ant-contrib-developers

[Ant-contrib-developers] Support for specifying path to compiler executable

[Dave T. <dav...@gm...>]

Hi all -

   Is there current support in the cc or compiler tags for specifying the
exact path (including binary name) of the compiler binary to use?  Usually
this will just be the name of the compiler flavore (eg. 'gcc'), but I've
run into one use case already where this will not work - namely using
Netbeans on Cygwin/Windows and trying to start a g++ instance through
ant/cpptasks:  Cygwin installs g++ as a symlink, which works fine from the
Cygwin command line, but cpptasks calls the Win32 CreateProcess() which
doesn't know anything about Cygwin and fails.  It would be great to be able
to say <cc ccpath="/usr/bin/gcc++-3.exe"/> or something like that.

   Workaround involves parsing the output of an 'ls' command - less than
savory.

   -David

[Ant-contrib-developers] cpptasks -- outputting actual gcc command

[JonathonS <the...@gm...>]

Hi all,

I was wondering if there is a way to output the actual gcc command
that is used when compiling source files.  If I enable verbose on Ant
using the "-v" flag, it still does not display the actual gcc command
used.  Note that -v does however output the gcc *linker* command-line,
but this does not help me.

Thanks,
J

[Ant-contrib-developers] [Help] RE: Exception on running client xml

[SourceForge.net <no...@so...>]

The following forum message was posted by  at http://sourceforge.net/projects/ant-contrib/forums/forum/113701/topic/2066721:

Hi, did you manage to solve this problem ?  Please let me know. 

[Ant-contrib-developers] [Help] RE: for list task eats multiple dollar signs

[SourceForge.net <no...@so...>]

The following forum message was posted by chris_holman at http://sourceforge.net/projects/ant-contrib/forums/forum/113701/topic/4660869:

I've done some more digging and it seems that this is an ant "feature":

[code]<property name="lineOfTestLine" value="$ $$ $$$ $$$$ $$$$$ $$$$$$" />
<echo message="${lineOfTestLine}" />
[/code] 

Results in:

[code][echo] $ $ $$ $$ $$$ $$$
[/code]

I've posted in the ant user forum.

[Ant-contrib-developers] [Help] for list task eats multiple dollar signs

[SourceForge.net <no...@so...>]

The following forum message was posted by chris_holman at http://sourceforge.net/projects/ant-contrib/forums/forum/113701/topic/4660869:

This script:


[code]		<for delimiter="${line.separator}" param="lineOfTestLine" list="$$$
$$$ ${line.separator}$$$$$$ $$$$ ${line.separator}$$$$$$ $$$$">
			<sequential>
				<echo level="info" message="lineOfTestLine: @{lineOfTestLine}" />
			</sequential>
		</for>
[/code]

Results in this output:


[code]     [echo] lineOfTestLine: $ $ 
     [echo] lineOfTestLine: $$ $ 
     [echo] lineOfTestLine: $$ $
[/code]

Anyone know why its eating all those dollar signs?

Regards,
Chris

Re: [Ant-contrib-developers] To determine Linux Kernel version using cpptasks

[Alex B. <ab...@ca...>]

Hi Varun,

The only thing I know of is using the os.version property.

Alexander Barker
Senior Programmer
California Lutheran University
Information Systems and Services
60 West Olsen Road
Thousand Oaks, CA 91360
805.493.3912


On 07/28/2011 03:33 AM, Varun Thacker wrote:
> Hi,
>
> I would like to know if there is a way to find out the specific kernel
> of a OS using cpptasks. Something similar to a shell "uname -r" command.
>
> -- 
>
>
> Regards,
> Varun Thacker
> http://varunthacker.wordpress.com
>
>

[Ant-contrib-developers] To determine Linux Kernel version using cpptasks

[Varun T. <var...@gm...>]

Hi,

I would like to know if there is a way to find out the specific kernel of a
OS using cpptasks. Something similar to a shell "uname -r" command.

-- 


Regards,
Varun Thacker
http://varunthacker.wordpress.com

Re: [Ant-contrib-developers] Status

[Stefan B. <ste...@fr...>]

On 2011-07-08, Matthew Inger wrote:

> Then add into the mix, supporting Java 1.5 and Java 1.6.

I'm completely on the fence here, I know I've had a customer not too
long ago where Java5 would have been too modern, but they'd still have
the choice of using an older version of ant-contrib (if at all).

Is there anything in ant-contrib that requires Java5 or Java6 right now?
If not, why would you want to introduce it for a 1.0 release right now?
Wouldn't it make sense to have a stable release and break compatibility
after that?

Having said that, I don't really care as I haven't really been active
around here for years now (has anybody?) and don't even use ant-contrib
myself.  Count me in when you need a hand for the release and I'm sure
Java5 and Ant 1.8.x would be great for most users.

> I'm willing to get everything working on Ant 1.8.2 (the most recent
> version), and put out a final release. However, it will require
> specific versions of Ant, JDK and Ivy.

Sounds OK.

> As those other products get updated, newer Ant-Contrib would support
> only the latest versions of those.

Like I said previously I don't think there has been a backwards
incompatible change in Ant that affected ant-contrib - Ivy has certainly
changed when it joined the ASF (package names changed).  If you need any
help to keep Ant itself in check, just yell.

