Re: Out of memory errors since recent changes.

Implementation seems strange. Usually thread local objects are used as following: private static ThreadLocal threadLocalProperties=new InheritableThreadLocal(); // notice "static" there public static ThreadLocalProperties getLocalProperties() { // static too ThreadLocalProperties properties=(

AW: cvs commit: ant/docs/manual/CoreTypes selectors.html

Should we place a warning (or note :) in the wrapper scripts if CLASSPATH is set? Jan > -Ursprüngliche Nachricht- > Von: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] > Gesendet am: Freitag, 14. Januar 2005 00:10 > An: [EMAIL PROTECTED] > Betreff: cvs commit: ant/docs/manual/CoreTypes selec

Re: DynamicConfigurator and friends

On Thu, 13 Jan 2005, Brian Topping <[EMAIL PROTECTED]> wrote: > I'm trying to use DynamicConfigurator and need a bit of > clarification. I'd like to make the child text of a Task (defined > with a taskdef) be usable as a String by the Task. That would be the addText(String) method, no DynamicFoo

cvs commit: ant/src/main/org/apache/tools/ant/taskdefs MacroDef.java MacroInstance.java

bodewig 2005/01/14 00:12:04 Modified:src/main/org/apache/tools/ant PropertyHelper.java src/main/org/apache/tools/ant/taskdefs MacroDef.java MacroInstance.java Log: 2005 Revision ChangesPath 1.21 +1 -1 ant/src/main/org/apach

cvs commit: ant/xdocs external.xml

bodewig 2005/01/14 00:14:12 Modified:docs contributors.html external.html xdocsexternal.xml Log: Add pointer to AntForm Revision ChangesPath 1.54 +1 -1 ant/docs/contributors.html Index: contributors.html

Property name policy (was Re: cvs commit: ant/docs/manual/CoreTypes selectors.html)

On 13 Jan 2005, <[EMAIL PROTECTED]> wrote: > I have declared a policy on what we will name new properties. Is > this a good formal policy to have? > If new properties get added (it happens), expect them to appear > under the "ant." and "org.apache.tools.ant" prefixes, unless the > developers ha

Re: Out of memory errors since recent changes.

Martijn Kruithof wrote: I can see from traces I added to the code that thousands of LocalPropertyStacks are created via the |InheritableThreadLocal|, hundreds of them seem to be never cleaned up, at the same time the number of threads remains more or less constant. When rolling back the last two

Re: Property name policy (was Re: cvs commit: ant/docs/manual/CoreTypes selectors.html)

Sounds good + 1 Cheers, Antoine > On 13 Jan 2005, <[EMAIL PROTECTED]> wrote: > > > I have declared a policy on what we will name new properties. Is > > this a good formal policy to have? > > > > > If new properties get added (it happens), expect them to appear > > under the "ant." and "org.ap

cvs commit: ant/src/main/org/apache/tools/ant/taskdefs/optional Javah.java

bodewig 2005/01/14 01:45:54 Modified:docs/manual clonevm.html sysclasspath.html src/main/org/apache/tools/ant/taskdefs Javadoc.java src/main/org/apache/tools/ant/taskdefs/optional Javah.java Log: Finalize and document bootclasspath changes Revision

cvs commit: ant WHATSNEW

bodewig 2005/01/14 02:03:50 Modified:.WHATSNEW Log: Some formatting changes and removing some duplicates Revision ChangesPath 1.713 +24 -112 ant/WHATSNEW Index: WHATSNEW === RCS file: /

cvs commit: ant WHATSNEW

bodewig 2005/01/14 02:09:23 Modified:.WHATSNEW Log: Silently ignore messages being written to System.err/out while a Listener is processing a different message Revision ChangesPath 1.714 +5 -0 ant/WHATSNEW Index: WHATSNEW =

cvs commit: ant/src/main/org/apache/tools/ant Project.java

bodewig 2005/01/14 02:09:23 Modified:src/main/org/apache/tools/ant Project.java Log: Silently ignore messages being written to System.err/out while a Listener is processing a different message Revision ChangesPath 1.183 +15 -5 ant/src/main/org/apache/tools/ant/Pr

cvs commit: ant WHATSNEW

bodewig 2005/01/14 02:12:31 Modified:.Tag: ANT_16_BRANCH WHATSNEW Log: merge Revision ChangesPath No revision No revision 1.503.2.159 +12 -99ant/WHATSNEW Index: WHATSNEW ===

cvs commit: ant/src/main/org/apache/tools/ant Project.java

bodewig 2005/01/14 02:12:31 Modified:src/main/org/apache/tools/ant Tag: ANT_16_BRANCH Project.java Log: merge Revision ChangesPath No revision No revision 1.154.2.15 +16 -5 ant/src/main/org/apache/tools/a

