java-gnome-hackers — Hacking on the bindings themselves: design issues and co-ordination

RE: [java-gnome-hackers] more build problems!

[Jeffrey M. <Jef...@Br...>]

> Should read "The build continued."
> 
> I do suspect that there is in fact a glib type called "gint" because I
> just got another error with the same problem, but what the heck, I'm
> throwing caution to the wind...

There is a glib type called gint and it is used throughout the 
entire bindings.  If we do not resolve the base issue you will
probably have numerous problems.  The code you pointed out works
fine on both of my systems.

Can you please send the entire build line for a gdk JNI file?

-Jeff

 NOTE: THIS IS A CONFIDENTIAL COMMUNICATION. This transmission is intended
only for the use of the individuals or entity to which it is addressed. If
you are not the intended recipient, or the person responsible for delivering
the message to the intended recipient, please return or delete it
immediately. Although this e-mail and any attachments are believed to be
free of any virus or other defect, it is the responsibility of the recipient
to ensure that it is virus free and no responsibility is accepted by us for
any loss or damage arising in any way from its unauthorized modification or
use.

Re: [java-gnome-hackers] more build problems!

[Ray A. <ra...@do...>]

> Well, I went ahead and switched the line to read 
> 
> return (jint)gdk_display_get_window_at_pointer(
> 	(GdkDisplay*)display,&(jint)x, &(jint)y);
> 
> The build 

Should read "The build continued."

I do suspect that there is in fact a glib type called "gint" because I
just got another error with the same problem, but what the heck, I'm
throwing caution to the wind...


Ray Auge <ra...@do...> :)
---------------------------------

"Anything is more stable than Windows.  Even a relationship based purely
on sex." - Pamela Auge

Re: [java-gnome-hackers] more build problems!

[Ray A. <ra...@do...>]

On Tue, 2004-03-09 at 09:54, Ray Auge wrote:
> On Tue, 2004-03-09 at 09:46, Ray Auge wrote:
> > I've gotten jhbuild environment going (I've logged into gnome2.6), I've
> > downloaded j-g from cvs. And I still get this error after several
> > attempts to figure out the problem:
> > 
> > ]$ cd <j-g_source_root>
> > ]$ jhbuild run bash
> > ]$ ./genConfigure.in      [ok]
> > ]$ ./genMakefile.in       [ok]
> > ]$ ./configure            [ok]
> > ]$ make
> > ...
> > 
> > ../../gdk/src/jni/org_gnu_gdk_Display.c: In function
> > `Java_org_gnu_gdk_Display_gdk_1display_1get_1window_1at_1pointer':
> > ../../gdk/src/jni/org_gnu_gdk_Display.c:275: invalid lvalue in unary `&'
> > ../../gdk/src/jni/org_gnu_gdk_Display.c:275: invalid lvalue in unary `&'
> > make[2]: *** [../../gdk/src/jni/org_gnu_gdk_Display.o] Error 1
> > make[2]: Leaving directory
> > `/home/Development/java-gnome/java-gnome-cvs/java-gnome/gtk/src'
> > make[1]: *** [distro] Error 2
> > make[1]: Leaving directory
> > `/home/Development/java-gnome/java-gnome-cvs/java-gnome/gtk'
> > make: *** [all] Error 1
> 
> 
> > 
> > Any thoughts?
> > 
> In fact the return looks like this:
> 
> return (jint)gdk_display_get_window_at_pointer(
> 	(GdkDisplay*)display,&(gint)x, &(gint)y);
> 
> Are those "gint"s supposed to be "jint"s???
> 

Well, I went ahead and switched the line to read 

return (jint)gdk_display_get_window_at_pointer(
	(GdkDisplay*)display,&(jint)x, &(jint)y);

The build 
> > 
> > Ray Auge <ra...@do...> :)
> > ---------------------------------
> > 
> > "Anything is more stable than Windows.  Even a relationship based purely
> > on sex." - Pamela Auge
> > 
> > 
> > 
> > -------------------------------------------------------
> > This SF.Net email is sponsored by: IBM Linux Tutorials
> > Free Linux tutorial presented by Daniel Robbins, President and CEO of
> > GenToo technologies. Learn everything from fundamentals to system
> > administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> > _______________________________________________
> > java-gnome-hackers mailing list
> > jav...@li...
> > https://lists.sourceforge.net/lists/listinfo/java-gnome-hackers
> Ray Auge <ra...@do...> :)
> ---------------------------------
> 
> "Anything is more stable than Windows.  Even a relationship based purely
> on sex." - Pamela Auge
> 
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by: IBM Linux Tutorials
> Free Linux tutorial presented by Daniel Robbins, President and CEO of
> GenToo technologies. Learn everything from fundamentals to system
> administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> _______________________________________________
> java-gnome-hackers mailing list
> jav...@li...
> https://lists.sourceforge.net/lists/listinfo/java-gnome-hackers
Ray Auge <ra...@do...> :)
---------------------------------

"Anything is more stable than Windows.  Even a relationship based purely
on sex." - Pamela Auge

Re: [java-gnome-hackers] more build problems!