Stefan

Re: [Ant-contrib-developers] Status

[Nick W. <nic...@ni...>]

I may ruffle some feathers here, but I'm going to be frank because I think ultimately it makes for better communication.

I don't think any *responsible* (yes that might be asking a lot) enterprise should still be using Java 1.4 or Java 5. Java 1.4 was end of life three years ago and Java 5 was end of life two years ago, and they both have security vulnerabilities and stability issues that Sun/Oracle considers critical. Furthermore, Java 6 came out five years ago, and Java 7 comes out in just 20 days. Could you imagine a company still using Windows 95, Windows 98SE or Windows NT and then calling up their software vendor and asking why the 2011-year software they just bought doesn't work on their antiquated, no-longer-supported operating system? They'd get laughed at. It's simply not reasonable, just like it's not reasonable for users of Java software written today to expect that software to run on Java 1.4 or Java 5. We recently told 200 customers that if they want to upgrade from the 2007 version of our software to the 2010 version, they were going to have to upgrade from Java 1.4 to Java 6. They complained a little bit, but they didn't decide to stop using our software. They upgraded.

People who care about using ant-contrib will still use ant-contrib if you end Java 1.4 and Java 5 support.

I know that writing open source software may not be a business most people are in to make money. But there's still a measure of success in the open source world, just like there's a measure of success in the corporate world. I see that success as 1) regularly releasing new, stable versions of your product that provide new features and take advantage of new technologies, and 2) having a product that people want to use and come download from you so that they can do so. Just like a business has to make decisions to achieve their measure of success, so must open source communities. I think supporting only Java 6 and 7 and the latest versions of Ant and Ivy is the right business decision to make to keep ant-contrib successful, and I support that decision.

And let's keep one very important fact in mind: If someone, for some strange reason, needs to build an old (or new) version of a product that must be built on Java 1.4, they can still do so in Ant even if Ant is running in Java 6. We do.

Just my two cents.

Nick

On Jul 7, 2011, at 11:01 PM, Matthew Inger wrote:

> A few people have asked the status of the ant-contrib project.
> 
> I have not had much time to work on it recently.  Also,  I got kind of frustrated by the constant API changes
> in Ant itself which kept breaking various tasks.  As a result, I had to maintain several different versions of the
> code just to support different Ant versions.
> 
> On top of that, we have 1 task that relies on Ivy as well, and that underwent version changes.
> 
> Then add into the mix, supporting Java 1.5 and Java 1.6.
> 
> It all because too many variables.
> 
> I'm willing to get everything working on Ant 1.8.2 (the most recent version), and put out a final
> release.  However, it will require specific versions of Ant, JDK and Ivy.  As those other products
> get updated, newer Ant-Contrib would support only the latest versions of those.  That's not to say
> ant-contrib would not be backward compatible, but that depends on the API changes of it's dependent
> products.
> 
> Thoughts?
>  
> ----
> mat...@ya...
> "Once you start down the dark path, forever will it
> dominate your destiny. Consume you it will " - Yoda
> ------------------------------------------------------------------------------
> All of the data generated in your IT infrastructure is seriously valuable.
> Why? It contains a definitive record of application performance, security 
> threats, fraudulent activity, and more. Splunk takes this data and makes 
> sense of it. IT sense. And common sense.
> http://p.sf.net/sfu/splunk-d2d-c2_______________________________________________
> Ant-contrib-developers mailing list
> Ant...@li...
> https://lists.sourceforge.net/lists/listinfo/ant-contrib-developers

Re: [Ant-contrib-developers] Status

[Dale A. <da...@gr...>]

I agree with picking a minimum version of Ant and Java. Ant 1.8.x is good.
I'd try to stay with Java 1.5 just because there are still a lot of
enterprise build systems using it even though it hit end of life quite a
while ago.

I also haven't had much time to help out, but let me know what I can do to
help get a solid release out.

Dale



On Thu, Jul 7, 2011 at 10:01 PM, Matthew Inger <mat...@ya...> wrote:

> A few people have asked the status of the ant-contrib project.
>
> I have not had much time to work on it recently.  Also,  I got kind of
> frustrated by the constant API changes
> in Ant itself which kept breaking various tasks.  As a result, I had to
> maintain several different versions of the
> code just to support different Ant versions.
>
> On top of that, we have 1 task that relies on Ivy as well, and that
> underwent version changes.
>
> Then add into the mix, supporting Java 1.5 and Java 1.6.
>
> It all because too many variables.
>
> I'm willing to get everything working on Ant 1.8.2 (the most recent
> version), and put out a final
> release.  However, it will require specific versions of Ant, JDK and Ivy.
>  As those other products
> get updated, newer Ant-Contrib would support only the latest versions of
> those.  That's not to say
> ant-contrib would not be backward compatible, but that depends on the API
> changes of it's dependent
> products.
>
> Thoughts?
>
> ----
> mat...@ya...
> "Once you start down the dark path, forever will it
> dominate your destiny. Consume you it will " - Yoda
>
>
> ------------------------------------------------------------------------------
> All of the data generated in your IT infrastructure is seriously valuable.
> Why? It contains a definitive record of application performance, security
> threats, fraudulent activity, and more. Splunk takes this data and makes
> sense of it. IT sense. And common sense.
> http://p.sf.net/sfu/splunk-d2d-c2
> _______________________________________________
> Ant-contrib-developers mailing list
> Ant...@li...
> https://lists.sourceforge.net/lists/listinfo/ant-contrib-developers
>
>

Re: [Ant-contrib-developers] Status

[Stefan B. <ste...@fr...>]

On 2011-07-08, Matthew Inger wrote:

> Also, I got kind of frustrated by the constant API changes in Ant
> itself which kept breaking various tasks. As a result, I had to
> maintain several different versions of the code just to support
> different Ant versions.

