Re: svn commit: r1696822 - in /velocity/tools/trunk: src/site/xdoc/ velocity-tools-struts/src/main/java/org/apache/velocity/tools/struts/ velocity-tools-view/src/main/java/org/apache/velocity/tools/vi

2015-08-31 Thread Christopher Schultz
Claude, On 8/20/15 3:12 PM, Claude Brisson wrote: > Yes, we should be using Velocity logging facade, but I just don't see > for now how we should do this while preserving tools serialization. A little late to the (logging) party, and this may be a bit off-topic, but why bother using a

Re: svn commit: r1696822 - in /velocity/tools/trunk: src/site/xdoc/ velocity-tools-struts/src/main/java/org/apache/velocity/tools/struts/ velocity-tools-view/src/main/java/org/apache/velocity/tools/vi

2015-08-31 Thread Mike Kienenberger
Chris, Read the whole thread. My understanding is that serialization of static commons-logging was the original problem needing fixed. slf4j doesn't have that issue. On Mon, Aug 31, 2015 at 1:41 PM, Christopher Schultz wrote: > Claude, > > On 8/20/15 3:12 PM,

Re: svn commit: r1696822 - in /velocity/tools/trunk: src/site/xdoc/ velocity-tools-struts/src/main/java/org/apache/velocity/tools/struts/ velocity-tools-view/src/main/java/org/apache/velocity/tools/vi

2015-08-31 Thread Christopher Schultz
Mike, On 8/31/15 1:48 PM, Mike Kienenberger wrote: > Read the whole thread. Yeah, I'm catching-up now. > My understanding is that serialization of static commons-logging was > the original problem needing fixed. slf4j doesn't have that issue. I didn't think that static members were serialized.

Re: svn commit: r1696822 - in /velocity/tools/trunk: src/site/xdoc/ velocity-tools-struts/src/main/java/org/apache/velocity/tools/struts/ velocity-tools-view/src/main/java/org/apache/velocity/tools/vi

2015-08-31 Thread Claude Brisson
On 31/08/2015 19:52, Christopher Schultz wrote: Mike, On 8/31/15 1:48 PM, Mike Kienenberger wrote: Read the whole thread. Yeah, I'm catching-up now. My understanding is that serialization of static commons-logging was the original problem needing fixed. slf4j doesn't have that issue. I

Re: svn commit: r1696822 - in /velocity/tools/trunk: src/site/xdoc/ velocity-tools-struts/src/main/java/org/apache/velocity/tools/struts/ velocity-tools-view/src/main/java/org/apache/velocity/tools/vi

2015-08-31 Thread Claude Brisson
I should add that it implies shipping a new major version for velocity-tools. I'm not against it, as we can take the opportunity to drop the deprecated package o.a.v.tools.view.tools. Also, do we maintain the maven-velocity-tools-plugin subproject? Right now, it fails with: [ERROR] Failed

Re: svn commit: r1696822 - in /velocity/tools/trunk: src/site/xdoc/ velocity-tools-struts/src/main/java/org/apache/velocity/tools/struts/ velocity-tools-view/src/main/java/org/apache/velocity/tools/vi

2015-08-30 Thread Claude Brisson
Some remarks after a little investigation. 1) Using the ServletContext logger doesn't seem to be a widespread practice among slf4j folks. There is a slf4j-servletcontext implementation on github, but it is not referenced on the official slf4j site. Moreover, I tested it, and it does not

Re: svn commit: r1696822 - in /velocity/tools/trunk: src/site/xdoc/ velocity-tools-struts/src/main/java/org/apache/velocity/tools/struts/ velocity-tools-view/src/main/java/org/apache/velocity/tools/vi

2015-08-30 Thread Nathan Bubna
Sounds reasonable to me. Thanks for tackling this, Claude! On Sun, Aug 30, 2015 at 9:07 AM, Claude Brisson cla...@renegat.net wrote: Some remarks after a little investigation. 1) Using the ServletContext logger doesn't seem to be a widespread practice among slf4j folks. There is a