[Ray A. <ra...@do...>]

On Tue, 2004-03-09 at 09:46, Ray Auge wrote:
> I've gotten jhbuild environment going (I've logged into gnome2.6), I've
> downloaded j-g from cvs. And I still get this error after several
> attempts to figure out the problem:
> 
> ]$ cd <j-g_source_root>
> ]$ jhbuild run bash
> ]$ ./genConfigure.in      [ok]
> ]$ ./genMakefile.in       [ok]
> ]$ ./configure            [ok]
> ]$ make
> ...
> 
> ../../gdk/src/jni/org_gnu_gdk_Display.c: In function
> `Java_org_gnu_gdk_Display_gdk_1display_1get_1window_1at_1pointer':
> ../../gdk/src/jni/org_gnu_gdk_Display.c:275: invalid lvalue in unary `&'
> ../../gdk/src/jni/org_gnu_gdk_Display.c:275: invalid lvalue in unary `&'
> make[2]: *** [../../gdk/src/jni/org_gnu_gdk_Display.o] Error 1
> make[2]: Leaving directory
> `/home/Development/java-gnome/java-gnome-cvs/java-gnome/gtk/src'
> make[1]: *** [distro] Error 2
> make[1]: Leaving directory
> `/home/Development/java-gnome/java-gnome-cvs/java-gnome/gtk'
> make: *** [all] Error 1


> 
> Any thoughts?
> 
In fact the return looks like this:

return (jint)gdk_display_get_window_at_pointer(
	(GdkDisplay*)display,&(gint)x, &(gint)y);

Are those "gint"s supposed to be "jint"s???

> 
> Ray Auge <ra...@do...> :)
> ---------------------------------
> 
> "Anything is more stable than Windows.  Even a relationship based purely
> on sex." - Pamela Auge
> 
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by: IBM Linux Tutorials
> Free Linux tutorial presented by Daniel Robbins, President and CEO of
> GenToo technologies. Learn everything from fundamentals to system
> administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> _______________________________________________
> java-gnome-hackers mailing list
> jav...@li...
> https://lists.sourceforge.net/lists/listinfo/java-gnome-hackers
Ray Auge <ra...@do...> :)
---------------------------------

"Anything is more stable than Windows.  Even a relationship based purely
on sex." - Pamela Auge

[java-gnome-hackers] more build problems!

[Ray A. <ra...@do...>]

I've gotten jhbuild environment going (I've logged into gnome2.6), I've
downloaded j-g from cvs. And I still get this error after several
attempts to figure out the problem:

]$ cd <j-g_source_root>
]$ jhbuild run bash
]$ ./genConfigure.in      [ok]
]$ ./genMakefile.in       [ok]
]$ ./configure            [ok]
]$ make
...

../../gdk/src/jni/org_gnu_gdk_Display.c: In function
`Java_org_gnu_gdk_Display_gdk_1display_1get_1window_1at_1pointer':
../../gdk/src/jni/org_gnu_gdk_Display.c:275: invalid lvalue in unary `&'
../../gdk/src/jni/org_gnu_gdk_Display.c:275: invalid lvalue in unary `&'
make[2]: *** [../../gdk/src/jni/org_gnu_gdk_Display.o] Error 1
make[2]: Leaving directory
`/home/Development/java-gnome/java-gnome-cvs/java-gnome/gtk/src'
make[1]: *** [distro] Error 2
make[1]: Leaving directory
`/home/Development/java-gnome/java-gnome-cvs/java-gnome/gtk'
make: *** [all] Error 1


Any thoughts?


Ray Auge <ra...@do...> :)
---------------------------------

"Anything is more stable than Windows.  Even a relationship based purely
on sex." - Pamela Auge

RE: [java-gnome-hackers] Update on my progress

[Jehanzeb P. <jk...@ho...>]

Thanks a lot for the input. Before I saw your email I was banging my head on 
my desk from all the pain the Gnome failed installations had caused me. Your 
email showed a ray of hope. Unfortunately, even after installing all the 
packages you have listed the configure script for java-gnome failed. I have 
been at it for the last 2 weeks. I have been hogging all the bandwidth of 
the office and to no avail. The only thing coming to my mind is that someone 
on the group who is using a Pentium 4 machine and a working development 
environment to work on java-gnome would use Partimage to make an image of 
his work partition and let me have it via ftp. Its easier for me to download 
one large file rather than 100 small files.

With regards to the discussion going on the mailing list, my first 
impression og the java-gnome project when I looked at the page in July 2003 
was that it could be used by Java developers just like SWT and since GTK+ is 
cross-platform.

I know that we have to keep the code in sync with Gnome, but I must say that 
the last two weeks have taught me that installing the new Gnome libraries is 
NOT a task we should be expecting of all Java developers. Not all Java 
developers are Linux geeks, actually I feel that since Java abstracts a lot 
of the low level stuff from its developers, most Java developers including 
me have a much harder time when it comes to installing C source code 
downloaded from CVS, like I have had to to get java-gnome compiled.

If we don't make it easy for new developers to start using java-gnome, let 
alone learn the API itself, I don't think that this project will get too 
far.

Any input?

