Hello again,

sorry to reply to myself, but I wanted to install ant1.7 and found out
that it can't be installed in parallel with ant (1.8); as we might end
up with a situation where certain packages will build correctly with Ant
1.7 and others with 1.8, I'm not sure this is viable.


Any idea (and decision) how to best resolve (quickly) the situation?

Eric


--
To UNSUBSCRIBE, email to debian-java-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 4B8A581F.8030906@Lavar.de">http://lists.debian.org/4B8A581F.8030906@Lavar.de

On Sun, Feb 28, 2010 at 12:38 PM, Eric Lavarde <Eric@lavar.de> wrote:
> Well, just after ant 1.8 has been uploaded, we've got at least 10 FTBFS
> bugs.

But only some of them can be attributed to ant. Maven-invoker-plugin
uses maven (not ant) for building! Some FTBFS are caused by asm3 and
groovy uploads.

> I'm not against upgrading to ant, but I would had prefered to not be obliged
> to work in urgency to fix a problem caused by somebody else's upload.

Such things can always happen. Just replace ant by ant1.7 for a quick fix.

We should think about having 2 versions of ant in Debian.

Cheers,
Torsten


--
To UNSUBSCRIBE, email to debian-java-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: a90bfcf1002280406l6c063f3ekeba8464bf457e8ef@mail.g mail.com">http://lists.debian.org/a90bfcf1002280406l6c063f3ekeba8464bf457e8ef@mail.g mail.com

On Sun, Feb 28, 2010 at 1:06 PM, Torsten Werner <twerner@debian.org> wrote:
> We should think about having 2 versions of ant in Debian.

I mean: having 2 upstream major versions of ant by default and not
just as a quick fix as we have it now. But the unversioned ant should
always point to the latest stable release. What do you think about
that?

Torsten


--
To UNSUBSCRIBE, email to debian-java-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: a90bfcf1002280410r40a9673dq4e15018a6fb2c448@mail.g mail.com">http://lists.debian.org/a90bfcf1002280410r40a9673dq4e15018a6fb2c448@mail.g mail.com

Torsten Werner wrote:

On Sun, Feb 28, 2010 at 1:06 PM, Torsten Werner <twerner@debian.org> wrote:

We should think about having 2 versions of ant in Debian.


I mean: having 2 upstream major versions of ant by default and not
just as a quick fix as we have it now. But the unversioned ant should
always point to the latest stable release. What do you think about
that?

Torsten


If that means being able to install both versions side by side, and
being able to redirect usage of ant using ANT_HOME (cdbs), then fine
with me.


It can be slightly more difficult but please document.

Cheers,
Eric


--
To UNSUBSCRIBE, email to debian-java-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 4B8A9976.20003@Lavar.de">http://lists.debian.org/4B8A9976.20003@Lavar.de

On Sun, Feb 28, 2010 at 5:27 PM, Eric Lavarde <Eric@lavar.de> wrote:
> If that means being able to install both versions side by side, and being
> able to redirect usage of ant using ANT_HOME (cdbs), then fine with me.

No, currently it is just a quick workaround for the dirty bug in gij.
That means you can either build-depend on ant/ant-optional or
ant1.7/ant1.7-optional but you cannot install both at the same time.

Torsten


--
To UNSUBSCRIBE, email to debian-java-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: a90bfcf1002280914h5da71decm7097d0f3ea76f3ae@mail.g mail.com">http://lists.debian.org/a90bfcf1002280914h5da71decm7097d0f3ea76f3ae@mail.g mail.com

reassign 571532 gij
thanks

Please never reassign bugs without giving an explaination. Ant's
upstream developers argue
<https://issues.apache.org/bugzilla/show_bug.cgi?id=48894> that this
is a bug in gij or its libraries. We have a workaround now but the bug
in gij is still there.

