commit-classpath
[Top][All Lists]
Advanced

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

Re: Reverting changes


From: Mark Wielaard
Subject: Re: Reverting changes
Date: Fri, 02 Apr 2004 19:01:32 +0200

Hi Etienne,

On Mon, 2004-03-29 at 09:29, Etienne Gagnon wrote:
> As you requested, I have reverted all the changes that have caused
> troubles.

Thanks! But you also reverted a lot more that wasn't causing trouble.
Some people did have comments about particular parts but not everything
Grzegorz and you did was bad or causing troubles. I actually think most
of what you wanted to do was good. But some needed some more discussion
and/or be adapted to the be incorporated. The main thing people objected
to was committing without discussing changes first on the mailinglist.

Working together is hard sometimes. I tried to add some guidelines for
working together on GNU Classpath to the Hacker's Guide. Hope that is
helpful. And I do realize that it wasn't properly documented in the
past. Sorry for not having done this earlier.

Lets go through what you reverted and see how we can get these
improvements in at the end.

--- ChangeLog   28 Mar 2004 20:22:07 -0000      1.1966
+++ ChangeLog   29 Mar 2004 07:07:23 -0000      1.1967
@@ -1,167 +1,13 @@
-2004-03-28  Etienne M. Gagnon <address@hidden>
-
-       * HACKING: Updated information about regenerating auto* files and
-       documented the existence of autogen.sh.

That was a good change (in combination with the addition of a working
autogen.sh script of course). I have now moved all info in this file
into the doc/hacking.texinfo to keep all this kind of information in the
same place.

-2004-03-28  Etienne M. Gagnon <address@hidden>
+2004-03-28  Etienne M. Gagnon  <address@hidden>
 
        * doc/vmintegration.texinfo, doc/www.gnu.org/stories.wml: Update
        SableVM related descriptions.

You didn't revert this since it was a good change.
But to update the webpages you need to do an extra dance as described in
doc/www.gnu.org/README. Normally we let Patrik take care of this since
he is more or less or official webmaster. I have taken care of this now
since I was updating the other guide anyway.

BTW. We probably should move the descriptions of the Compilers/Runtime
Environments to stories and just keep a little reference to the tool in
question in the integration guide (or just remove the redundant
information at all from that document).

-2004-03-28  Grzegorz B. Prokopski <address@hidden>
-
-        * Merged vm/reference implementation of these reflection elements,
-        that can be realized in Java:
-        vm/reference/java/lang/reflect/Constructor.java,
-        vm/reference/java/lang/reflect/Field.java,
-        vm/reference/java/lang/reflect/Makefile.am,
-        vm/reference/java/lang/reflect/Method.java,
-        vm/reference/java/lang/reflect/ReflectUtil.java.

Thanks for reverting this. There is code in there that is good and that
I think is a valuable contribution to GNU Classpath. Let see if we can
get answers to the questions raised in
http://mail.gnu.org/archive/html/commit-classpath/2004-03/msg00105.html
And see if we can agree on a revised patch with the other VM-hackers
(like Jeroen) who have some experience with integrating these things.


-2004-03-27  Grzegorz B. Prokopski <address@hidden>
-
-        * configure.ac: Added --with-jvm= handling defaulting to none.
-       * acinclude.m4: Don't build .zip for --with-jvm=sablevm.

I think this one wasn't very useful indeed. It might be nice to try to
come up with something like Michael and I suggested that takes as an
argument the source or installation dir of a particular VM so we don't
have to encode any VM/Compiler specific things in GNU Classpath. On the
other hand if this is only about --disable-zip then it is probably not
worth to spend much time on. Just document that configure option in the
SableVM INSTALL document then.

-2004-03-27  Grzegorz B. Prokopski <address@hidden>
-
-       * java/lang/Object.java (static): Remove static initializer.
-       * java/lang/reflect/Array.java (static): Remove static initializer.
-       * java/lang/Throwable.java (nl): Remove static field.
-       (StaticData): New private static inner class.
-       (stackTraceStringBuffer): Use StaticData.nl.

