On Thu, 2008-01-31 at 19:21 +0000, Mark Thomas wrote: > [EMAIL PROTECTED] wrote: > > Modified: tomcat/tc6.0.x/trunk/STATUS.txt > > URL: > > http://svn.apache.org/viewvc/tomcat/tc6.0.x/trunk/STATUS.txt?rev=616973&r1=616972&r2=616973&view=diff > > ============================================================================== > > --- tomcat/tc6.0.x/trunk/STATUS.txt (original) > > +++ tomcat/tc6.0.x/trunk/STATUS.txt Wed Jan 30 17:12:44 2008 > > @@ -43,33 +43,37 @@ > > Patch provided by Anthony Berglas > > http://svn.apache.org/viewvc?rev=616556&view=rev > > +1: markt > > - -1: > > + -1: remm: I didn't test the performance issue. If it exists, then the > > patch could be improved > > + (move any check logic to the Compiler class - and maybe try to > > handle JAR updates ?). > I'll look at options for moving the logic. I did look at handling JAR > updates and it was looking as if lots of changes would be required. Is > there any reason to handle them for this case when for the general case we > just reload the app?
Moving the check to where the other tests are done in Compiler seems easy. I don't understand the reload strategy: reloading won't clean up the work folder, you would need to redeploy. > > * Fix http://issues.apache.org/bugzilla/show_bug.cgi?id=43878 > > Use a single class loader when not reloading JSPs and tag files to > > reduce perm > > gen usage. > > http://svn.apache.org/viewvc?rev=616563&view=rev > > +1: markt > > + +0: remm: The patch will now no longer blow up with the background > > reloading thread, > > + but the use case intersects with precompilation, so it's > > probably not very useful. > Fair point. If that is the general consensus I'll revert it and mark the > bug as WONTFIX. It's +0 in the comment. > > * Fix http://issues.apache.org/bugzilla/show_bug.cgi?id=44282 > > Do call to getClassLoader() in a privileged block. > > http://svn.apache.org/viewvc?rev=616953&view=rev > > +1: markt > > + +0: remm: do we really want to fix these sort of "bugs" ? > Yes. It might be trivial but it is still a bug. With this sort of fix ? Ok, it's not much added resources, but anything seems a waste when you consider the issue. Rémy --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