Matt, could you please clarify what sort of changes that have been?

Apache Gump has been building ant-contrib's trunk against Ant's trunk
for years without any failure (that wouldn't have been addressed
immediately) at all.

Stefan

Re: [Ant-contrib-developers] Status

[Raja N. K. <nag...@te...>]

Hi,

As ant-contrib has become very useful to many, why this code and tasks 
not taken as part of ant it self, so that, once save lot of time and IT 
energy wastage.

Regards,
Nagendra
-- 


  Raja Nagendra Kumar


    Founder & C.T.O


    http://www.tejasoft.com


        " When Teja is tasked - The job gets done!! "


      Work: +91-80-28436591
      Mobile: +91-9886723872

	


      Email: nag...@te... <mailto:nag...@te...>
      Skype/Yahoo: nagendra.raja
      http://www.linkedin.com/in/nagkumar

	


      *TejaSoft Innovations Pvt. Ltd.* <http://www.tejasoft.com>
      17c/17, Sector-II, Nandi Gardens
      J.P Nagar 9th Phase
      Bangalore, Karnataka 560062 India
      <http://maps.google.com/maps?q=17C/17%2C+Sector-II%2C+Nandi+Gardens%2CJ.P+Nagar+9th+Phase%2CBangalore%2CKarnataka+560062%2CIndia&hl=en>








On 7/8/2011 9:31 AM, Matthew Inger wrote:
> A few people have asked the status of the ant-contrib project.
>
> I have not had much time to work on it recently.  Also,  I got kind of 
> frustrated by the constant API changes
> in Ant itself which kept breaking various tasks.  As a result, I had 
> to maintain several different versions of the
> code just to support different Ant versions.
>
> On top of that, we have 1 task that relies on Ivy as well, and that 
> underwent version changes.
>
> Then add into the mix, supporting Java 1.5 and Java 1.6.
>
> It all because too many variables.
>
> I'm willing to get everything working on Ant 1.8.2 (the most recent 
> version), and put out a final
> release.  However, it will require specific versions of Ant, JDK and 
> Ivy.  As those other products
> get updated, newer Ant-Contrib would support only the latest versions 
> of those.  That's not to say
> ant-contrib would not be backward compatible, but that depends on the 
> API changes of it's dependent
> products.
>
> Thoughts?
> ----
> mat...@ya...
> "Once you start down the dark path, forever will it
> dominate your destiny. Consume you it will " - Yoda
>
>
> ------------------------------------------------------------------------------
> All of the data generated in your IT infrastructure is seriously valuable.
> Why? It contains a definitive record of application performance, security
> threats, fraudulent activity, and more. Splunk takes this data and makes
> sense of it. IT sense. And common sense.
> http://p.sf.net/sfu/splunk-d2d-c2
>
>
> _______________________________________________
> Ant-contrib-developers mailing list
> Ant...@li...
> https://lists.sourceforge.net/lists/listinfo/ant-contrib-developers

[Ant-contrib-developers] Status

[Matthew I. <mat...@ya...>]

A few people have asked the status of the ant-contrib project.

I have not had much time to work on it recently.  Also,  I got kind of frustrated by the constant API changes
in Ant itself which kept breaking various tasks.  As a result, I had to maintain several different versions of the
code just to support different Ant versions.

On top of that, we have 1 task that relies on Ivy as well, and that underwent version changes.

Then add into the mix, supporting Java 1.5 and Java 1.6.

It all because too many variables.

I'm willing to get everything working on Ant 1.8.2 (the most recent version), and put out a final
release.  However, it will require specific versions of Ant, JDK and Ivy.  As those other products
get updated, newer Ant-Contrib would support only the latest versions of those.  That's not to say
ant-contrib would not be backward compatible, but that depends on the API changes of it's dependent
products.

Thoughts?
 
----
mat...@ya...
"Once you start down the dark path, forever will it
dominate your destiny.  Consume you it will " - Yoda

Re: [Ant-contrib-developers] C++ code compiles and links with make, compiles but won't link with cpptasks

[Nick W. <nic...@ni...>]

Follow-up: I did a little more Googling and was reminded about -lstdc++, so I added the following:

<linkerarg value="-lstdc++" />

And it works now. Compiles and links perfectly.

But my questions still stand: Why does it use g++ to compile and gcc to link when I asked for g++ for both of them? Do I need to file a bug? Or am I doing something wrong?

Thanks!