Torsten



--
To UNSUBSCRIBE, email to debian-gcc-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: a90bfcf1003140144y6795e250ma3ba7945fd14df6e@mail.g mail.com">http://lists.debian.org/a90bfcf1003140144y6795e250ma3ba7945fd14df6e@mail.g mail.com

severity 571532 normal
tags 571532 + moreinfo
tags 571532 + help
thanks

On 14.03.2010 10:44, Torsten Werner wrote:

reassign 571532 gij
thanks

Please never reassign bugs without giving an explaination.


right, there was no information when it was reassigned to gij.


Ant's upstream developers argue
<https://issues.apache.org/bugzilla/show_bug.cgi?id=48894> that this
is a bug in gij or its libraries. We have a workaround now but the bug
in gij is still there.

Torsten








--
To UNSUBSCRIBE, email to debian-gcc-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: 4B9CDBC6.6050306@debian.org">http://lists.debian.org/4B9CDBC6.6050306@debian.org

On Sun, Mar 14, 2010 at 1:51 PM, Matthias Klose <doko@debian.org> wrote:
> right, there was no information when it was reassigned to gij.

This bug never got reassigned to gij. It was reported against gij in
the first place.

Torsten



--
To UNSUBSCRIBE, email to debian-gcc-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: a90bfcf1003140638n5f0ece33n66962c017c53e3f@mail.gm ail.com">http://lists.debian.org/a90bfcf1003140638n5f0ece33n66962c017c53e3f@mail.gm ail.com

Hi.

The problem is triggered by this line, for any small file,
which seems to me allowed by java specification

srcChannel.transferTo(0, FileUtils.BUF_SIZE,destChannel);

http://java.sun.com/j2se/1.4.2/docs/api/java/nio/channels/FileChannel.html:

An attempt is made to read up to count bytes from the source channel and
write them to this channel's file starting at the given position. An
invocation of this method may or may not transfer all of the requested
bytes; whether or not it does so depends upon the natures and states of
the channels. Fewer than the requested number of bytes will be
transferred
if the source channel has fewer than count bytes remaining, or if the
source channel is non-blocking and has fewer than count bytes
immediately available in its input buffer.

It is implemented in
libjava/gnu/java/nio/channels/FileChannelImpl.java

via smallTransferTo() via map() via mapImpl().

The specification of map() explicitely states:

Many of the details of memory-mapped files are inherently dependent upon
the underlying operating system and are therefore unspecified. The
behavior of this method when the requested region is not completely
contained within this channel's file is unspecified.

The transferTo(0, 8192, ...) is translated into mmap() of 8192
bytes from file of size 1701 on architecture with pagesize 4096.

This system call succeeds, the later acces to 2nd page fails.
Moreover, it looks like transferTo(0, 4096, ...) would write
4096 bytes to destination file from file with size 1701.

IMHO, the implementaion of smallTransferTo() should be fixed.

Petr



--
To UNSUBSCRIBE, email to debian-gcc-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: Pine.LNX.4.62.1003151036030.24151@sci.felk.cvut.cz ">http://lists.debian.org/Pine.LNX.4.62.1003151036030.24151@sci.felk.cvut.cz

Moreover, it looks like transferTo(0, 4096, ...) would write
4096 bytes to destination file from file with size 1701.


It can be shown by this snippet:

-----------------------------------------
import java.io.FileInputStream;
import java.io.FileOutputStream;
import java.nio.channels.FileChannel;

public class Tra {

public static void main(String[] args) throws Exception {
if (args.length != 2) {
System.out.println("arguments: sourcefile destfile");
System.exit(1);
}
FileChannel in = new FileInputStream(args[0]).getChannel(),
out = new FileOutputStream(args[1]).getChannel();

in.transferTo(0, 4096, out);
}
}
---------------------------------------------



Petr



--
To UNSUBSCRIBE, email to debian-gcc-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Archive: Pine.LNX.4.62.1003152001470.24623@sci.felk.cvut.cz ">http://lists.debian.org/Pine.LNX.4.62.1003152001470.24623@sci.felk.cvut.cz