapshot wouldn't have that issue,
> David> but it's been a longstanding issue.  Haven't had time to
> David> look at it.  Try configuring GNUJAXP "--with-gcj" and
> David> see if it happens for you too.
> What version of gcj are you using?
> Maybe there is a bug in the @filelist handling.
> We'd like to hear about it.  Could you send a stack trace, and/or a
> simple way to reproduce the problem?
> Thanks,
> Tom

reply via email to

[Prev in Thread] Current Thread [Next in Thread]