diff options
Diffstat (limited to 'docs/release/README-1.6.7.adoc')
| -rw-r--r-- | docs/release/README-1.6.7.adoc | 10 |
1 files changed, 5 insertions, 5 deletions
diff --git a/docs/release/README-1.6.7.adoc b/docs/release/README-1.6.7.adoc index 464f34006..c2b102027 100644 --- a/docs/release/README-1.6.7.adoc +++ b/docs/release/README-1.6.7.adoc @@ -13,7 +13,7 @@ one of your pointcuts that is hurting your weaving performance. In 1.6.7 it is possible to turn on timers for pointcuts. These timers show the time spent in the weaver matching the pointcut components against join points. The details on this feature are here: -http://andrewclement.blogspot.com/2009/11/aspectj-profiling-pointcut-matching.html[Profiling +https://andrewclement.blogspot.com/2009/11/aspectj-profiling-pointcut-matching.html[Profiling pointcut matching]. Basically by turning on the options '-timers -verbose' on the command line (or via Ant), output will be produced that looks a little like this: @@ -36,7 +36,7 @@ Time:4ms (jps:#662) matching against It shows the component, the number of joinpoints (jps) the weaver attempted to match it against and how many milliseconds were spent performing those matches. The options can also be turned on -http://contraptionsforprogramming.blogspot.com/2009/11/getting-aspectj-pointcut-matching-timer.html[]through +https://contraptionsforprogramming.blogspot.com/2009/11/getting-aspectj-pointcut-matching-timer.html[]through AJDT. Armed with this information you can optimize your pointcuts or post on the mailing list asking for help. The timers can even be turned on for load time weaving. @@ -48,7 +48,7 @@ be done on the matching algorithms. These have remained unchanged for a few years, but in 1.6.7 have received a bit of an overhaul. 'Fast match' has been implemented for the execution() pointcut, drastically reducing weave times for heavy users of execution - more details -http://andrewclement.blogspot.com/2009/11/aspectj-how-much-faster-is-aspectj-167.html[here]. +https://andrewclement.blogspot.com/2009/11/aspectj-how-much-faster-is-aspectj-167.html[here]. The pointcut cost calculator (which is used to sort pointcuts to optimize matching speed) has been reviewed and after determining that this() ought to be considered cheaper than call() - any user combining @@ -67,7 +67,7 @@ The processing of include/exclude entries in aop.xml has been rewritten. It now optimizes for many more common patterns. If a pattern is optimized then there is no need to ask the weaver to do an expensive include/exclude match. More details -http://andrewclement.blogspot.com/2009/12/aspectj-167-and-faster-load-time.html[here]. +https://andrewclement.blogspot.com/2009/12/aspectj-167-and-faster-load-time.html[here]. == Less need to tweak options for load time weaving @@ -87,7 +87,7 @@ will now do the right thing out of the box. All those changes above, and some additional tweaks, mean we are now using less memory than ever before and getting things done more quickly. -http://andrewclement.blogspot.com/2009/12/aspectj-167-and-faster-load-time.html[This +https://andrewclement.blogspot.com/2009/12/aspectj-167-and-faster-load-time.html[This post] discusses the details. From that article, the graph below shows the speed and memory consumption of the various AspectJ 1.6 releases when load time weaving a small application loading in Tomcat. For each |