Nick

On Jul 7, 2011, at 4:44 AM, Nick Williams wrote:

> Environment: MacBook Pro Intel Core Due running Mac OS X Snow Leopard 10.6.7, And 1.8.2 and cpptasks 1.0b5.
> 
> I have a make file (which I can provide on request) that compiles all of the .cpp files in the directory to .o files then links them as the shared library file librawsockets.jnilib (.jnilib is the Mac extension for .so files that are specific to Java native code ... but it's the same file type). The output when I run this make file is:
> 
> g++ -Wall -O2 -pipe -D_REENTRANT  -I/System/Library/Frameworks/JavaVM.framework/Home/include -c Constants.cpp -o Constants.o
> g++ -Wall -O2 -pipe -D_REENTRANT  -I/System/Library/Frameworks/JavaVM.framework/Home/include -c RawSocketImpl.cpp -o RawSocketImpl.o
> RawSocketImpl.cpp:216: warning: ‘sockaddr* initIPv4SocketAddress(JNIEnv*, sockaddr_in*, _jbyteArray*)’ defined but not used
> RawSocketImpl.cpp:230: warning: ‘sockaddr* initIPv6SocketAddress(JNIEnv*, sockaddr_in6*, _jbyteArray*)’ defined but not used
> g++ -Wall -O2 -pipe -D_REENTRANT  -I/System/Library/Frameworks/JavaVM.framework/Home/include -c SocketUtilities.cpp -o SocketUtilities.o
> g++  -o librawsockets.jnilib Constants.o RawSocketImpl.o SocketUtilities.o  -dynamiclib -noprebind -single_module -framework JavaVM
> ld: warning: option -noprebind is obsolete and being ignored
> 
> The file is created successfully and all of the native code works with my Java code perfectly. Now, I'm trying to port this to Ant because I'm using Ant to compile my Java code and run my unit tests and I want Ant to do this, too. So here's what I have so far:
> 
> 	<cc name="g++" objdir="${rawsockets.dir.output.native}" outtype="shared" runtime="dynamic" rebuild="true" relentless="true"
> 		warnings="diagnostic" outfile="librawsockets.jnilib" optimize="full" failonerror="true">
> 		<defineset define="_REENTRANT" />
> 		<includepath path="${jdk.include.1.6}" />
> 		<compilerarg value="-pipe" />
> 		<fileset dir="/Users/Nicholas/RawSocketsNativeSrc" includes="*.cpp" />
> 		<linker name="g++">
> 			<linkerarg value="-noprebind" />
> 			<linkerarg value="-single_module" />
> 			<linkerarg value="-framework" />
> 			<linkerarg value="JavaVM" />
> 		</linker>
> 	</cc>
> 
> And my output is this (it compiles exactly the same but doesn't link):
> 
>       [cc] 3 total files to be compiled.
>       [cc] g++ -pipe -c -O2 -fPIC -D_REENTRANT -W -Wall -I/System/Library/Java/JavaVirtualMachines/1.6.0.jdk/Contents/Home/include /Users/Nicholas/RawSocketsNativeSrc/RawSocketImpl.cpp /Users/Nicholas/RawSocketsNativeSrc/Constants.cpp /Users/Nicholas/RawSocketsNativeSrc/SocketUtilities.cpp
>       [cc] /Users/Nicholas/RawSocketsNativeSrc/RawSocketImpl.cpp:216: warning: ‘sockaddr* initIPv4SocketAddress(JNIEnv*, sockaddr_in*, _jbyteArray*)’ defined but not used
>       [cc] /Users/Nicholas/RawSocketsNativeSrc/RawSocketImpl.cpp:230: warning: ‘sockaddr* initIPv6SocketAddress(JNIEnv*, sockaddr_in6*, _jbyteArray*)’ defined but not used
>       [cc] Starting link
>       [cc] gcc -Wl,-noprebind -Wl,-single_module -framework JavaVM -prebind -dynamiclib -o liblibrawsockets.jnilib.so out/production/native/RawSocketImpl.o out/production/native/Constants.o out/production/native/SocketUtilities.o
>       [cc] ld: warning: option -noprebind is obsolete and being ignored
>       [cc] Undefined symbols:
>       [cc]   "___cxa_rethrow", referenced from:
>       [cc]       std::_Rb_tree<std::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int>, std::_Select1st<std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> >, std::less<std::basic_string<char, std::char_traits<char>, std::allocator<char> > >, std::allocator<std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> > >::_M_insert(std::_Rb_tree_node_base*, std::_Rb_tree_node_base*, std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> const&)in Constants.o
>       [cc]   "operator new(unsigned long)", referenced from:
>       [cc]       std::_Rb_tree<std::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int>, std::_Select1st<std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> >, std::less<std::basic_string<char, std::char_traits<char>, std::allocator<char> > >, std::allocator<std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> > >::_M_insert(std::_Rb_tree_node_base*, std::_Rb_tree_node_base*, std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> const&)in Constants.o
>       [cc]       std::vector<_jobject*, std::allocator<_jobject*> >::_M_insert_aux(__gnu_cxx::__normal_iterator<_jobject**, std::vector<_jobject*, std::allocator<_jobject*> > >, _jobject* const&)in SocketUtilities.o
>       (and on and on and on)
> BUILD FAILED
> 
> I even tried this:
> 
> 	<cc name="g++" objdir="${rawsockets.dir.output.native}" outtype="shared" runtime="dynamic" rebuild="true" relentless="true"
> 		warnings="diagnostic" outfile="rawsockets" optimize="full" failonerror="true">
> 		<defineset define="_REENTRANT" />
> 		<includepath path="${jdk.include.1.6}" />
> 		<compilerarg value="-pipe" />
> 		<fileset dir="/Users/Nicholas/RawSocketsNativeSrc" includes="*.cpp" />
> 		<linker name="g++">
> 			<linkerarg value="-single_module" />
> 			<linkerarg value="-framework" />
> 			<linkerarg value="JavaVM" />
> 		</linker>
> 	</cc>
> 
> But I still get this:
> 
>       [cc] 3 total files to be compiled.
>       [cc] g++ -pipe -c -O2 -fPIC -D_REENTRANT -W -Wall -I/System/Library/Java/JavaVirtualMachines/1.6.0.jdk/Contents/Home/include /Users/Nicholas/RawSocketsNativeSrc/RawSocketImpl.cpp /Users/Nicholas/RawSocketsNativeSrc/Constants.cpp /Users/Nicholas/RawSocketsNativeSrc/SocketUtilities.cpp
>       [cc] /Users/Nicholas/RawSocketsNativeSrc/RawSocketImpl.cpp:216: warning: ‘sockaddr* initIPv4SocketAddress(JNIEnv*, sockaddr_in*, _jbyteArray*)’ defined but not used
>       [cc] /Users/Nicholas/RawSocketsNativeSrc/RawSocketImpl.cpp:230: warning: ‘sockaddr* initIPv6SocketAddress(JNIEnv*, sockaddr_in6*, _jbyteArray*)’ defined but not used
>       [cc] Starting link
>       [cc] gcc -Wl,-single_module -framework JavaVM -prebind -dynamiclib -o librawsockets.so out/production/native/RawSocketImpl.o out/production/native/Constants.o out/production/native/SocketUtilities.o
>       [cc] Undefined symbols:
>       [cc]   "___cxa_rethrow", referenced from:
>       [cc]       std::_Rb_tree<std::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int>, std::_Select1st<std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> >, std::less<std::basic_string<char, std::char_traits<char>, std::allocator<char> > >, std::allocator<std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> > >::_M_insert(std::_Rb_tree_node_base*, std::_Rb_tree_node_base*, std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> const&)in Constants.o
>       [cc]   "operator new(unsigned long)", referenced from:
>       [cc]       std::_Rb_tree<std::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int>, std::_Select1st<std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> >, std::less<std::basic_string<char, std::char_traits<char>, std::allocator<char> > >, std::allocator<std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> > >::_M_insert(std::_Rb_tree_node_base*, std::_Rb_tree_node_base*, std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> const&)in Constants.o
>       [cc]       std::vector<_jobject*, std::allocator<_jobject*> >::_M_insert_aux(__gnu_cxx::__normal_iterator<_jobject**, std::vector<_jobject*, std::allocator<_jobject*> > >, _jobject* const&)in SocketUtilities.o
>       (and on and on and on)
> BUILD FAILED
> 
> I can manually change the name of the .so/.jnilib file, so I'm not worried about that. But why won't this thing link, I wondered? The things that I see are different are the fact that the linker is using gcc (even though the compiler is g++ and I specifically told it to use g++) and that the options -Wl, and -prebind are added. I suspected that gcc was the problem, so I changed my makefile to use gcc to link (but keep using g++ to compile) just like the Ant output and it failed the exact same way, with all of the Undefined symbols errors.
> 
> Is this a bug that it's mixing the gcc linker with the g++ compiler (and ignoring my request for the g++ linker)? Or am I doing something wrong?
> 
> Thanks!
> 
> Nick