I liked this patch and Grzegorz fixed the small indentation issues with
it. Could that be recommitted? I can do this if neither of you has time
or trouble reverting a reverted commit.

-2004-03-27  Etienne M. Gagnon <address@hidden>
-
-       * native/fdlibm/dtoa.c, native/fdlibm/e_acos.c,
-       native/fdlibm/e_asin.c, native/fdlibm/e_atan2.c,
-       native/fdlibm/e_exp.c, native/fdlibm/e_fmod.c,
-       native/fdlibm/e_log.c, native/fdlibm/e_pow.c,
-       native/fdlibm/e_rem_pio2.c, native/fdlibm/e_remainder.c,
-       native/fdlibm/e_scalb.c, native/fdlibm/e_sqrt.c,
-       native/fdlibm/fdlibm.h, native/fdlibm/k_cos.c,
-       native/fdlibm/k_rem_pio2.c, native/fdlibm/k_sin.c,
-       native/fdlibm/k_tan.c, native/fdlibm/mprec.c,
-       native/fdlibm/mprec.h, native/fdlibm/s_atan.c,
-       native/fdlibm/s_ceil.c, native/fdlibm/s_copysign.c,
-       native/fdlibm/s_cos.c, native/fdlibm/s_fabs.c,
-       native/fdlibm/s_floor.c, native/fdlibm/s_rint.c,
-       native/fdlibm/s_scalbn.c, native/fdlibm/s_sin.c,
-       native/fdlibm/s_tan.c, native/fdlibm/sf_fabs.c,
-       native/fdlibm/sf_rint.c, native/fdlibm/strtod.c,
-       native/fdlibm/w_acos.c, native/fdlibm/w_asin.c,
-       native/fdlibm/w_atan2.c, native/fdlibm/w_exp.c,
-       native/fdlibm/w_fmod.c, native/fdlibm/w_log.c,
-       native/fdlibm/w_pow.c, native/fdlibm/w_remainder.c,
-       native/fdlibm/w_sqrt.c, native/jni/classpath/jcl.c,
-       native/jni/classpath/jcl.h, native/jni/classpath/jnilink.c,
-       native/jni/classpath/jnilink.h,
-       native/jni/classpath/native_state.c,
-       native/jni/classpath/native_state.h,
-       native/jni/classpath/primlib.c, native/jni/classpath/primlib.h,
-       native/jni/gtk-peer/gdkfont.h,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GdkClasspathFontPeer.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GdkClasspathFontPeerMetrics.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GdkFontMetrics.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GdkGlyphVector.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GdkGraphics.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GdkGraphics2D.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GdkPixbufDecoder.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GtkButtonPeer.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GtkCanvasPeer.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GtkCheckboxMenuItemPeer.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GtkCheckboxPeer.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GtkChoicePeer.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GtkClipboard.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GtkComponentPeer.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GtkEmbeddedWindowPeer.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GtkEvents.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GtkFileDialogPeer.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GtkImagePainter.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GtkLabelPeer.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GtkListPeer.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GtkMainThread.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GtkMenuBarPeer.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GtkMenuComponentPeer.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GtkMenuItemPeer.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GtkMenuPeer.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GtkPanelPeer.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GtkPopupMenuPeer.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GtkScrollBarPeer.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GtkScrollPanePeer.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GtkTextAreaPeer.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GtkTextComponentPeer.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GtkTextFieldPeer.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GtkToolkit.c,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GtkWindowPeer.c,
-       native/jni/gtk-peer/gthread-jni.c, native/jni/gtk-peer/gtkpeer.h,
-       native/jni/java-awt/gnu_java_awt_EmbeddedWindow.c,
-       native/jni/java-io/FileDescriptor.c,
-       native/jni/java-io/java_io_File.c,
-       native/jni/java-io/java_io_ObjectInputStream.c,
-       native/jni/java-io/java_io_VMObjectStreamClass.c,
-       native/jni/java-io/javaio.c, native/jni/java-io/javaio.h,
-       native/jni/java-lang/java_lang_Double.c,
-       native/jni/java-lang/java_lang_Math.c,
-       native/jni/java-lang/java_lang_Object.c,
-       native/jni/java-lang/java_lang_VMDouble.c,
-       native/jni/java-lang/java_lang_VMFloat.c,
-       native/jni/java-lang/java_lang_VMSystem.c,
-       native/jni/java-lang/java_lang_reflect_Array.c,
-       native/jni/java-net/gnu_java_net_PlainDatagramSocketImpl.c,
-       native/jni/java-net/gnu_java_net_PlainSocketImpl.c,
-       native/jni/java-net/java_net_InetAddress.c,
-       native/jni/java-net/java_net_NetworkInterface.c,
-       native/jni/java-net/javanet.c, native/jni/java-net/javanet.h,
-       native/jni/java-nio/gnu_java_nio_NIOServerSocket.c,
-       native/jni/java-nio/java_nio.c,
-       native/jni/java-nio/java_nio_DirectByteBufferImpl.c,
-       native/jni/java-nio/java_nio_FileChannelImpl.c,
-       native/jni/java-nio/java_nio_FileLockImpl.c,
-       native/jni/java-util/java_util_TimeZone.c,
-       native/target/Linux/target_native.h,
-       native/target/Linux/target_native_file.h,
-       native/target/Linux/target_native_io.h,
-       native/target/Linux/target_native_math_float.h,
-       native/target/Linux/target_native_math_int.h,
-       native/target/Linux/target_native_misc.h,
-       native/target/Linux/target_native_network.h,
-       native/target/generic/target_generic.h,
-       native/target/generic/target_generic_file.h,
-       native/target/generic/target_generic_io.h,
-       native/target/generic/target_generic_math_float.h,
-       native/target/generic/target_generic_math_int.h,
-       native/target/generic/target_generic_misc.h,
-       native/target/generic/target_generic_network.h,
-       native/testsuite/guile-jvm.c, native/vmi/vmi.c, native/vmi/vmi.h:
-       Indented using GNU indent.

