Community
Participate
Working Groups
I have been using Mars and Luna on a set of m2e/wtp/egit/java8 projects (different workspace, but same project files on disk). When using Luna the incremental build is always fast and happens in the background (unnoticeable). The same incremental build in Mars takes 3+ minutes, causes a blocking dialog to appear with build status, and is very frustrating. I turned on -debug for Mars and found the time is consumed in the MavenBuilder. While I haven't double checked it seems this only occurs on a project that has 6k resources. Although both Luna and Mars build output say they are processing 6k resources, the Luna build finishes in seconds. What are the next steps for providing the information needed to solve the problem? How can I help?
Newer version of maven embedded in m2e requires more memory in some scenarios (a known Maven problem/regression). See if eclipse jvm heap usage is running close to -Xmx and increase max heap size if it is. Beyond that, we'll need a standalone example project and exact steps we can use to reproduce the problem. Ideally, the test project should work with m2e installed on eclipse sdk without any additional components. Specifically, if m2e-wtp/wtp are required to reproduce the problem I will move the bug to m2e-wtp for further investigation. If you are interested (or prefer) to profile m2e yourself, I will be happy to provide pointers on m2e-dev mailing list.
Just tried it with -Xmx2048 in Eclipse.ini. This did not improve the performance and I noticed the heap never really gets beyond 300-400m. I also had the "Progress" view open and noticed the build went through every item in /src/main/resources and listed "Refresh: /src/main/sources/abc/xyz.ext" for about 1 second per file. Might be easiest for me to do a profile and if the root cause is not immediately obvious then I can work on a standalone reproduction project.
I am having difficulties getting a profile going connecting visualvm straight to Eclipse. Any suggestions?
Created attachment 252753 [details] screenshot What os and java do you use? I've had no problems profiling eclipse when running oracle/sun vms.
oracle jdk 1.8 on OSX... it may eventually complete but so far the initial instrumentation from visualvm is crashing at worst and hanging at best. Possibly related to not being able to detect what class to start profiling from (visualvm can't auto detect). I am trying org.eclipse.m2e.** and see if that helps it get going (was trying ** before)
I can confirm the problem. Incremental compilation after any code change: Luna: 1-2seconds or less. Mars: 45minutes Neon M3: 40minutes Platform: OSX, Oracle JDK8 Project has 3 packages with huge number of classes (mostly generated code). Lambdas are actively used. Because of this issue I forced to use IDEA and I really hate it.
(In reply to Fedor Ilinykh from comment #6) > I can confirm the problem. > > Incremental compilation after any code change: > > Luna: 1-2seconds or less. > Mars: 45minutes > Neon M3: 40minutes > > Platform: OSX, Oracle JDK8 > > Project has 3 packages with huge number of classes (mostly generated code). > Lambdas are actively used. > > Because of this issue I forced to use IDEA and I really hate it. I also encountered this problem, do you solve? thank you.
(In reply to think gem from comment #7) > (In reply to Fedor Ilinykh from comment #6) > > I can confirm the problem. > > > > Incremental compilation after any code change: > > > > Luna: 1-2seconds or less. > > Mars: 45minutes > > Neon M3: 40minutes > > > > Platform: OSX, Oracle JDK8 > > > > Project has 3 packages with huge number of classes (mostly generated code). > > Lambdas are actively used. > > > > Because of this issue I forced to use IDEA and I really hate it. > > I also encountered this problem, do you solve? thank you. I have found the reason, because a compilation phase yui a plug-in.But there is no this kind of circumstance Luna. thank you.
This bug hasn't had any activity in quite some time. Maybe the problem got resolved, was a duplicate of something else, or became less pressing for some reason - or maybe it's still relevant but just hasn't been looked at yet. As such, we're closing this bug. If you have further information on the current state of the bug, please add it and reopen this bug. The information can be, for example, that the problem still occurs, that you still want the feature, that more information is needed, or that the bug is (for whatever reason) no longer relevant. -- The automated Eclipse Genie.
Moved to https://github.com/eclipse-m2e/m2e-core/issues/