- Pathan



>From: Luca De Rugeriis <pie...@li...>
>Reply-To: jav...@li...
>To: j-g hackers <jav...@li...>
>Subject: RE: [java-gnome-hackers] Update on my progress
>Date: Mon, 08 Mar 2004 18:40:19 +0100
>
>Il lun, 2004-03-08 alle 14:04, Jeffrey Morgan ha scritto:
> > > Hello everyone.
> > >
> > > Sorry I have not been able to add any value to the project yet. I am
> > > struggling with updating my Gnome installation. Spent the
> > > whole week trying
> > > to get Garnome to work, then downloaded directly from the
> > > Gnome website only
> > > to find out that real problem was that my RedHat9 partition
> > > was completed
> > > filled up. I created some space on my hard drive and now have
> > > started the
> > > Garnome installation process again. It is very  frustrating
> > > downloading on a
> > > slow connection especially in a country where the electricity
> > > goes out in
> > > the middle of an installation process (happened to me twice
> > > and then I had
> > > to do a clean install).
> > >
> > > I hope that the Garnome installation will be complete and
> > > ready by the end
> > > of the day today. I have downloaded the java-gnome code from
> > > CVS using
> > > Eclipse 3 and it seems to be installing fine with the
> > > exception of the
> > > required Gnome version.
> > >
>Pathan,
>if you are on a slow connection, maybe you'll find useful to know that
>you only need a small subset of the gnome packages to build j-g.
>I'm on a 56k too, and I didn't bother using garnome, (however jhbuild is
>better): it requires _days_ of downloading.
>Instead this is what I have done:
>I've grabbed:
>
>fontconfig-2.2.92
>freetype-2.1.7
>GConf-2.4.0.1
>glib (cvs head)
>gnome-keyring-0.1.4
>gnome-vfs-2.5.8
>gtk+ (cvs head)
>libbonobo-2.5.4
>libglade-2.3.1
>libgnome-2.5.90
>libgnomecanvas-2.5.90
>libgnomeui-2.5.90.1
>libgtkhtml-2.5.0
>ORBit2-2.9.2
>pango (cvs head)
>
>This is the minimum set of packages to actually compile j-g head.
>We're talking about 10 megs of download. (about 1h with a 56k).
>You'll find the above tarballs in the gnome ftp.
>You'll find the remaing ones one the redhat/fedora cds.
>
>Hope this helps.
>--
>Luca De Rugeriis <pie...@li...>
>
>
>
>-------------------------------------------------------
>This SF.Net email is sponsored by: IBM Linux Tutorials
>Free Linux tutorial presented by Daniel Robbins, President and CEO of
>GenToo technologies. Learn everything from fundamentals to system
>administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
>_______________________________________________
>java-gnome-hackers mailing list
>jav...@li...
>https://lists.sourceforge.net/lists/listinfo/java-gnome-hackers

_________________________________________________________________
Protect your PC - get McAfee.com VirusScan Online 
http://clinic.mcafee.com/clinic/ibuy/campaign.asp?cid=3963

RE: [java-gnome-hackers] points against CNI

[Vikram H. <Vik...@Br...>]

On the same lines (with Mark), it will be valuable to keep JNI/JVM
combination alive, or even better have it co-exist with the CNI. Option 2
seems better choice provided (JNI-to-CNI) conversion can be automated. Or
else option 1 can be a good starting point in the migration path to option
2. 

If option 2 works, eventually all the existing code should be available in
both JNI and CNI. Now assuming new development after this point will be done
in CNI. It might be worth to write some scripts to generate JNI from CNI (or
vice-versa), hence supporting both JNI and CNI.

Decommissioning JNI would mean not supporting "JVM runnable java-gnome
code". Question out there is Supporting JVM a priority ? On the other hand,
supporting "both" JNI (to support JVM) and CNI (for smaller size executable
code and speed) can be a daunting task. Unless CNI to JNI or the other way
around automated code generation can be implemented reliably.

To the best of my knowledge, GCJ does not support few libraries, the ones of
concern are RMI or EJB/Corba related. So for example, if we want to build a
Java GUI client in java-gnome that talks to J2EE application server. A GCJ
converted executable might not be able to talk to App Server running in a
JVM. This can be one of the reasons to support running java-gnome projects
in a JVM.

-Vikram

-----Original Message-----
From: Mark Howard [mailto:mh...@ca...]
Sent: Monday, March 08, 2004 12:25 PM
To: jav...@li...
Subject: [java-gnome-hackers] points against CNI


There is a major point against CNI that I think we haven't discussed
before:

Java programmers know and like java tools.

These include ant, eclipse, eclipse profiler plugin, java debuggers...

Most of our users will be java programmers. If they are not then nothing
good will ever be produced with java-gnome in which case all our efforts
would be mostly pointless. 
If we do switch fully to CNI then most of these tools would probably not
be useful. This means that the entry barrier to using java-gnome is
increased even further.

I'm not trying to say we shouldn't switch to CNI, I just think we need
to take longer to decide about it. What do you think?


-- 
  .''`. Mark Howard
 : :' :
 `. `'  http://www.tildemh.com 
   `-   mh...@de... | mh...@ti... | mh...@ca... 