There was only a question of whether this was bad for native/fdlibm
and/or native/gtk-peer/ The rest was actually fine. Lets wait what
Graydon says about native/gtk-peer/ and then redo the indenting for the
non-contested files.

-2004-03-27  Etienne M. Gagnon <address@hidden>
-
-       * autogen.sh: Useful script for hackers to call auto* tools on
-       fresh CVS copy.
-       * Makefile.am: Only indent C files below native/ directory.

Both these were useful additions in principle. The C indent target has
to be adapted to not run over native/fdlibm, but is fine otherwise.
autogen.sh probably could need some input from Brian, Michael and
Dalibor who have some more experience with the build system. But it was
a good idea in general.

-2004-03-27  Etienne M. Gagnon <address@hidden>
-
-       * .cvsignore: Ignore some additional generated files.
-       * config.guess, config.sub, depcomp, install-sh, ltmain.sh,
-       missing, mkinstalldirs: Removed these generated files.

In combination with the above autogen.sh script fine.
The only criticism was that it wasn't properly announced before it was
committed which caused some people a nasty surprise when they updated.
But now that people know about it, it can go in (or could have stayed
in).

-2004-03-27  Etienne M. Gagnon <address@hidden>
-
-       * .indent.pro: Added typedef file for GNU indent.
-       * Makefile.am: Added indent target.
-       * include/jni.h.in, native/fdlibm/fdlibm.h,
-       native/jni/gtk-peer/gnu_java_awt_peer_gtk_GdkGlyphVector.c,
-       native/vmi/vmi.h: Separate typedef declarations from struct/union
-       declarations.
 
Same as above. Fine except for the native/fdlibm.h one (although that
can probably be pushed upstream if it is an important clarification).

Cheers,

Mark

Attachment: signature.asc
Description: This is a digitally signed message part


reply via email to

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