Re: [External] : Re: Very poor code indexing performances
Hi, I'm sorry that the mentioned PR caused some errors for files without a corresponding parser (thanks for fixing it), but I'm wondering why this particular PR should harm the indexing performance so much? To solve the problem, is it possible to start NB with a fresh userdir, invoke the "Profile the IDE (Alt+Shift+Y)" action while the initial code indexing is in progress and share the generated profiler snapshot? It should show us what is going on and perhaps point to the particular bottleneck. Thanks, Dusan On 5/25/22 6:01 PM, Matthias Bläsing wrote: Hi, this might be related: https://urldefense.com/v3/__https://github.com/apache/netbeans/pull/3516__;!!ACWV5N9M2RV99hQ!KnspVMNRiQEALD0gHmymGj_vmEmLt0G8eyHVoGRmwCTvsK048VPbYu13bBG33y3fV1s0mLexiv_aETsNloFjEcoivWc$ I also had the feeling, that performance suffered from it, but Dusan neither reacted to a direct fix for things, that were broken by that PR, nor to later requests for other performance fixes. Might be worth trying to revert that and see if performance improves (this might need revert too as it was the followup fix: https://urldefense.com/v3/__https://github.com/apache/netbeans/pull/3583__;!!ACWV5N9M2RV99hQ!KnspVMNRiQEALD0gHmymGj_vmEmLt0G8eyHVoGRmwCTvsK048VPbYu13bBG33y3fV1s0mLexiv_aETsNloFjN3vvYRs$ ) If reverting the PR helps, we should consider that. Greetings Matthias Am Mittwoch, dem 25.05.2022 um 15:25 + schrieb Jean-Marc Borer: Hello, Since NB 13 and Java 17, I am experiencing again poor code indexing performances as it used to be with older version of NB. 12.4 was running fine. I checked the IDE logs and see only entries like: INFO [null]: Last record repeated again. WARNING [org.netbeans.modules.maven.nodes.DependenciesNode]: Could not determine module name for artifact null WARNING [null]: Last record repeated 2 more times. INFO [org.netbeans.ui.metrics.debugger]: Debugger session crystal-position-application-acc-nimbus started at localhost for Java INFO [org.netbeans.modules.git.status]: GitInterceptor.refreshTask: Scanning in progress, trying again in 10?000ms INFO [null]: Last record repeated again. INFO [org.netbeans.modules.subversion.FileStatusCache]: FileStatusCache.refreshTask: Scanning in progress, trying again in 10?000ms INFO [org.netbeans.modules.git.status]: GitInterceptor.refreshTask: Scanning in progress, trying again in 10?000ms INFO [org.netbeans.modules.subversion.FileStatusCache]: FileStatusCache.refreshTask: Scanning in progress, trying again in 10?000ms INFO [org.netbeans.modules.git.status]: GitInterceptor.refreshTask: Scanning in progress, trying again in 10?000ms INFO [org.netbeans.modules.subversion.FileStatusCache]: FileStatusCache.refreshTask: Scanning in progress, trying again in 10?000ms INFO [org.netbeans.modules.git.status]: GitInterceptor.refreshTask: Scanning in progress, trying again in 10?000ms INFO [org.netbeans.modules.subversion.FileStatusCache]: FileStatusCache.refreshTask: Scanning in progress, trying again in 10?000ms INFO [org.netbeans.modules.git.status]: GitInterceptor.refreshTask: Scanning in progress, trying again in 10?000ms WARNING [org.netbeans.modules.java.source.parsing.VanillaPartialReparser]: Javac returned startpos: 17?726 > endpos: -1 I I have not idea what happens. Indeed I have a bunch of small JAXB generated files with my projects, but I used not to be an issue. I am running on a SSD disk. *Product Version:* Apache NetBeans IDE 13 *Java:* 17.0.2; OpenJDK 64-Bit Server VM 17.0.2+8-LTS *Runtime:* OpenJDK Runtime Environment 17.0.2+8-LTS *System:* Windows 10 version 10.0 running on amd64; Cp1252; en_US (nb) *User directory:* C:\Users\borerjc\AppData\Roaming\NetBeans\13 *Cache directory:* C:\projects\.netbeans\cache\13 Any help would be appreciated. Cheers, JMB - To unsubscribe, e-mail: dev-unsubscr...@netbeans.apache.org For additional commands, e-mail: dev-h...@netbeans.apache.org For further information about the NetBeans mailing lists, visit: https://urldefense.com/v3/__https://cwiki.apache.org/confluence/display/NETBEANS/Mailing*lists__;Kw!!ACWV5N9M2RV99hQ!KnspVMNRiQEALD0gHmymGj_vmEmLt0G8eyHVoGRmwCTvsK048VPbYu13bBG33y3fV1s0mLexiv_aETsNloFjpHgeRc8$ - To unsubscribe, e-mail: dev-unsubscr...@netbeans.apache.org For additional commands, e-mail: dev-h...@netbeans.apache.org For further information about the NetBeans mailing lists, visit: https://cwiki.apache.org/confluence/display/NETBEANS/Mailing+lists
Re: Very poor code indexing performances