-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
_______________________________________________
java-gnome-hackers mailing list
jav...@li...
https://lists.sourceforge.net/lists/listinfo/java-gnome-hackers

 NOTE: THIS IS A CONFIDENTIAL COMMUNICATION. This transmission is intended
only for the use of the individuals or entity to which it is addressed. If
you are not the intended recipient, or the person responsible for delivering
the message to the intended recipient, please return or delete it
immediately. Although this e-mail and any attachments are believed to be
free of any virus or other defect, it is the responsibility of the recipient
to ensure that it is virus free and no responsibility is accepted by us for
any loss or damage arising in any way from its unauthorized modification or
use.

RE: [java-gnome-hackers] points against CNI

[Jeffrey M. <Jef...@Br...>]

> There is a major point against CNI that I think we haven't discussed
> before:
> 
> Java programmers know and like java tools.
> 
> These include ant, eclipse, eclipse profiler plugin, java debuggers...
> 
> Most of our users will be java programmers. If they are not 
> then nothing
> good will ever be produced with java-gnome in which case all 
> our efforts
> would be mostly pointless. 
> If we do switch fully to CNI then most of these tools would 
> probably not
> be useful. This means that the entry barrier to using java-gnome is
> increased even further.

ant and eclipse will still work but you would not be able to 
use the java debuggers.  You would have to use gdb instead.

> I'm not trying to say we shouldn't switch to CNI, I just think we need
> to take longer to decide about it. What do you think?

You are absolutely right Mark.  This decision will impact the future
of this project.  Essentially what we are doing is trying to answer
the question of what role java-gnome will fill in the future.  We
need to thoroughly debate this prior to any move.

-Jeff






 NOTE: THIS IS A CONFIDENTIAL COMMUNICATION. This transmission is intended
only for the use of the individuals or entity to which it is addressed. If
you are not the intended recipient, or the person responsible for delivering
the message to the intended recipient, please return or delete it
immediately. Although this e-mail and any attachments are believed to be
free of any virus or other defect, it is the responsibility of the recipient
to ensure that it is virus free and no responsibility is accepted by us for
any loss or damage arising in any way from its unauthorized modification or
use.

Re: [java-gnome-hackers] Finally, got gnome build with jhbuild!

[Mark H. <mh...@ca...>]

On Mon, Mar 08, 2004 at 12:12:04PM -0500, Ray Auge wrote:
> I finally got jhbuild finished and tested via startx. I works fine.
> 
> Now what would be the best approach for tackling a build of j-g 2.6.*.
> 
> A) do I simply run the j-g build (./configure; make; make install) from
> inside the jhbuild environment?
Yes. 

> B) should I direct the build to place j-g in the build env. (./configure
> --prefix=/opt/gnome2)
It's up to you. Personally, I don't install java-gnome - I just point
classpath and LD_LIBRARY path to the places in my source tree where the
files have been built.

> C) when running eclipse, should I start it from within the build env,
> then link to the jars (ie jhbuild run eclipse)?
not sure.

-- 
  .''`. Mark Howard
 : :' :
 `. `'  http://www.tildemh.com 
   `-   mh...@de... | mh...@ti... | mh...@ca... 

Re: [java-gnome-hackers] jhbuild problems??

[Mark H. <mh...@ca...>]

On Mon, Mar 08, 2004 at 10:31:18AM -0500, Ray Auge wrote:
> Hello All,
> 
> I know that this is not really the place to ask about jhbuild problems,
> but I'm following through on Mark's offer.
> 
> I've successfully finished the jhbuild with
> 
> #] jhbuild build --no-network
> 
> And after configuring GDM with a session to access it, I simply cannot.

are you running /some/path/jhbuild run gnome-session in your gdm
session? what error do you get?

> 
> Two questions.
> 
> 1) Is it necessary to be able to login to the newly built gnome. (I
> really just want to make sure the build actually works).
no, jhbuild shell will get you to a shell with the environment variables
set. you can then open java-gnome apps in a gnome 2.4 env.
>
> 2) Is it safe to skip these modules
> ['epiphany','gnomemeeting','memprof','mozilla']. I've not done so yet,
Should be ok.


-- 
  .''`. Mark Howard
 : :' :
 `. `'  http://www.tildemh.com 
   `-   mh...@de... | mh...@ti... | mh...@ca... 

Re: [java-gnome-hackers] jhbuild problems??

[Ray A. <ra...@do...>]

IGNORE THIS ONE, sorry, email problems.



On Mon, 2004-03-08 at 10:31, Ray Auge wrote:
> Hello All,
> 
> I know that this is not really the place to ask about jhbuild problems,
> but I'm following through on Mark's offer.
> 
> I've successfully finished the jhbuild with
> 
> #] jhbuild build --no-network
> 
> And after configuring GDM with a session to access it, I simply cannot.
> 
> Two questions.
> 
> 1) Is it necessary to be able to login to the newly built gnome. (I
> really just want to make sure the build actually works).
> 
> 2) Is it safe to skip these modules
> ['epiphany','gnomemeeting','memprof','mozilla']. I've not done so yet,
> but it would cut down my build time in half and I don't really see why
> they are required for the default gnome build.
> 
> All I really want to do is start using the new j-g bindings. Sheesh! ;)
> 
> I've attached my .jhbuildrc file incase I've made an error.
> 
> Thanks for your patience.
-- 
Ray Auge <ra...@do...>

[java-gnome-hackers] jhbuild problems??

