Re: Release plans
On 04/10/2022 03:53, Christopher Schultz wrote: Mark, On 10/3/22 03:33, Mark Thomas wrote: Hi all, Given the regression in the previous set of releases and that it is the beginning of October, my intention is to start the October release round shortly. I have a couple of things I want to look at before I tag the releases as well as running my usual checks with the unit tests. Hopefully, I'll be in a position to tag later today. I believe the other RMs are at ApacheCon this week so - unless there are objections - I'm intending to tag and release the full set of releases myself: 10.1.x, 10.0.x, 9.0.x and 8.5.x. Thanks for doing this. If I had seen it earlier today, I would have done the 8.5 release. No problem. Rémy just arrived in NOLA and we had dinner together instead of doing these releases :) I am very jealous. Hope you all have a great conference. Mark - To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
[Bug 64775] mod_jk is sending both Content-Length and Transfer-Encoding
https://bz.apache.org/bugzilla/show_bug.cgi?id=64775 --- Comment #5 from Mark Thomas --- If the application is adding an content-length header then this is primarily an application issue. We could clean this up in Tomcat and/or mod_jk but generally the expectation is that applications are well-behaved and we don't validate the headers they set. -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
[Bug 66294] New: Util.getContextClassLoader() can be a hotspot
https://bz.apache.org/bugzilla/show_bug.cgi?id=66294 Bug ID: 66294 Summary: Util.getContextClassLoader() can be a hotspot Product: Tomcat 9 Version: 9.0.65 Hardware: All OS: Linux Status: NEW Severity: regression Priority: P2 Component: EL Assignee: dev@tomcat.apache.org Reporter: mxhen...@amazon.de Target Milestone: - Created attachment 38401 --> https://bz.apache.org/bugzilla/attachment.cgi?id=38401&action=edit Benchmark which shows performance difference between getting class loader with/without privileges In our application Util.getContextClassLoader() is a hotspot with almost all the cost coming from AccessController.doPrivileged. By adding instrumentation we've confirmed that the native implementation of AccessController.doPrivileged itself and not PrivilegedGetTccl.run() is expensive. In an example request to our application we had 17000 calls to Util.getContextClassLoader() and recorded this CPU time in the methods: 0.682 ms for javax.el.Util$PrivilegedGetTccl.run() 27.800 ms for javax.el.Util.getContextClassLoader() This potential hotspot was introduced four years ago when a change was made to get the context class loader of the Thread with privileges enabled: https://bz.apache.org/bugzilla/show_bug.cgi?id=62080 We'd like to fix this hotspot by calling Thread.currentThread().getContextClassLoader() without privileges enabled again. This would be a revert of the change here: https://github.com/apache/tomcat/commit/7c359957f0b600b92e964b1a2497374e6cac4bc3#diff-f5b7c13f66b3d070a6b1c1713ad4b60fd70b26f579a746edc8fe9f19109243b2L93 Since this change addressed a rare bug, we probably have to make this change backwards compatible by adding a configuration switch. I have also included a basic benchmark which compares the performance of the two approaches. The benchmark shows a difference by factor 100. I ran the benchmark with JDK 8 and 11. Since SecurityManager and related functionality are on track to deprecation, this change may be mandatory in the future, irrespective of the performance benefit: https://openjdk.org/jeps/411 -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
Re: [tomcat] branch 9.0.x updated: I'm going to RM 9.0.68
On Mon, Oct 3, 2022 at 9:05 PM wrote: > > This is an automated email from the ASF dual-hosted git repository. > > markt pushed a commit to branch 9.0.x > in repository https://gitbox.apache.org/repos/asf/tomcat.git > > > The following commit(s) were added to refs/heads/9.0.x by this push: > new 3356e003c3 I'm going to RM 9.0.68 > 3356e003c3 is described below > > commit 3356e003c3b3a0d139ad2ac2355258542a0023fc > Author: Mark Thomas > AuthorDate: Mon Oct 3 20:04:44 2022 +0100 > > I'm going to RM 9.0.68 Thanks for backing me up ! In the end I could have done it today, but you never know for certain when you're travelling ;) Remy > --- > webapps/docs/changelog.xml | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/webapps/docs/changelog.xml b/webapps/docs/changelog.xml > index d27ed6b481..35afdb94a8 100644 > --- a/webapps/docs/changelog.xml > +++ b/webapps/docs/changelog.xml > @@ -104,7 +104,7 @@ >They eventually become mixed with the numbered issues (i.e., numbered >issues do not "pop up" wrt. others). > --> > - > + > > > > > > - > To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org > For additional commands, e-mail: dev-h...@tomcat.apache.org > - To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
Re: Release plans
On Tue, Oct 4, 2022 at 9:07 AM Mark Thomas wrote: > > > > On 04/10/2022 03:53, Christopher Schultz wrote: > > Mark, > > > > On 10/3/22 03:33, Mark Thomas wrote: > >> Hi all, > >> > >> Given the regression in the previous set of releases and that it is > >> the beginning of October, my intention is to start the October release > >> round shortly. > >> > >> I have a couple of things I want to look at before I tag the releases > >> as well as running my usual checks with the unit tests. Hopefully, > >> I'll be in a position to tag later today. > >> > >> I believe the other RMs are at ApacheCon this week so - unless there > >> are objections - I'm intending to tag and release the full set of > >> releases myself: 10.1.x, 10.0.x, 9.0.x and 8.5.x. > > > > Thanks for doing this. If I had seen it earlier today, I would have done > > the 8.5 release. > > No problem. > > > Rémy just arrived in NOLA and we had dinner together instead of doing > > these releases :) > > I am very jealous. Hope you all have a great conference. There's one (unexpected) problem though: there is no recording at all. I didn't expect it (it's 2022 now ;) ) and there was no notice about this, so I don't have any equipment with me (but I have 100% and more at home :D ). Chris will do audio. I might try the teacup phone holder thing again for my talks. Remy > Mark > > - > To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org > For additional commands, e-mail: dev-h...@tomcat.apache.org > - To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
[Bug 66279] Tomcat
https://bz.apache.org/bugzilla/show_bug.cgi?id=66279 --- Comment #4 from Christopher Schultz --- Also, can we please get a better description than "Tomcat"? -- You are receiving this mail because: You are the assignee for the bug. - To unsubscribe, e-mail: dev-unsubscr...@tomcat.apache.org For additional commands, e-mail: dev-h...@tomcat.apache.org