[Ant-contrib-developers] C++ code compiles and links with make, compiles but won't link with cpptasks

[Nick W. <nic...@ni...>]

Environment: MacBook Pro Intel Core Due running Mac OS X Snow Leopard 10.6.7, And 1.8.2 and cpptasks 1.0b5.

I have a make file (which I can provide on request) that compiles all of the .cpp files in the directory to .o files then links them as the shared library file librawsockets.jnilib (.jnilib is the Mac extension for .so files that are specific to Java native code ... but it's the same file type). The output when I run this make file is:

g++ -Wall -O2 -pipe -D_REENTRANT  -I/System/Library/Frameworks/JavaVM.framework/Home/include -c Constants.cpp -o Constants.o
g++ -Wall -O2 -pipe -D_REENTRANT  -I/System/Library/Frameworks/JavaVM.framework/Home/include -c RawSocketImpl.cpp -o RawSocketImpl.o
RawSocketImpl.cpp:216: warning: ‘sockaddr* initIPv4SocketAddress(JNIEnv*, sockaddr_in*, _jbyteArray*)’ defined but not used
RawSocketImpl.cpp:230: warning: ‘sockaddr* initIPv6SocketAddress(JNIEnv*, sockaddr_in6*, _jbyteArray*)’ defined but not used
g++ -Wall -O2 -pipe -D_REENTRANT  -I/System/Library/Frameworks/JavaVM.framework/Home/include -c SocketUtilities.cpp -o SocketUtilities.o
g++  -o librawsockets.jnilib Constants.o RawSocketImpl.o SocketUtilities.o  -dynamiclib -noprebind -single_module -framework JavaVM
ld: warning: option -noprebind is obsolete and being ignored