[Ray A. <ra...@do...>]

Hello All,

I know that this is not really the place to ask about jhbuild problems,
but I'm following through on Mark's offer.

I've successfully finished the jhbuild with

#] jhbuild build --no-network

And after configuring GDM with a session to access it, I simply cannot.

Two questions.

1) Is it necessary to be able to login to the newly built gnome. (I
really just want to make sure the build actually works).

2) Is it safe to skip these modules
['epiphany','gnomemeeting','memprof','mozilla']. I've not done so yet,
but it would cut down my build time in half and I don't really see why
they are required for the default gnome build.

All I really want to do is start using the new j-g bindings. Sheesh! ;)

I've attached my .jhbuildrc file incase I've made an error.

Thanks for your patience.

-- 
Ray Auge <ra...@do...>

[java-gnome-hackers] Finally, got gnome build with jhbuild!

[Ray A. <ra...@do...>]

I finally got jhbuild finished and tested via startx. I works fine.

Now what would be the best approach for tackling a build of j-g 2.6.*.

A) do I simply run the j-g build (./configure; make; make install) from
inside the jhbuild environment?

B) should I direct the build to place j-g in the build env. (./configure
--prefix=/opt/gnome2)

C) when running eclipse, should I start it from within the build env,
then link to the jars (ie jhbuild run eclipse)?


Thanks for the help.

-- 
Ray Auge <ra...@do...>

RE: [java-gnome-hackers] Update on my progress

[Luca De R. <pie...@li...>]

Il lun, 2004-03-08 alle 14:04, Jeffrey Morgan ha scritto:
> > Hello everyone.
> > 
> > Sorry I have not been able to add any value to the project yet. I am 
> > struggling with updating my Gnome installation. Spent the 
> > whole week trying 
> > to get Garnome to work, then downloaded directly from the 
> > Gnome website only 
> > to find out that real problem was that my RedHat9 partition 
> > was completed 
> > filled up. I created some space on my hard drive and now have 
> > started the 
> > Garnome installation process again. It is very  frustrating 
> > downloading on a 
> > slow connection especially in a country where the electricity 
> > goes out in 
> > the middle of an installation process (happened to me twice 
> > and then I had 
> > to do a clean install).
> > 
> > I hope that the Garnome installation will be complete and 
> > ready by the end 
> > of the day today. I have downloaded the java-gnome code from 
> > CVS using 
> > Eclipse 3 and it seems to be installing fine with the 
> > exception of the 
> > required Gnome version.
> > 
Pathan,
if you are on a slow connection, maybe you'll find useful to know that
you only need a small subset of the gnome packages to build j-g.
I'm on a 56k too, and I didn't bother using garnome, (however jhbuild is
better): it requires _days_ of downloading.
Instead this is what I have done:
I've grabbed:

fontconfig-2.2.92
freetype-2.1.7
GConf-2.4.0.1
glib (cvs head)
gnome-keyring-0.1.4
gnome-vfs-2.5.8
gtk+ (cvs head)
libbonobo-2.5.4
libglade-2.3.1
libgnome-2.5.90
libgnomecanvas-2.5.90
libgnomeui-2.5.90.1
libgtkhtml-2.5.0
ORBit2-2.9.2
pango (cvs head)

This is the minimum set of packages to actually compile j-g head.
We're talking about 10 megs of download. (about 1h with a 56k).
You'll find the above tarballs in the gnome ftp.
You'll find the remaing ones one the redhat/fedora cds.

Hope this helps.
-- 
Luca De Rugeriis <pie...@li...>

[java-gnome-hackers] points against CNI

[Mark H. <mh...@ca...>]

There is a major point against CNI that I think we haven't discussed
before:

Java programmers know and like java tools.

These include ant, eclipse, eclipse profiler plugin, java debuggers...

Most of our users will be java programmers. If they are not then nothing
good will ever be produced with java-gnome in which case all our efforts
would be mostly pointless. 
If we do switch fully to CNI then most of these tools would probably not
be useful. This means that the entry barrier to using java-gnome is
increased even further.

I'm not trying to say we shouldn't switch to CNI, I just think we need
to take longer to decide about it. What do you think?


-- 
  .''`. Mark Howard
 : :' :
 `. `'  http://www.tildemh.com 
   `-   mh...@de... | mh...@ti... | mh...@ca... 

[java-gnome-hackers] migration to CNI

[Jeffrey M. <Jef...@Br...>]

In the next day or two I will be posting an email to the developers list to
start a discussion on the possibility of moving to CNI for our native layer.
I have been busy trying to determine how we can best perform this mirgation.
I believe there are three ways we can make the migration and I am attaching
a small example that demonstrates the fundamentals of each of these three
approaches.  I look forward to your views on each approach.

The attached example has four directories with four implementations of the
same application.  The first directory (jni) demonstrates the example using
a pure JNI approach.  The remaining three directories demonstrate the
three approaches we might take.


Option 1 (cni1)
---------------
The first approach provides a very simple CNI layer that then makes a
call to our existing JNI code.  I am fairly certain that the CNI layer
can be generated.