Since talking performance…Just curious…is it conceivable to add some performance test for use during build time testing and gather metrics during a given build so with each build/release to get performance historical over time? Get Outlook for iOS<https://aka.ms/o0ukef> From: Michael Bien Sent: Wednesday, May 25, 2022 6:41:57 PM To: dev@netbeans.apache.org ; John Neffenger Subject: Re: Very poor code indexing performances On 26.05.22 01:18, John Neffenger wrote: > On 5/25/22 10:27 AM, Michael Bien wrote: >> async-profiler[1] combined with flame graphs can be helpful to find >> bottlenecks of long running tasks (> few seconds). > > +1 for async-profiler! It can let you get a much bigger picture than > is possible with most other profilers, and it often shows you > immediately what's wrong. For example: > > JDK 13 Performance > https://jgneff.github.io/framebufferY8/2020-05/#jdk-13-performance > > It would be a nice addition to the NetBeans profiler. :-) there is an issue for that (for flame graphs specifically) https://github.com/apache/netbeans/issues/3849 I suppose it would be nice to have an async-profiler plugin for discoverabiity reasons, but I personally would keep using it's CLI since it doesn't require any setup anyway. -mbien > > John > > - > To unsubscribe, e-mail: dev-unsubscr...@netbeans.apache.org > For additional commands, e-mail: dev-h...@netbeans.apache.org > > For further information about the NetBeans mailing lists, visit: > https://cwiki.apache.org/confluence/display/NETBEANS/Mailing+lists > > > - To unsubscribe, e-mail: dev-unsubscr...@netbeans.apache.org For additional commands, e-mail: dev-h...@netbeans.apache.org For further information about the NetBeans mailing lists, visit: https://cwiki.apache.org/confluence/display/NETBEANS/Mailing+lists
Re: Very poor code indexing performances
On 26.05.22 01:18, John Neffenger wrote: On 5/25/22 10:27 AM, Michael Bien wrote: async-profiler[1] combined with flame graphs can be helpful to find bottlenecks of long running tasks (> few seconds). +1 for async-profiler! It can let you get a much bigger picture than is possible with most other profilers, and it often shows you immediately what's wrong. For example: JDK 13 Performance https://jgneff.github.io/framebufferY8/2020-05/#jdk-13-performance It would be a nice addition to the NetBeans profiler. :-) there is an issue for that (for flame graphs specifically) https://github.com/apache/netbeans/issues/3849 I suppose it would be nice to have an async-profiler plugin for discoverabiity reasons, but I personally would keep using it's CLI since it doesn't require any setup anyway. -mbien John - To unsubscribe, e-mail: dev-unsubscr...@netbeans.apache.org For additional commands, e-mail: dev-h...@netbeans.apache.org For further information about the NetBeans mailing lists, visit: https://cwiki.apache.org/confluence/display/NETBEANS/Mailing+lists - To unsubscribe, e-mail: dev-unsubscr...@netbeans.apache.org For additional commands, e-mail: dev-h...@netbeans.apache.org For further information about the NetBeans mailing lists, visit: https://cwiki.apache.org/confluence/display/NETBEANS/Mailing+lists
Re: Very poor code indexing performances
On 5/25/22 10:27 AM, Michael Bien wrote: async-profiler[1] combined with flame graphs can be helpful to find bottlenecks of long running tasks (> few seconds). +1 for async-profiler! It can let you get a much bigger picture than is possible with most other profilers, and it often shows you immediately what's wrong. For example: JDK 13 Performance https://jgneff.github.io/framebufferY8/2020-05/#jdk-13-performance It would be a nice addition to the NetBeans profiler. :-) John - To unsubscribe, e-mail: dev-unsubscr...@netbeans.apache.org For additional commands, e-mail: dev-h...@netbeans.apache.org For further information about the NetBeans mailing lists, visit: https://cwiki.apache.org/confluence/display/NETBEANS/Mailing+lists
Re: Very poor code indexing performances
Hi, this might be related: https://github.com/apache/netbeans/pull/3516 I also had the feeling, that performance suffered from it, but Dusan neither reacted to a direct fix for things, that were broken by that PR, nor to later requests for other performance fixes. Might be worth trying to revert that and see if performance improves (this might need revert too as it was the followup fix: https://github.com/apache/netbeans/pull/3583) If reverting the PR helps, we should consider that. Greetings Matthias Am Mittwoch, dem 25.05.2022 um 15:25 + schrieb Jean-Marc Borer: > Hello, > > Since NB 13 and Java 17, I am experiencing again poor code indexing > performances as it used to be with older version of NB. 12.4 was running > fine. > > I checked the IDE logs and see only entries like: > INFO [null]: Last record repeated again. > WARNING [org.netbeans.modules.maven.nodes.DependenciesNode]: Could not > determine module name for artifact null > WARNING [null]: Last record repeated 2 more times. > INFO [org.netbeans.ui.metrics.debugger]: Debugger session > crystal-position-application-acc-nimbus started at localhost for Java > INFO [org.netbeans.modules.git.status]: GitInterceptor.refreshTask: > Scanning in progress, trying again in 10?000ms > INFO [null]: Last record repeated again. > INFO [org.netbeans.modules.subversion.FileStatusCache]: > FileStatusCache.refreshTask: Scanning in progress, trying again in 10?000ms > INFO [org.netbeans.modules.git.status]: GitInterceptor.refreshTask: > Scanning in progress, trying again in 10?000ms > INFO [org.netbeans.modules.subversion.FileStatusCache]: > FileStatusCache.refreshTask: Scanning in progress, trying again in 10?000ms > INFO [org.netbeans.modules.git.status]: GitInterceptor.refreshTask: > Scanning in progress, trying again in 10?000ms > INFO [org.netbeans.modules.subversion.FileStatusCache]: > FileStatusCache.refreshTask: Scanning in progress, trying again in 10?000ms > INFO [org.netbeans.modules.git.status]: GitInterceptor.refreshTask: > Scanning in progress, trying again in 10?000ms > INFO [org.netbeans.modules.subversion.FileStatusCache]: > FileStatusCache.refreshTask: Scanning in progress, trying again in 10?000ms > INFO [org.netbeans.modules.git.status]: GitInterceptor.refreshTask: > Scanning in progress, trying again in 10?000ms > WARNING [org.netbeans.modules.java.source.parsing.VanillaPartialReparser]: > Javac returned startpos: 17?726 > endpos: -1 > I > I have not idea what happens. Indeed I have a bunch of small JAXB generated > files with my projects, but I used not to be an issue. > > I am running on a SSD disk. > > *Product Version:* Apache NetBeans IDE 13 > > *Java:* 17.0.2; OpenJDK 64-Bit Server VM 17.0.2+8-LTS > > *Runtime:* OpenJDK Runtime Environment 17.0.2+8-LTS > > *System:* Windows 10 version 10.0 running on amd64; Cp1252; en_US (nb) > > *User directory:* C:\Users\borerjc\AppData\Roaming\NetBeans\13 > > *Cache directory:* C:\projects\.netbeans\cache\13 > > > Any help would be appreciated. > > > Cheers, > > > JMB - To unsubscribe, e-mail: dev-unsubscr...@netbeans.apache.org For additional commands, e-mail: dev-h...@netbeans.apache.org For further information about the NetBeans mailing lists, visit: https://cwiki.apache.org/confluence/display/NETBEANS/Mailing+lists
Very poor code indexing performances
Hello, Since NB 13 and Java 17, I am experiencing again poor code indexing performances as it used to be with older version of NB. 12.4 was running fine. I checked the IDE logs and see only entries like: INFO [null]: Last record repeated again. WARNING [org.netbeans.modules.maven.nodes.DependenciesNode]: Could not determine module name for artifact null WARNING [null]: Last record repeated 2 more times. INFO [org.netbeans.ui.metrics.debugger]: Debugger session crystal-position-application-acc-nimbus started at localhost for Java INFO [org.netbeans.modules.git.status]: GitInterceptor.refreshTask: Scanning in progress, trying again in 10?000ms INFO [null]: Last record repeated again. INFO [org.netbeans.modules.subversion.FileStatusCache]: FileStatusCache.refreshTask: Scanning in progress, trying again in 10?000ms INFO [org.netbeans.modules.git.status]: GitInterceptor.refreshTask: Scanning in progress, trying again in 10?000ms INFO [org.netbeans.modules.subversion.FileStatusCache]: FileStatusCache.refreshTask: Scanning in progress, trying again in 10?000ms INFO [org.netbeans.modules.git.status]: GitInterceptor.refreshTask: Scanning in progress, trying again in 10?000ms INFO [org.netbeans.modules.subversion.FileStatusCache]: FileStatusCache.refreshTask: Scanning in progress, trying again in 10?000ms INFO [org.netbeans.modules.git.status]: GitInterceptor.refreshTask: Scanning in progress, trying again in 10?000ms INFO [org.netbeans.modules.subversion.FileStatusCache]: FileStatusCache.refreshTask: Scanning in progress, trying again in 10?000ms INFO [org.netbeans.modules.git.status]: GitInterceptor.refreshTask: Scanning in progress, trying again in 10?000ms WARNING [org.netbeans.modules.java.source.parsing.VanillaPartialReparser]: Javac returned startpos: 17?726 > endpos: -1 I I have not idea what happens. Indeed I have a bunch of small JAXB generated files with my projects, but I used not to be an issue. I am running on a SSD disk. *Product Version:* Apache NetBeans IDE 13 *Java:* 17.0.2; OpenJDK 64-Bit Server VM 17.0.2+8-LTS *Runtime:* OpenJDK Runtime Environment 17.0.2+8-LTS *System:* Windows 10 version 10.0 running on amd64; Cp1252; en_US (nb) *User directory:* C:\Users\borerjc\AppData\Roaming\NetBeans\13 *Cache directory:* C:\projects\.netbeans\cache\13 Any help would be appreciated. Cheers, JMB