Re: Avoiding logging loops

I've changed Ant's CVS HEAD as well as the 1.6 branch to now silently ignore messages written by Listeners while they are supposed to log something else. Stefan - To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands,

AW: Property name policy (was Re: cvs commit: ant/docs/manual/Cor eTypes selectors.html)

> > > I have declared a policy on what we will name new properties. Is > > > this a good formal policy to have? > > > > > > > > > If new properties get added (it happens), expect them to appear > > > under the "ant." and "org.apache.tools.ant" prefixes, unless the > > > developers have a very go

Re: AW: cvs commit: ant/docs/manual/CoreTypes selectors.html

[EMAIL PROTECTED] wrote: Should we place a warning (or note :) in the wrapper scripts if CLASSPATH is set? aahh, maybe too brutal. we dont want to be as irritating as windows XP, do we "you have no network", "updates available", etc, etc :)

Re: Property name policy (was Re: cvs commit: ant/docs/manual/CoreTypes selectors.html)

Stefan Bodewig wrote: On 13 Jan 2005, <[EMAIL PROTECTED]> wrote: I have declared a policy on what we will name new properties. Is this a good formal policy to have? If new properties get added (it happens), expect them to appear under the "ant." and "org.apache.tools.ant" prefixes, unless the de

Re: AW: Property name policy (was Re: cvs commit: ant/docs/manual/Cor eTypes selectors.html)

[EMAIL PROTECTED] wrote: I have declared a policy on what we will name new properties. Is this a good formal policy to have? If new properties get added (it happens), expect them to appear under the "ant." and "org.apache.tools.ant" prefixes, unless the developers have a very good reason to use an

Re: Property name policy

On Fri, 14 Jan 2005, Jan Materne <[EMAIL PROTECTED]> wrote: >> > maybe we should reserve the "build." prefix as well - >> > build.sysclasspath and CVS HEAD's build.clonevm. > > I wouldnt do that - I cant count the numbers of buildfiles where I > have seen build.dir and build.classes ... True. S

DO NOT REPLY [Bug 32973] - batchtest does not work well with custum formatter

DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG· RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT . ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND· INSERTED IN THE BUG DATABASE. http://issues.apache.org/bugzilla/show_bu

DO NOT REPLY [Bug 32815] - Invalid Subant BuildException: "subant task calling a target that depends on its parent target" incorectly handles task scope

DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG· RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT . ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND· INSERTED IN THE BUG DATABASE. http://issues.apache.org/bugzilla/show_bu

Re: Out of memory errors since recent changes.

--- Peter Reilly <[EMAIL PROTECTED]> wrote: > Is this a specific test, or is it all the tests? > > Peter Peter, it's when running the full suite; it's an accumulation problem. Mine usually dies around oata.taskdefs.F* , usually FixCRLFTest but sometimes FilterTest, and Martijn also mentioned E

RE: DynamicConfigurator and friends

> From: Stefan Bodewig [mailto:[EMAIL PROTECTED] > > > > ... arbitrary XML goes here ... > > > > and then serialize the DocumentFragment held by > XMLFragement into a String. Thanks for the explanation Stefan. I didn't grasp the fact that a DynamicConfigurator was still introspected

Re: DynamicConfigurator and friends

On Fri, 14 Jan 2005, Dominique Devienne <[EMAIL PROTECTED]> wrote: > There's one last twist about text handling in Ant that I > discovered with a modified version of though: Ant > merges text nodes for an element, calling addText() only > once, so bde does not call addText("b") and > later addTex