PROS
----
1) This approach will allow us to maintain both a JNI and a CNI
build.
2) Although it looks like we are still using JNI we are in fact
using CNI due to a flag provided in the Makefile.  We should see
a performance increase in the bindings as a result.
3) If we do decide to move forward then options 2 and/or 3 can be 
performed incrementally.
4) This is the least risk approach.

CONS
----
1) If we do support both a JNI and CNI build it will make our build
process more complicated that it is already.
2) Our bindings will become larger (in size and code) than they are
already.



Option 2 (cni2)
---------------
Option 2 takes the same approach as Option 1 one step further.  It
creates a JNIEnv but instead of making a call to the external JNI C
code it moves the JNI code into the new C++ file.  We might be able
to generate this code but it would not be as straight forward as the
first approach.  As you can see, the JNI code is converted to a C++
JNI implementation.

PROS
----
1) We can completely eliminate the JNI C code and have only one set
of native code.
2) We should see more performance improvements over option 1.
3) The conversion of the code from the current JNI files to the new
CNI files is very simple.

CONS
----
1) The shared object was larger than option 1.
2) We will no longer have JNI capabilities.



Option 3 (cni3)
---------------
The third option takes a pure CNI approach. There is no attempt to 
salvage our current JNI code.

PROS
----
1) This approach yields the smallest shared object of any other
approach
2) We will have a pure CNI native layer.

CONS
----
1) This is the highest risk conversion and will require the most
work on our part.


-Jeff



 <<test.tar.gz>> 






 NOTE: THIS IS A CONFIDENTIAL COMMUNICATION. This transmission is intended
only for the use of the individuals or entity to which it is addressed. If
you are not the intended recipient, or the person responsible for delivering
the message to the intended recipient, please return or delete it
immediately. Although this e-mail and any attachments are believed to be
free of any virus or other defect, it is the responsibility of the recipient
to ensure that it is virus free and no responsibility is accepted by us for
any loss or damage arising in any way from its unauthorized modification or
use.

RE: [java-gnome-hackers] Update on my progress

[Jeffrey M. <Jef...@Br...>]

We look forward to your contributions.  Please let us know
if there is anything we can do.

-Jeff

> Hello everyone.
> 
> Sorry I have not been able to add any value to the project yet. I am 
> struggling with updating my Gnome installation. Spent the 
> whole week trying 
> to get Garnome to work, then downloaded directly from the 
> Gnome website only 
> to find out that real problem was that my RedHat9 partition 
> was completed 
> filled up. I created some space on my hard drive and now have 
> started the 
> Garnome installation process again. It is very  frustrating 
> downloading on a 
> slow connection especially in a country where the electricity 
> goes out in 
> the middle of an installation process (happened to me twice 
> and then I had 
> to do a clean install).
> 
> I hope that the Garnome installation will be complete and 
> ready by the end 
> of the day today. I have downloaded the java-gnome code from 
> CVS using 
> Eclipse 3 and it seems to be installing fine with the 
> exception of the 
> required Gnome version.
> 
> Hope to start working on something tomorrow.
> 
> - Pathan
> 
> _________________________________________________________________
> Add photos to your messages with MSN 8. Get 2 months FREE*. 
> http://join.msn.com/?page=features/featuredemail
> 
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by: IBM Linux Tutorials
> Free Linux tutorial presented by Daniel Robbins, President and CEO of
> GenToo technologies. Learn everything from fundamentals to system
> administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> _______________________________________________
> java-gnome-hackers mailing list
> jav...@li...
> https://lists.sourceforge.net/lists/listinfo/java-gnome-hackers
> 

 NOTE: THIS IS A CONFIDENTIAL COMMUNICATION. This transmission is intended
only for the use of the individuals or entity to which it is addressed. If
you are not the intended recipient, or the person responsible for delivering
the message to the intended recipient, please return or delete it
immediately. Although this e-mail and any attachments are believed to be
free of any virus or other defect, it is the responsibility of the recipient
to ensure that it is virus free and no responsibility is accepted by us for
any loss or damage arising in any way from its unauthorized modification or
use.

[java-gnome-hackers] Update on my progress

[Jehanzeb P. <jk...@ho...>]

Hello everyone.

Sorry I have not been able to add any value to the project yet. I am 
struggling with updating my Gnome installation. Spent the whole week trying 
to get Garnome to work, then downloaded directly from the Gnome website only 
to find out that real problem was that my RedHat9 partition was completed 
filled up. I created some space on my hard drive and now have started the 
Garnome installation process again. It is very  frustrating downloading on a 
slow connection especially in a country where the electricity goes out in 
the middle of an installation process (happened to me twice and then I had 
to do a clean install).

I hope that the Garnome installation will be complete and ready by the end 
of the day today. I have downloaded the java-gnome code from CVS using 
Eclipse 3 and it seems to be installing fine with the exception of the 
required Gnome version.

Hope to start working on something tomorrow.