The file is created successfully and all of the native code works with my Java code perfectly. Now, I'm trying to port this to Ant because I'm using Ant to compile my Java code and run my unit tests and I want Ant to do this, too. So here's what I have so far:

	<cc name="g++" objdir="${rawsockets.dir.output.native}" outtype="shared" runtime="dynamic" rebuild="true" relentless="true"
		warnings="diagnostic" outfile="librawsockets.jnilib" optimize="full" failonerror="true">
		<defineset define="_REENTRANT" />
		<includepath path="${jdk.include.1.6}" />
		<compilerarg value="-pipe" />
		<fileset dir="/Users/Nicholas/RawSocketsNativeSrc" includes="*.cpp" />
		<linker name="g++">
			<linkerarg value="-noprebind" />
			<linkerarg value="-single_module" />
			<linkerarg value="-framework" />
			<linkerarg value="JavaVM" />
		</linker>
	</cc>

And my output is this (it compiles exactly the same but doesn't link):

       [cc] 3 total files to be compiled.
       [cc] g++ -pipe -c -O2 -fPIC -D_REENTRANT -W -Wall -I/System/Library/Java/JavaVirtualMachines/1.6.0.jdk/Contents/Home/include /Users/Nicholas/RawSocketsNativeSrc/RawSocketImpl.cpp /Users/Nicholas/RawSocketsNativeSrc/Constants.cpp /Users/Nicholas/RawSocketsNativeSrc/SocketUtilities.cpp
       [cc] /Users/Nicholas/RawSocketsNativeSrc/RawSocketImpl.cpp:216: warning: ‘sockaddr* initIPv4SocketAddress(JNIEnv*, sockaddr_in*, _jbyteArray*)’ defined but not used
       [cc] /Users/Nicholas/RawSocketsNativeSrc/RawSocketImpl.cpp:230: warning: ‘sockaddr* initIPv6SocketAddress(JNIEnv*, sockaddr_in6*, _jbyteArray*)’ defined but not used
       [cc] Starting link
       [cc] gcc -Wl,-noprebind -Wl,-single_module -framework JavaVM -prebind -dynamiclib -o liblibrawsockets.jnilib.so out/production/native/RawSocketImpl.o out/production/native/Constants.o out/production/native/SocketUtilities.o
       [cc] ld: warning: option -noprebind is obsolete and being ignored
       [cc] Undefined symbols:
       [cc]   "___cxa_rethrow", referenced from:
       [cc]       std::_Rb_tree<std::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int>, std::_Select1st<std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> >, std::less<std::basic_string<char, std::char_traits<char>, std::allocator<char> > >, std::allocator<std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> > >::_M_insert(std::_Rb_tree_node_base*, std::_Rb_tree_node_base*, std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> const&)in Constants.o
       [cc]   "operator new(unsigned long)", referenced from:
       [cc]       std::_Rb_tree<std::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int>, std::_Select1st<std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> >, std::less<std::basic_string<char, std::char_traits<char>, std::allocator<char> > >, std::allocator<std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> > >::_M_insert(std::_Rb_tree_node_base*, std::_Rb_tree_node_base*, std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> const&)in Constants.o
       [cc]       std::vector<_jobject*, std::allocator<_jobject*> >::_M_insert_aux(__gnu_cxx::__normal_iterator<_jobject**, std::vector<_jobject*, std::allocator<_jobject*> > >, _jobject* const&)in SocketUtilities.o
       (and on and on and on)
BUILD FAILED

I even tried this:

	<cc name="g++" objdir="${rawsockets.dir.output.native}" outtype="shared" runtime="dynamic" rebuild="true" relentless="true"
		warnings="diagnostic" outfile="rawsockets" optimize="full" failonerror="true">
		<defineset define="_REENTRANT" />
		<includepath path="${jdk.include.1.6}" />
		<compilerarg value="-pipe" />
		<fileset dir="/Users/Nicholas/RawSocketsNativeSrc" includes="*.cpp" />
		<linker name="g++">
			<linkerarg value="-single_module" />
			<linkerarg value="-framework" />
			<linkerarg value="JavaVM" />
		</linker>
	</cc>

But I still get this:

       [cc] 3 total files to be compiled.
       [cc] g++ -pipe -c -O2 -fPIC -D_REENTRANT -W -Wall -I/System/Library/Java/JavaVirtualMachines/1.6.0.jdk/Contents/Home/include /Users/Nicholas/RawSocketsNativeSrc/RawSocketImpl.cpp /Users/Nicholas/RawSocketsNativeSrc/Constants.cpp /Users/Nicholas/RawSocketsNativeSrc/SocketUtilities.cpp
       [cc] /Users/Nicholas/RawSocketsNativeSrc/RawSocketImpl.cpp:216: warning: ‘sockaddr* initIPv4SocketAddress(JNIEnv*, sockaddr_in*, _jbyteArray*)’ defined but not used
       [cc] /Users/Nicholas/RawSocketsNativeSrc/RawSocketImpl.cpp:230: warning: ‘sockaddr* initIPv6SocketAddress(JNIEnv*, sockaddr_in6*, _jbyteArray*)’ defined but not used
       [cc] Starting link
       [cc] gcc -Wl,-single_module -framework JavaVM -prebind -dynamiclib -o librawsockets.so out/production/native/RawSocketImpl.o out/production/native/Constants.o out/production/native/SocketUtilities.o
       [cc] Undefined symbols:
       [cc]   "___cxa_rethrow", referenced from:
       [cc]       std::_Rb_tree<std::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int>, std::_Select1st<std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> >, std::less<std::basic_string<char, std::char_traits<char>, std::allocator<char> > >, std::allocator<std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> > >::_M_insert(std::_Rb_tree_node_base*, std::_Rb_tree_node_base*, std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> const&)in Constants.o
       [cc]   "operator new(unsigned long)", referenced from:
       [cc]       std::_Rb_tree<std::basic_string<char, std::char_traits<char>, std::allocator<char> >, std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int>, std::_Select1st<std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> >, std::less<std::basic_string<char, std::char_traits<char>, std::allocator<char> > >, std::allocator<std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> > >::_M_insert(std::_Rb_tree_node_base*, std::_Rb_tree_node_base*, std::pair<std::basic_string<char, std::char_traits<char>, std::allocator<char> > const, int> const&)in Constants.o
       [cc]       std::vector<_jobject*, std::allocator<_jobject*> >::_M_insert_aux(__gnu_cxx::__normal_iterator<_jobject**, std::vector<_jobject*, std::allocator<_jobject*> > >, _jobject* const&)in SocketUtilities.o
       (and on and on and on)
BUILD FAILED

I can manually change the name of the .so/.jnilib file, so I'm not worried about that. But why won't this thing link, I wondered? The things that I see are different are the fact that the linker is using gcc (even though the compiler is g++ and I specifically told it to use g++) and that the options -Wl, and -prebind are added. I suspected that gcc was the problem, so I changed my makefile to use gcc to link (but keep using g++ to compile) just like the Ant output and it failed the exact same way, with all of the Undefined symbols errors.

Is this a bug that it's mixing the gcc linker with the g++ compiler (and ignoring my request for the g++ linker)? Or am I doing something wrong?

Thanks!

Nick

[Ant-contrib-developers] [Help] RE: Any plans for a final release?

[SourceForge.net <no...@so...>]

The following forum message was posted by guido_leenders at http://sourceforge.net/projects/ant-contrib/forums/forum/113701/topic/1554714:

We are using it a lot. It is in general reliable. Some bugs, but fixing myself
is too much work. I found little to no documentation / comment in the code
studied.

[Ant-contrib-developers] [Help] RE: Any plans for a final release?

[SourceForge.net <no...@so...>]

The following forum message was posted by davidchanck at http://sourceforge.net/projects/ant-contrib/forums/forum/113701/topic/1554714:

Is no one using this?  Why is this still in beta?

[Ant-contrib-developers] [Help] &quot; bool&quot; and &quot; islessthan...

[SourceForge.net <no...@so...>]

The following forum message was posted by stephenk at http://sourceforge.net/projects/ant-contrib/forums/forum/113701/topic/3792513:

Back in the day there was (apparently) a BooleanConditionTask that extended BooleanConditionBase (that enabled support for the <, > etc conditions)

What happened to this?  Is there any way to use the \"islessthan\" and \"isgreaterthan\" conditions as part of if\'s or otherwise?

Re: [Ant-contrib-developers] Using mingw with cpptasks

[Barker, A. <ab...@ca...>]

If your goal is cross platform building with ant write a Makefile for the native sections and call the exec task for make.  It will save you a lot of time trying to figure all the quirky platform dependent bugs in the cpptasks.  If your only going to be building on a specific system and need to cross compile with mingw then the symlinks are the only method that I know of unless you want to extend/patch the compiler tasks to add mingw support.

I hope that helps,

Alexander Barker
Senior Programmer
California Lutheran University
Information Systems and Services
60 West Olsen Road
Thousand Oaks, CA 91360
805.493.3912


On 07/20/2010 08:41 AM, Bailey, Darragh wrote:
>
>
> Wondering if it is possible to use mingw with cpptasks? 
>
> Since mingw is really just a modified version of gcc I would expect that it can as cpptasks handles gcc already, however since in some cases Linux distributions use different names for the mingw gcc, i.e. x86_64-pc-mingw32-gcc, how can I get cpptasks to use the mingw one instead of the system default gcc?
>
> Right now all I can do is create symlinks under a local directory (i.e. gcc -> /usr/bin/x86_64-pc-mingw32-gcc) and use PATH to make sure the mingw gcc is called before the standard gcc. Which is not particularly ideal.
>
> Any better solutions available?
>
> --
> Regards,
> Darragh Bailey
>
> Systems Software Engineer
> Hewlett Packard Galway Ltd.
>
> Postal Address:    Hewlett Packard Galway Limited, Ballybrit Business Park, Galway
> Registered Office: Hewlett Packard Galway Limited, 63-74 Sir John Rogerson's Quay Dublin 2
> Registered Number: 361933 
>
>
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by Sprint
> What will you do first with EVO, the first 4G phone?
> Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first
> _______________________________________________
> Ant-contrib-developers mailing list
> Ant...@li...
> https://lists.sourceforge.net/lists/listinfo/ant-contrib-developers

[Ant-contrib-developers] Using mingw with cpptasks

[Bailey, D. <db...@hp...>]

Wondering if it is possible to use mingw with cpptasks? 