RE: DynamicConfigurator and friends

> From: Stefan Bodewig [mailto:[EMAIL PROTECTED] > > On Fri, 14 Jan 2005, Dominique Devienne <[EMAIL PROTECTED]> wrote: > > > There's one last twist about text handling in Ant that I > > discovered with a modified version of though: Ant > > merges text nodes for an element, calling addText() onl

Re: DynamicConfigurator and friends

On Fri, 14 Jan 2005, Dominique Devienne <[EMAIL PROTECTED]> wrote: >> From: Stefan Bodewig [mailto:[EMAIL PROTECTED] >> It is a problem, but one that would be rather hard to fix without >> breaking BWC. > > With a DynamicConfigurator-based impl of XMLFragment, yes. > > But that's not the only wa

cvs commit: ant/src/main/org/apache/tools/ant PropertyHelper.java

peterreilly2005/01/14 08:19:43 Modified:src/main/org/apache/tools/ant PropertyHelper.java Log: Attempt to cut down on memory usage of local properties for threads that do not use properties. This is a quick fix and not tested too much. The whole thing may be scrapped... Re

cvs commit: ant/src/main/org/apache/tools/ant/taskdefs MacroDef.java MacroInstance.java Ant.java

peterreilly2005/01/14 08:49:35 Modified:src/main/org/apache/tools/ant PropertyHelper.java src/main/org/apache/tools/ant/taskdefs MacroDef.java MacroInstance.java Ant.java Log: revert local property patch until memory leakage problem is fixed

Re: Out of memory errors since recent changes.

Ok, I have seen the problem on my machine (after removing export ANT_OPTS="-Xms200m -Xmx1000m" from my .bashrc file). I tryed to make a solution that required less memory but it did not work. So I pulled the local property impl (for the moment). Jose's solution is possibly the way to go at the mome

cvs commit: ant/src/main/org/apache/tools/ant PropertyHelper.java

peterreilly2005/01/14 09:03:40 Modified:src/main/org/apache/tools/ant PropertyHelper.java Log: date Revision ChangesPath 1.24 +1 -1 ant/src/main/org/apache/tools/ant/PropertyHelper.java Index: PropertyHelper.java ==

cvs commit: ant/src/main/org/apache/tools/ant/taskdefs Ant.java

peterreilly2005/01/14 09:23:38 Modified:src/main/org/apache/tools/ant/taskdefs Ant.java Log: unneeded import Revision ChangesPath 1.116 +0 -1 ant/src/main/org/apache/tools/ant/taskdefs/Ant.java Index: Ant.java ==

Re: AW: Property name policy (was Re: cvs commit: ant/docs/manual/Cor eTypes selectors.html)

[EMAIL PROTECTED] wrote: I have declared a policy on what we will name new properties. Is this a good formal policy to have? If new properties get added (it happens), expect them to appear under the "ant." and "org.apache.tools.ant" prefixes, unless the developers have a very good

Re: cvs commit: ant/src/main/org/apache/tools/ant Project.java

Would it not be better to do something like this: if (loggingMessage) { PrintStream tempErr = new PrintStream (new FileOutputStream (FileDescriptor.err)); tempErr.println (message); } That way the message is not swallowed & lost? -- Jeffrey E. Care ([EMAIL PROTECTED]) WebSpher

Re: Property name policy (was Re: cvs commit: ant/docs/manual/CoreTypes selectors.html)

+1 on the general policy for "ant.*" and "org.apache.tools.ant.*" -1 for "build.*"; I'm sure there are many projects that use "build.this" or "build.that" -- Jeffrey E. Care ([EMAIL PROTECTED]) WebSphere Build SWAT Team Lead WebSphere Build Tooling Lead (Project Mantis) https://w3.opensource.ibm

DO NOT REPLY [Bug 33105] New: - Provide an if/unless optional attribute with all tasks

DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG· RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT . ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND· INSERTED IN THE BUG DATABASE. http://issues.apache.org/bugzilla/show_bu

DO NOT REPLY [Bug 33105] - Provide an if/unless optional attribute with all tasks

DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG· RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT . ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND· INSERTED IN THE BUG DATABASE. http://issues.apache.org/bugzilla/show_bu