- Pathan

_________________________________________________________________
Add photos to your messages with MSN 8. Get 2 months FREE*. 
http://join.msn.com/?page=features/featuredemail

Re: [java-gnome-hackers] Re: [Java-gnome-checkins] CVS: java-gnome/glib/src/java/org/gnu/glib CustomEvents.java,1.2,1.3

[Luca De R. <pie...@li...>]

Il ven, 2004-03-05 alle 17:46, Mark Howard ha scritto:
> Fixed in cvs. My compiler didn't give this warning; or if it did it was
> only a warning, not an error.
Strange, it gave me the same error with jikes-1.18 (a bit old),
and with sun javac build 1.4.2_03-b02.
Anyway thanks for the fix :)
-- 
Luca De Rugeriis <pie...@li...>

Re: [java-gnome-hackers] Re: [Java-gnome-checkins] CVS: java-gnome/glib/src/java/org/gnu/glib CustomEvents.java,1.2,1.3

[Mark H. <mh...@ca...>]

Fixed in cvs. My compiler didn't give this warning; or if it did it was
only a warning, not an error.

On Fri, Mar 05, 2004 at 03:26:59AM +0100, Luca De Rugeriis wrote:
> Il gio, 2004-03-04 alle 18:49, Mark Howard ha scritto:
> > Update of /cvsroot/java-gnome/java-gnome/glib/src/java/org/gnu/glib
> > In directory sc8-pr-cvs1.sourceforge.net:/tmp/cvs-serv32314/glib/src/java/org/gnu/glib
> > 
> > Modified Files:
> > 	CustomEvents.java 
> > Log Message:
> > Reverted to old version of CustomEvents with a couple of enhancements:
> >   - wake up the main loop after setting pending (I didn't know about this
> > 	function before - that's why we were using the alternative version)
> >   - Better lock handling in the java code
> 
> Mark, after this commit it fails to compile with:
> 
> ../../glib/src/java/org/gnu/glib/CustomEvents.java:88: variable listCopy
> might not have been initialized
>                         ((Runnable) listCopy.removeFirst()).run();
> 
> Sorry, I couldn't look into it at the moment.
> 
> -- 
> Luca De Rugeriis <pie...@li...>
> 
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by: IBM Linux Tutorials
> Free Linux tutorial presented by Daniel Robbins, President and CEO of
> GenToo technologies. Learn everything from fundamentals to system
> administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> _______________________________________________
> java-gnome-hackers mailing list
> jav...@li...
> https://lists.sourceforge.net/lists/listinfo/java-gnome-hackers
> 

-- 
  .''`. Mark Howard
 : :' :
 `. `'  http://www.tildemh.com 
   `-   mh...@de... | mh...@ti... | mh...@ca... 

Re: [java-gnome-hackers] build problems - plus TODO items

[Luca De R. <pie...@li...>]

Il mar, 2004-03-02 alle 01:03, Jeffrey Morgan ha scritto:
> On Mon, 2004-03-01 at 17:57, Mark Howard wrote:
> > Is there any chance we can get the destroy and delete events done too?
> > 
> > What do you think to adding a 
> > boolean lifeCycleQuery( LifeCycleEvent event ) 
> > method to LifeCycleListener?
> > If this is ok, I can make the changes. 
> > (but I hope you can think of a better method name)
> > 
> > On Mon, Mar 01, 2004 at 04:33:48PM -0500, Jeffrey Morgan wrote:
> > > 1) gcj.jar checks that work for all developers
> > I've added another hack - Does this work for everyone?
> 
> This works for me.
For me too,
sorry if that has created some fuss :)
-- 
Luca De Rugeriis <pie...@li...>

[java-gnome-hackers] Re: [Java-gnome-checkins] CVS: java-gnome/glib/src/java/org/gnu/glib CustomEvents.java,1.2,1.3

[Luca De R. <pie...@li...>]

Il gio, 2004-03-04 alle 18:49, Mark Howard ha scritto:
> Update of /cvsroot/java-gnome/java-gnome/glib/src/java/org/gnu/glib
> In directory sc8-pr-cvs1.sourceforge.net:/tmp/cvs-serv32314/glib/src/java/org/gnu/glib
> 
> Modified Files:
> 	CustomEvents.java 
> Log Message:
> Reverted to old version of CustomEvents with a couple of enhancements:
>   - wake up the main loop after setting pending (I didn't know about this
> 	function before - that's why we were using the alternative version)
>   - Better lock handling in the java code

Mark, after this commit it fails to compile with:

../../glib/src/java/org/gnu/glib/CustomEvents.java:88: variable listCopy
might not have been initialized
                        ((Runnable) listCopy.removeFirst()).run();

Sorry, I couldn't look into it at the moment.

-- 
Luca De Rugeriis <pie...@li...>

[java-gnome-hackers] new release

[Jeffrey M. <Jef...@Br...>]

Mark,

I just put the new release on the sourceforge site and on the gnome ftp
site.
Could you please take the time sometime today to send out the announcement.

Thanks
-Jeff

 NOTE: THIS IS A CONFIDENTIAL COMMUNICATION. This transmission is intended
only for the use of the individuals or entity to which it is addressed. If
you are not the intended recipient, or the person responsible for delivering
the message to the intended recipient, please return or delete it
immediately. Although this e-mail and any attachments are believed to be
free of any virus or other defect, it is the responsibility of the recipient
to ensure that it is virus free and no responsibility is accepted by us for
any loss or damage arising in any way from its unauthorized modification or
use.

