diff options
Diffstat (limited to 'docs/dist/doc/README-1919.html')
| -rw-r--r-- | docs/dist/doc/README-1919.html | 57 |
1 files changed, 26 insertions, 31 deletions
diff --git a/docs/dist/doc/README-1919.html b/docs/dist/doc/README-1919.html index a146c53a7..9861baf97 100644 --- a/docs/dist/doc/README-1919.html +++ b/docs/dist/doc/README-1919.html @@ -42,24 +42,24 @@ <li>Structured concurrency (incubator)</li> </ul> -Please note that the upstream Eclipse Java Compiler (ECJ) which the AspectJ Compiler (AJC) is a fork of still has some -open issues concerning Java 19 preview feature support, see the list in -<a href="https://github.com/eclipse/org.aspectj/issues/184#issuecomment-1272254940">this comment</a>. AJC therefore -inherits the same problems for the specific cases described in the linked issues. +<p> + Please note that the upstream Eclipse Java Compiler (ECJ) which the AspectJ Compiler (AJC) is a fork of still has some + open issues concerning Java 19 preview feature support, see the list in + <a href="https://github.com/eclipse/org.aspectj/issues/184#issuecomment-1272254940">this comment</a>. AJC therefore + inherits the same problems for the specific cases described in the linked issues. +</p> <h2>Improvements</h2> -<p> - In annotation style aspects, asynchronous <tt>proceed()</tt> calls in <tt>@Around</tt> advice now works in threads - created from within the advice. Previously, this was only working in native syntax aspects. There is still a - limitation with regard to asynchronous proceed, if you do not create the thread in the advice but want to use e.g. - an <tt>ExecutorService</tt> with its own thread pool. This still is not working in annotation style aspects, only in - native syntax ones. -</p> -<p> - See <a href="https://github.com/eclipse/org.aspectj/issues/128">issue #128</a> and - <a href="https://github.com/eclipse/org.aspectj/pull/132">pull request #132</a> for more details. -</p> +<ul> + <li> + Improve condy (constant dynamic) support. Together with some custom compilation or weaving options, this helps to + avoid a problem when using JaCoCo together with AspectJ, see + <a href="https://github.com/eclipse/org.aspectj/issues/170#issuecomment-1214163297">this comment in #170</a> for + more details. + </li> +</ul> + <h2>Code examples</h2> <p> @@ -78,16 +78,11 @@ inherits the same problems for the specific cases described in the linked issues </a> </li> <li> - <a href="https://github.com/eclipse/org.aspectj/tree/master/tests/features199/java18"> + <a href="https://github.com/eclipse/org.aspectj/tree/master/tests/features1919/java19"> Pattern matching for switch (preview 3), record patterns (preview 1) </a> </li> <li> - <a href="https://github.com/eclipse/org.aspectj/tree/master/tests/bugs199/github_128"> - Asynchronous proceed in native vs. annotation style syntax - </a> - </li> - <li> Please note that presently there is no specific sample code for virtual threads and structured concurrency in the AspectJ code base, because these are just new APIs, no Java language features. You can find sample code for these concurrency features elsewhere, e.g. in the corresponding JEPs. In AspectJ, they should just work transparently @@ -99,17 +94,17 @@ inherits the same problems for the specific cases described in the linked issues <ul> <li> - Fix a bug which led to <tt>NullPointerException</tt>s if too many JAR archives were on the classpath. Too many here - means the value system property <tt>org.aspectj.weaver.openarchives</tt> (1,000 by default). The AspectJ compiler is - meant to close archives upon cache exhaustion and then re-open them if it needs them again later. Re-opening was - broken, now the compiler works reliably even for cache sizes as small as 20. See issue - <a href="https://github.com/eclipse/org.aspectj/issues/125">#125</a>. + Fix (or rather work around) an old bug occurring when compiling or weaving code using ITD to declare annotations + with <tt>SOURCE</tt> retention on types, methods, constructors or fields. While declaring such annotations does not + make sense to begin with, at least the AspectJ weaver or compiler should handle the situation gracefully, which now + it does by simply ignoring errors caused by it. See <a href="https://bugs.eclipse.org/bugs/show_bug.cgi?id=366085"> + Bugzilla #366085</a> and <a href="https://github.com/eclipse/org.aspectj/pull/196">pull request #196</a>. Better + than this workaround would be for the compiler or weaver to actually print a warning when meeting source level + annotations in declare statements. Hence, follow-up issue + <a href="https://github.com/eclipse/org.aspectj/issues/201">#201</a> was created. </li> <li> - Improvements for <tt>if()</tt> pointcuts in annotation syntax, see issues - <a href="https://github.com/eclipse/org.aspectj/issues/115">#115</a>, - <a href="https://github.com/eclipse/org.aspectj/issues/120">#120</a>, - <a href="https://github.com/eclipse/org.aspectj/issues/122">#122</a>. + Remove legacy AspectJ Browser code and documentation. </li> <li> Thanks to Andrey Turbanov for several clean code contributions. @@ -155,7 +150,7 @@ inherits the same problems for the specific cases described in the linked issues <hr> <p> - <b>Available:</b> 1.9.19 on DD-MMM-2022 + <b>Available:</b> 1.9.19 on 21-Dec-2022 </p> </body> |