Dear Mark,
* native/jni/classpath/jcl.h, native/jni/classpath/jcl.c,
native/jni/java-io/java_io_VMFile.c:
Inserted native layer macros for OS functions, added parameter for old
memory size to JCL_realloc() to be able to make an malloc() on system
where no
Hi,
gnu_java_net_PlainDatagramSocketImpl.c seems to contain a small error.
Please find a patch below. I also changed this into thiz for
C++-compatibility in gnu_java_net_PlainDatagramSocketImpl.c and
gnu_java_net_PlainSocketImpl.c .
Bye,
Torsten
--- gnu_java_net_PlainDatagramSocketImpl.c
Dear Classpath-members,
after my last update of the Classpath sources I detected some parameters
of functions which are unused and are marked by
__attribute__ ((__unused__)
Unfortunatelly this is not portable to some non-GNU compilers (as I
know... at least I have one C-compiler for ARM which
Dear Mark,
Agreed. Could you supply a patch that cleans up the current code to
separate the 8.3 thing and the file separator thing to explicitly
mention something like fileSystem8_3 and introduces a system property
that the runtime should set when such a file system is set?
Attached to this email
Dear Classpath members,
some time ago there was a discussion how to support systems with a
filesystem which are limited to 8+3-filenames (e. g. DOS-filesystem).
Not so long ago there were also some messages concerning this and the
introduction of a constant like ON_WINDOWS or isWin32. Now I
Dear Michael,
Wouldnt it be better to call it posix (note: no captial letter) as
it currently supports Linux, *BSD, AIX ... We can and should add Solaris
too. With autoconf its really simple. Then AICAS can remove their own
Solaris port totally and we have a clean non-redundant codebase for
Dear Tom
I've rebuilt all the include/*.h files with a newer gcjh.
This one includes JNICALL and JNIEXPORT in the output.
Thanks.
Is the filename java_nio_FileChannelImpl.c correct? I think
it sould be java_nio_channels_FileChannelImpl.c (matching the
name of the header file).
Sincerely,
Dear Classpath-members,
I yust made some changes in java_nio_FileChannelImpl.c: it seemed it was
an old implementation which does not match to the current implementation
of FileChannelImpl.java. I hope this is ok.
Sincerely,
Torsten
___
Classpath
Dear Stephen,
I don't know if you did this deliberately, but your latest change
to java/net/PlainSocketImpl.java to implement the connect(addr,timeout)
method has clobbered an existing implementation checked in a few
hours earlier by Michael Koch on my behalf.
The CVS log http://savannah.gnu.org
Dear Classpath members,
I yust added a fix to java.io.File.createTempFile() to limit the
filename to 8+3 characters, such that it also runs on DOS-filesystems.
To create a temporary file with an unique name I also limited the
prefix to max. 7 characters to be able to create a filename for a
Dear Brian,
I think you can call _javanet_set_option for the datagram case in
_javanet_create(), however you need to add the SO_BROADCAST case and
link up the setBroadcast to it in Java.
I added such a call in my copy here. I will commit the changes soon.
Sincerely,
Torsten
Dear Mark,
The 1.4 API doc for java.net.SocketOptions indicates that SO_BROADCAST
is set by default for DatagramSockets.
This also seems to be what libgcj does.
Dear Classpath members,
it seems the behaviour of the Classpath implementaiton of datagram
sockets and sending broadcast messages is different from the
implementation in the Sun JDK API. I detected that the Sun
implementation set the socket option SO_BROADCAST for a datagram
socket by default.
Dear Classpath members,
I added some files with the names
class-dependencies.conf
in some Java package directories. These files describe dependencies
of class/methods/fields as a property file. The data can be used if
a VM tool want to link classes to an static linked application,
like our
Dear Classpath members,
I yust check-in the second part of the native layer in:
native/jni/nio
native/jni/java-net
native/jni/java-util
Now most of the modifications are done now, but some fine
tuning and of course bug-fixes (if there are bugs) still
have to be done. I will do this as soon as
Dear Classpath members,
I checked-in some files with an abstraction layer for native
OS functions from our Jamaica VM to be able to port the Classpath
C sources to other OS than Unix. This is my first major change in
Classpath and there could still be some errors or not fixed things.
I will try
Hi,
I yust got a little bit confused, because of the behaviour of
java.io.File.isDirectory()/java.io.File.isFile(). If the
requested filename is a link to a file/directory it seems
always false is reported. With JDK true is reported. The
reason for this seems to be the OS function lstat which
is
Dear Classpath members,
already a (too) long time ago I added the files in native/target
as a proposal for a native function layer to make the C sources
independent from a specific OS/machine. This code is still not
active and because of many other things I had to do upto now
I was still not able
18 matches
Mail list logo