Re: svn commit: r1696822 - in /velocity/tools/trunk: src/site/xdoc/ velocity-tools-struts/src/main/java/org/apache/velocity/tools/struts/ velocity-tools-view/src/main/java/org/apache/velocity/tools/vi

2015-08-30 Thread Sergiu Dumitriu
Sorry, but I'm not sure I understand what you're proposing. Can you make a short summary? On 08/30/2015 12:07 PM, Claude Brisson wrote: Some remarks after a little investigation. 1) Using the ServletContext logger doesn't seem to be a widespread practice among slf4j folks. There is a

Re: svn commit: r1696822 - in /velocity/tools/trunk: src/site/xdoc/ velocity-tools-struts/src/main/java/org/apache/velocity/tools/struts/ velocity-tools-view/src/main/java/org/apache/velocity/tools/vi

2015-08-30 Thread Claude Brisson
Of course: switching the log system to slf4j, which means: - getting rid of the org.apache.velocity.runtime.log package - kepping loggers injection as it is, but switch to org.slf4j.Logger - getting rid of the four log-realted sub-modules, that is velocity-engine-commons-logging,

Re: svn commit: r1696822 - in /velocity/tools/trunk: src/site/xdoc/ velocity-tools-struts/src/main/java/org/apache/velocity/tools/struts/ velocity-tools-view/src/main/java/org/apache/velocity/tools/vi

2015-08-21 Thread Nathan Bubna
I like simple. :) On Fri, Aug 21, 2015 at 2:18 AM, Claude Brisson cla...@renegat.net wrote: It looks like that with slf4j, we could keep loggers injection even in session tools, so no need to go static: As of SLF4J version 1.5.3, logger instances survive serialization. Thus, serialization

Re: svn commit: r1696822 - in /velocity/tools/trunk: src/site/xdoc/ velocity-tools-struts/src/main/java/org/apache/velocity/tools/struts/ velocity-tools-view/src/main/java/org/apache/velocity/tools/vi

2015-08-21 Thread Claude Brisson
It looks like that with slf4j, we could keep loggers injection even in session tools, so no need to go static: As of SLF4J version 1.5.3, logger instances survive serialization. Thus, serialization of the host class no longer requires any special action, even when loggers are declared as

Re: svn commit: r1696822 - in /velocity/tools/trunk: src/site/xdoc/ velocity-tools-struts/src/main/java/org/apache/velocity/tools/struts/ velocity-tools-view/src/main/java/org/apache/velocity/tools/vi

2015-08-20 Thread Sergiu Dumitriu
Isn't the logging framework supposed to be pluggable, meaning that all calls go through org.apache.velocity.runtime.log.Log which is then passed to the actual logging library in use, be it log4j, slf4j, jcl or commons-logging? On 08/20/2015 01:16 PM, cbris...@apache.org wrote: Author: cbrisson

Re: svn commit: r1696822 - in /velocity/tools/trunk: src/site/xdoc/ velocity-tools-struts/src/main/java/org/apache/velocity/tools/struts/ velocity-tools-view/src/main/java/org/apache/velocity/tools/vi

2015-08-20 Thread Claude Brisson
Thanks for the review. Yes, we should be using Velocity logging facade, but I just don't see for now how we should do this while preserving tools serialization. For now I just mimicked a behavior present in some deprecated classes, re-introducing the dependency of velocity-tools-view on

Re: svn commit: r1696822 - in /velocity/tools/trunk: src/site/xdoc/ velocity-tools-struts/src/main/java/org/apache/velocity/tools/struts/ velocity-tools-view/src/main/java/org/apache/velocity/tools/vi

2015-08-20 Thread Sergiu Dumitriu
A recent discussion was proposing to switch to slf4j as the façade. On Aug 20, 2015 3:12 PM, Claude Brisson cla...@renegat.net wrote: Thanks for the review. Yes, we should be using Velocity logging facade, but I just don't see for now how we should do this while preserving tools