RE: [java-gnome-hackers] LifeCycleEvent changes

[Jeffrey M. <Jef...@Br...>]

Some users might use these events to free some sort of
resource.  That is why I believe we need to notify all
of the listeners.

> I thought the idea was that it should exit as soon as it is 
> supposed to.
> TBH, there is no documentation for this signal and I'm a 
> little confused
> about how it is supposed to be used. In particular, I'm not sure how
> destroy-event is used or even if it is relevant to java-gnome. I will
> try to find out more about it.
> 
> On Wed, Mar 03, 2004 at 12:07:36PM -0500, Jeffrey Morgan wrote:
> > Mark,
> > 
> > Shouldn't we loop through all of the listeners before 
> returning the return
> > code?  In 
> > the implementation that you just committed you are 
> returning as soon as a
> > listener
> > returns a code that is different from the default.  We 
> should probably
> > continue through
> > the listeners and return once all have been notified of the 
> event.  What do
> > you
> > think?
> > 
> > -Jeff
> > 
> > 
> > 
> > 
> > 
> > 
> > 
> >  NOTE: THIS IS A CONFIDENTIAL COMMUNICATION. This 
> transmission is intended
> > only for the use of the individuals or entity to which it 
> is addressed. If
> > you are not the intended recipient, or the person 
> responsible for delivering
> > the message to the intended recipient, please return or delete it
> > immediately. Although this e-mail and any attachments are 
> believed to be
> > free of any virus or other defect, it is the responsibility 
> of the recipient
> > to ensure that it is virus free and no responsibility is 
> accepted by us for
> > any loss or damage arising in any way from its unauthorized 
> modification or
> > use.
> > 
> > 
> > -------------------------------------------------------
> > SF.Net is sponsored by: Speed Start Your Linux Apps Now.
> > Build and deploy apps & Web services for Linux with
> > a free DVD software kit from IBM. Click Now!
> > http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click
> > _______________________________________________
> > java-gnome-hackers mailing list
> > jav...@li...
> > https://lists.sourceforge.net/lists/listinfo/java-gnome-hackers
> > 
> 
> -- 
>   .''`. Mark Howard
>  : :' :
>  `. `'  http://www.tildemh.com 
>    `-   mh...@de... | mh...@ti... | mh...@ca... 
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by: IBM Linux Tutorials
> Free Linux tutorial presented by Daniel Robbins, President and CEO of
> GenToo technologies. Learn everything from fundamentals to system
> administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
> _______________________________________________
> java-gnome-hackers mailing list
> jav...@li...
> https://lists.sourceforge.net/lists/listinfo/java-gnome-hackers
> 

 NOTE: THIS IS A CONFIDENTIAL COMMUNICATION. This transmission is intended
only for the use of the individuals or entity to which it is addressed. If
you are not the intended recipient, or the person responsible for delivering
the message to the intended recipient, please return or delete it
immediately. Although this e-mail and any attachments are believed to be
free of any virus or other defect, it is the responsibility of the recipient
to ensure that it is virus free and no responsibility is accepted by us for
any loss or damage arising in any way from its unauthorized modification or
use.

Re: [java-gnome-hackers] LifeCycleEvent changes

[Mark H. <mh...@ca...>]

I thought the idea was that it should exit as soon as it is supposed to.
TBH, there is no documentation for this signal and I'm a little confused
about how it is supposed to be used. In particular, I'm not sure how
destroy-event is used or even if it is relevant to java-gnome. I will
try to find out more about it.

On Wed, Mar 03, 2004 at 12:07:36PM -0500, Jeffrey Morgan wrote:
> Mark,
> 
> Shouldn't we loop through all of the listeners before returning the return
> code?  In 
> the implementation that you just committed you are returning as soon as a
> listener
> returns a code that is different from the default.  We should probably
> continue through
> the listeners and return once all have been notified of the event.  What do
> you
> think?
> 
> -Jeff
> 
> 
> 
> 
> 
> 
> 
>  NOTE: THIS IS A CONFIDENTIAL COMMUNICATION. This transmission is intended
> only for the use of the individuals or entity to which it is addressed. If
> you are not the intended recipient, or the person responsible for delivering
> the message to the intended recipient, please return or delete it
> immediately. Although this e-mail and any attachments are believed to be
> free of any virus or other defect, it is the responsibility of the recipient
> to ensure that it is virus free and no responsibility is accepted by us for
> any loss or damage arising in any way from its unauthorized modification or
> use.
> 
> 
> -------------------------------------------------------
> SF.Net is sponsored by: Speed Start Your Linux Apps Now.
> Build and deploy apps & Web services for Linux with
> a free DVD software kit from IBM. Click Now!
> http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click
> _______________________________________________
> java-gnome-hackers mailing list
> jav...@li...
> https://lists.sourceforge.net/lists/listinfo/java-gnome-hackers
> 

-- 
  .''`. Mark Howard
 : :' :
 `. `'  http://www.tildemh.com 
   `-   mh...@de... | mh...@ti... | mh...@ca...