Since mingw is really just a modified version of gcc I would expect that it can as cpptasks handles gcc already, however since in some cases Linux distributions use different names for the mingw gcc, i.e. x86_64-pc-mingw32-gcc, how can I get cpptasks to use the mingw one instead of the system default gcc?

Right now all I can do is create symlinks under a local directory (i.e. gcc -> /usr/bin/x86_64-pc-mingw32-gcc) and use PATH to make sure the mingw gcc is called before the standard gcc. Which is not particularly ideal.

Any better solutions available?

--
Regards,
Darragh Bailey

Systems Software Engineer
Hewlett Packard Galway Ltd.

Postal Address:    Hewlett Packard Galway Limited, Ballybrit Business Park, Galway
Registered Office: Hewlett Packard Galway Limited, 63-74 Sir John Rogerson's Quay Dublin 2
Registered Number: 361933 

[Ant-contrib-developers] [Help] RE: propertyregex global attribute not working

[SourceForge.net <no...@so...>]

The following forum message was posted by biaib at http://sourceforge.net/projects/ant-contrib/forums/forum/113701/topic/3713234:

hu, I just realize my mistake :

should have use [code] replace=\"\\1-min.js\"[/code] instead [code] select=\"\\1-min.js\" [/code].

[Ant-contrib-developers] [Help] propertyregex global attribute not working

[SourceForge.net <no...@so...>]

The following forum message was posted by biaib at http://sourceforge.net/projects/ant-contrib/forums/forum/113701/topic/3713234:

Hillo,

I must use a global replacement regex, but <propertyregex> only match the first occurence. Here\'s the code:

[code]<propertyregex property=\"append.order.min\"
	input=\"bla,blabla,tic\"
	regexp=\"([^,]+)\"
	select=\"\\1-min.js\"
	global=\"true\"
	defaultValue=\"fail\" />[/code]

append.order.min returns 
[code]\'bla-min.js\' [/code]
instead
[code]\'bla-min.js,blabla-min.js,tic-min.js\'[/code]


I\'ve check my regex with http://gskinner.com/RegExr/, and it work there.

the documentation say this about attribute global:
global : 	Should a replacement operation be performed on the entire string, rather than just the first occurance
Not required. default is false.


Am I missing something or this is a bug ?

[Ant-contrib-developers] g++ compile problem

[<men...@st...>]

Hi，

First,please forgive for my poor english.I am using Apache Ant 1.7.0 and 
am an ant beginner.My build environment is freebsd6.3 + diablo-jdk1.5.0 + 
g++4.4.3.

I found the obj file's size which compiled by ant different from which 
built directly by g++. As below:
%g++ -ggdb -Werror -Wreturn-type -c -DHAVE_CONFIG_H -D_DEBUG -I. -I.. 
-I../Zoic/mysql/include -I../protocol -I../protocol/Types 
-I../protocol/Login -o Cmd_Server.o command/Cmd_Server.cpp
%ll Cmd_Server.o
-rw-r--r--  1 lmy  lmy  176332 Apr 15 17:44 Cmd_Server.o

Compiled by ant as below:
%ll Cmd_Server.o
-rw-r--r--  1 lmy  lmy  176620 Apr 15 11:37 Cmd_Server.o
Here is history.xml:
<processor signature="g++ noversion nomachine -ggdb -Werror -Wreturn-type 
-c -DHAVE_CONFIG_H -D_DEBUG -I. -I.. -I../Zoic/mysql/include -I../protocol 
-I../protocol/Types -I../protocol/Login -Iprotocol/Types"> 

Does it matter ？Can any one explain this for me?

Further more,here is another problem:When I was building project by 
ant,CPU often went high level and the build was halted .And then I had to 
kill the java process to continue ant build.It troubled me a lot.Any one 
has idea?


Any help would be appreciated. 

[Ant-contrib-developers] statically link?

[Hei C. <str...@ya...>]

Hi,

I have a cc task and want to link one of the 3rd party library statically.

Is it correct to do something like this:
        <libset libs="aStaticLibaray" type="static" />
        <libset libs="dynamicLib1, dynamicLib2" />
?

Thanks in advance.


Cheers,
Hei



      

[Ant-contrib-developers] Fwd: PostMethod task + MyClient bugs???

[А.А. Д. <and...@tu...>]

Hello.

1) I have noticed the bug in the documentation see
apache\ant\docs\manual\tasks\http-client_type.html.
The RFC2109 should be used instead of RFC_2109 (which cases an
exception: java.lang.IllegalStateException: Unsupported cookie spec
RFC_2109)

2) I believe that there are some bugs in
"net/sf/antcontrib/net/httpclient/PostMethodTask.java".
I belive that the first parameter in
==================================
partArray[i] = new FilePart(filePart.getPath().getName(),
                                filePart.getPath().getName(),
                                filePart.getPath(),
                                filePart.getContentType(),
                                filePart.getCharSet());
==================================
Should be the name of file attribute (file-input control) not the
filename. (see attached patched file)

For example: Mozilla uses different names for file and for parameter:
-----------------------------------------
Content-Disposition: form-data; name="pic_file[]";
filename="AbstractMethodTask.class"
-----------------------------------------

3) Have anybody succeeded in uploading file to the server. I can't
still do it :(. But I can not also file a bug.


Thanks a lot.
Andrei.