Re: New Antlib
For the information of dev@, the new repo discussed has been created at https://gitbox.apache.org/repos/asf/ant-antlibs-s3.git . Additionally I committed the current state of my work before realizing that its notifications were set up improperly; this has now been fixed and future commits should be reported to Ant's notification list. Matt On Mon, Feb 21, 2022, 1:31 PM Matt Benson wrote: > Thanks, Stefan. Will do. > > Matt > > On Mon, Feb 21, 2022, 12:50 PM Stefan Bodewig wrote: > >> On 2022-02-20, Matt Benson wrote: >> >> > The plan from this thread is two years old, so I want to revisit. As >> > of now the extant antlibs seem to have their own repos. If I am >> > creating a new one, do I want to create its own repo now, or would we >> > rather create a sandbox repo and promote its children as they >> > "graduate" to their own repos? >> >> I believe there hasn't been any change to the list of antlibs in this >> two years. Neither of them is really active. And to be honest I don't >> expect any of the existing sandbox antilibs to ever leave that state. >> >> If you plan to bring the Antlib from sandox to proper anyway, I'd >> recommend starting with a new targetted repository reight from the >> start. This is not an incredibly strong preference, though. >> >> Stefan >> >> - >> To unsubscribe, e-mail: dev-unsubscr...@ant.apache.org >> For additional commands, e-mail: dev-h...@ant.apache.org >> >>
Re: New Antlib
Thanks, Stefan. Will do. Matt On Mon, Feb 21, 2022, 12:50 PM Stefan Bodewig wrote: > On 2022-02-20, Matt Benson wrote: > > > The plan from this thread is two years old, so I want to revisit. As > > of now the extant antlibs seem to have their own repos. If I am > > creating a new one, do I want to create its own repo now, or would we > > rather create a sandbox repo and promote its children as they > > "graduate" to their own repos? > > I believe there hasn't been any change to the list of antlibs in this > two years. Neither of them is really active. And to be honest I don't > expect any of the existing sandbox antilibs to ever leave that state. > > If you plan to bring the Antlib from sandox to proper anyway, I'd > recommend starting with a new targetted repository reight from the > start. This is not an incredibly strong preference, though. > > Stefan > > - > To unsubscribe, e-mail: dev-unsubscr...@ant.apache.org > For additional commands, e-mail: dev-h...@ant.apache.org > >
Re: New Antlib
On 2022-02-20, Matt Benson wrote: > The plan from this thread is two years old, so I want to revisit. As > of now the extant antlibs seem to have their own repos. If I am > creating a new one, do I want to create its own repo now, or would we > rather create a sandbox repo and promote its children as they > "graduate" to their own repos? I believe there hasn't been any change to the list of antlibs in this two years. Neither of them is really active. And to be honest I don't expect any of the existing sandbox antilibs to ever leave that state. If you plan to bring the Antlib from sandox to proper anyway, I'd recommend starting with a new targetted repository reight from the start. This is not an incredibly strong preference, though. Stefan - To unsubscribe, e-mail: dev-unsubscr...@ant.apache.org For additional commands, e-mail: dev-h...@ant.apache.org
Re: New Antlib
The plan from this thread is two years old, so I want to revisit. As of now the extant antlibs seem to have their own repos. If I am creating a new one, do I want to create its own repo now, or would we rather create a sandbox repo and promote its children as they "graduate" to their own repos? Matt On Sat, Mar 7, 2020, 4:33 PM Matt Benson wrote: > Thanks, Stefan! Sounds like a good plan. > > Matt > > On Sat, Mar 7, 2020, 9:52 AM Stefan Bodewig wrote: > >> On 2020-03-05, Matt Benson wrote: >> >> > What is our current protocol for creating a new Antlib? >> >> Do we have one? :-) >> >> > I have written some types for working with Amazon S3 objects as Ant >> > resources and think they could be generally useful to the community. >> >> Sounds good. >> >> You could start with a sandbox immediately if you wanted to, but after >> the move to git there really isn't any big difference. IMHO you should >> just create a new antlibs git repository and add something under >> "sandbox" and we talk about "graduating" the sandbox once it is ready >> for a release (candidate). >> >> Stefan >> >> - >> To unsubscribe, e-mail: dev-unsubscr...@ant.apache.org >> For additional commands, e-mail: dev-h...@ant.apache.org >> >>
Re: New Antlib
Thanks, Stefan! Sounds like a good plan. Matt On Sat, Mar 7, 2020, 9:52 AM Stefan Bodewig wrote: > On 2020-03-05, Matt Benson wrote: > > > What is our current protocol for creating a new Antlib? > > Do we have one? :-) > > > I have written some types for working with Amazon S3 objects as Ant > > resources and think they could be generally useful to the community. > > Sounds good. > > You could start with a sandbox immediately if you wanted to, but after > the move to git there really isn't any big difference. IMHO you should > just create a new antlibs git repository and add something under > "sandbox" and we talk about "graduating" the sandbox once it is ready > for a release (candidate). > > Stefan > > - > To unsubscribe, e-mail: dev-unsubscr...@ant.apache.org > For additional commands, e-mail: dev-h...@ant.apache.org > >
Re: New Antlib
On 2020-03-05, Matt Benson wrote: > What is our current protocol for creating a new Antlib? Do we have one? :-) > I have written some types for working with Amazon S3 objects as Ant > resources and think they could be generally useful to the community. Sounds good. You could start with a sandbox immediately if you wanted to, but after the move to git there really isn't any big difference. IMHO you should just create a new antlibs git repository and add something under "sandbox" and we talk about "graduating" the sandbox once it is ready for a release (candidate). Stefan - To unsubscribe, e-mail: dev-unsubscr...@ant.apache.org For additional commands, e-mail: dev-h...@ant.apache.org
New Antlib
Hello all, What is our current protocol for creating a new Antlib? I have written some types for working with Amazon S3 objects as Ant resources and think they could be generally useful to the community. Matt
DO NOT REPLY [Bug 28782] New: - Antlib ClassLoading issues in J2EE environment
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT http://issues.apache.org/bugzilla/show_bug.cgi?id=28782. ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND INSERTED IN THE BUG DATABASE. http://issues.apache.org/bugzilla/show_bug.cgi?id=28782 Antlib ClassLoading issues in J2EE environment Summary: Antlib ClassLoading issues in J2EE environment Product: Ant Version: 1.6.1 Platform: All OS/Version: All Status: NEW Severity: Normal Priority: Other Component: Core AssignedTo: [EMAIL PROTECTED] ReportedBy: [EMAIL PROTECTED] Hi all, Diving deeper into Ant's code I might have identified the problem... First of all the Definer uses the Project.createClassLoader method to get a ClassLoader in order to lookup the antlib.xml file and further instanciate the tasks. However, by default, this ClassLoader does not have any parent and uses the java.class.path sysprop as its path So I first try adding the project instance's ClassLoader as the parent of the newly created ClassLoader However this does not solved the problem... This is because later on, when lookin'up the antlib.xml using the getResources method, the AntClassLoader (through the findResources method) uses its inner class ResourceEnumeration to delegate the lookup but this Class does not take into account the parent classloader (which is different from the normal parent which is always null). Thus, only the java.class.path is taken into account when trying to find the antlib.xml, which obviously only works when in standalone mode. I additionaly changed the findResource method in AntClassLoader to use a CompoundEnumeration when a parent has been set and it finally worked :-) I guess I should enter a bug report... Cheers, Claude -Original Message- From: Claude Vedovini [mailto:[EMAIL PROTECTED] Sent: 4 May 2004 11:29 To: 'Ant Users List' Subject: RE: Antlib ClassLoading issues Hi, Actually I tried to modify the code for the ant task, initializing the coreloader of the called build this does not seems to be enough. Thanks for the hint, seems like I used the core loader for nothing since two years :-) Anyhow, could someone tell me how to set the classloader for the whole stuff? Thanks in advance, Claude -Original Message- From: Peter Reilly [mailto:[EMAIL PROTECTED] Sent: 4 May 2004 10:57 To: Ant Users List Subject: Re: Antlib ClassLoading issues Hi, I did not know that the coreloader was actually used. See: http://marc.theaimsgroup.com/?l=ant-devm=107451990930463w=2 for some discussion on this. Be that as is may, if the coreloader is in the code, it should be passed to the child projects. As a work-around, you could trigger the loading of the antlib tasks and types in the initial project: For example add a macro to the antlib : init macrodef name=init sequential /sequential /macrodef and call it in the initial project: project xmlns:s=antlib:com.company.project default=subs s:init/ target name=subs subant ../ /target /project It may be better to do something like this in any case, otherwise the tasks/types will be loaded in each of the sub-projects. Peter Claude Vedovini wrote: I all, I am just trying to take advantage of the Antlib feature and it seems that in certain circumstances it does not work as expected :-) Ok, here is the background: We built a nice continuous integration infrastructure using ant, which is driven by a web application (our build launcher, currently running with Websphere, Weblogic and Tomcat) we also developed some tasks that are automatically put in the context of the build by our launcher. Grosso-modo the launcher is a servlet that creates a Project instance, configures it and so on. As you can imagine it also uses the setCoreLoader method to give the Project the web-app ClassLoader (which is a non-delegating J2EE ClassLoader). All this used to work since two years or so... Those last days I tried to promote out own tasks to an Antlib library in order to take advantage of the namespace stuff and here is the problem I ran into: - When I declare my antlib: namespace in a sub build (called via the ant task) I got a ClassNotFoundException (and Ant tells me that I certainly forgot to add a jar in the lib folder) I took a look at the ant task and it seems that the CoreLoader of the calling project (which should further be used to load tasks I guess) is not passed through to the called project. Taskdefs are, which explains why it was working previously, but as I guess that antlibs are loaded on request (when the namespace declaration is met) and using the project's ClassLoader then
DO NOT REPLY [Bug 25226] New: - antlib availability
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT http://nagoya.apache.org/bugzilla/show_bug.cgi?id=25226. ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND INSERTED IN THE BUG DATABASE. http://nagoya.apache.org/bugzilla/show_bug.cgi?id=25226 antlib availability Summary: antlib availability Product: Ant Version: 1.6Beta Platform: All OS/Version: Other Status: NEW Severity: Enhancement Priority: Other Component: Core AssignedTo: [EMAIL PROTECTED] ReportedBy: [EMAIL PROTECTED] It is possible to write antlibs that are composed, for example, entirely of *def element sets and thus use no custom code. Currently the way to make these available via antlib namespaces, as I understand it, is to package the antlib in a jar file with an appropriate directory structure, and deploy it as you would any other Ant add-on. Would it make sense to create an ANT_HOME subdirectory specifically for jarless antlibs? This could be implemented as a classes type directory that would work for classes and/or resources. - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]