[
https://issues.apache.org/jira/browse/JCRVLT-809?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18008063#comment-18008063
]
Konrad Windszus edited comment on JCRVLT-809 at 7/18/25 12:47 PM:
------------------------------------------------------------------
[~joerghoh] Can you come up with a proposal?
I wouldn't know exactly how to speed up things as the applicable node type
depends upon
- property name
- property type
- multi-value or single-value
Maybe caching {{NodeType.getPropertyDefinitions()}} and
{{NodeType.getChildNodeDefinitions()}} inside {{EffectiveNodeType}} could help
here.
was (Author: kwin):
[~joerghoh] Can you come up with a proposal?
I wouldn't know exactly how to speed up things as the applicable node type
depends upon
- property name
- property type
- multi-value or single-value
> DocViewImporter: logIgnoredProtectedProperties can slow down package import
> ---------------------------------------------------------------------------
>
> Key: JCRVLT-809
> URL: https://issues.apache.org/jira/browse/JCRVLT-809
> Project: Jackrabbit FileVault
> Issue Type: Improvement
> Components: Packaging
> Reporter: Joerg Hoh
> Priority: Major
>
> In the case of Sling distribution the packages being created and imported are
> controlled by the framework, and IIUC protected properties are not part of
> the package; for that reason
> {{DocViewImporter.logIgnoredProtectedProperties}} will not write any WARN
> message.
> When checking the details of the import of a complex package (with almost
> 100k nodes and low to medium amount of properties per node), I came
> frequently across this type of stack, where it's checked if properties are
> protected. Commenting the 2 calls to
> {{DocViewImporter.logIgnoredProtectedProperties}} gave me an improvement of
> the import time in the range of 10%.
> For that I think that we should provide a chance to disable these checks,
> preferably via the API, so that it can configured per package I want to
> import. Disabling it globally might not be helpful, as we might have cases,
> where we cannot give the guarantee about no protected properties being part
> of the package(s).
> {noformat}
> at
> org.apache.commons.collections4.iterators.FilterIterator.hasNext(FilterIterator.java:105)
> at
> org.apache.jackrabbit.oak.plugins.nodetype.NodeTypeImpl.getDeclaredPropertyDefinitions(NodeTypeImpl.java:189)
> at
> org.apache.jackrabbit.oak.plugins.nodetype.NodeTypeImpl.internalGetPropertyDefinitions(NodeTypeImpl.java:521)
> at
> org.apache.jackrabbit.oak.plugins.nodetype.NodeTypeImpl.getPropertyDefinitions(NodeTypeImpl.java:316)
> at
> org.apache.jackrabbit.vault.util.EffectiveNodeType.lambda$getApplicablePropertyDefinition$1(EffectiveNodeType.java:111)
> at
> org.apache.jackrabbit.vault.util.EffectiveNodeType$$Lambda/0x000000080194cb00.apply(Unknown
> Source)
> at
> java.util.stream.ReferencePipeline$7$1.accept(java.base@21.0.2/ReferencePipeline.java:273)
> at
> java.util.ArrayList$ArrayListSpliterator.forEachRemaining(java.base@21.0.2/ArrayList.java:1708)
> at
> java.util.stream.AbstractPipeline.copyInto(java.base@21.0.2/AbstractPipeline.java:509)
> at
> java.util.stream.AbstractPipeline.wrapAndCopyInto(java.base@21.0.2/AbstractPipeline.java:499)
> at
> java.util.stream.ReduceOps$ReduceOp.evaluateSequential(java.base@21.0.2/ReduceOps.java:921)
> at
> java.util.stream.AbstractPipeline.evaluate(java.base@21.0.2/AbstractPipeline.java:234)
> at
> java.util.stream.ReferencePipeline.collect(java.base@21.0.2/ReferencePipeline.java:682)
> at
> org.apache.jackrabbit.vault.util.EffectiveNodeType.getApplicablePropertyDefinition(EffectiveNodeType.java:111)
> at
> org.apache.jackrabbit.vault.util.EffectiveNodeType.getApplicablePropertyDefinition(EffectiveNodeType.java:107)
> at
> org.apache.jackrabbit.vault.fs.impl.io.DocViewImporter.isPropertyProtected(DocViewImporter.java:1235)
> at
> org.apache.jackrabbit.vault.fs.impl.io.DocViewImporter.lambda$logIgnoredProtectedProperties$7(DocViewImporter.java:1217)
> at
> org.apache.jackrabbit.vault.fs.impl.io.DocViewImporter$$Lambda/0x000000080194c688.accept(Unknown
> Source)
> at
> java.util.stream.ForEachOps$ForEachOp$OfRef.accept(java.base@21.0.2/ForEachOps.java:184)
> at
> java.util.stream.ReferencePipeline$2$1.accept(java.base@21.0.2/ReferencePipeline.java:179)
> at
> java.util.Iterator.forEachRemaining(java.base@21.0.2/Iterator.java:133)
> at
> java.util.Spliterators$IteratorSpliterator.forEachRemaining(java.base@21.0.2/Spliterators.java:1939)
> at
> java.util.stream.AbstractPipeline.copyInto(java.base@21.0.2/AbstractPipeline.java:509)
> at
> java.util.stream.AbstractPipeline.wrapAndCopyInto(java.base@21.0.2/AbstractPipeline.java:499)
> at
> java.util.stream.ForEachOps$ForEachOp.evaluateSequential(java.base@21.0.2/ForEachOps.java:151)
> at
> java.util.stream.ForEachOps$ForEachOp$OfRef.evaluateSequential(java.base@21.0.2/ForEachOps.java:174)
> at
> java.util.stream.AbstractPipeline.evaluate(java.base@21.0.2/AbstractPipeline.java:234)
> at
> java.util.stream.ReferencePipeline.forEach(java.base@21.0.2/ReferencePipeline.java:596)
> at
> org.apache.jackrabbit.vault.fs.impl.io.DocViewImporter.logIgnoredProtectedProperties(DocViewImporter.java:1215)
> at
> org.apache.jackrabbit.vault.fs.impl.io.DocViewImporter.createNewNode(DocViewImporter.java:1181)
> at
> org.apache.jackrabbit.vault.fs.impl.io.DocViewImporter.addNode(DocViewImporter.java:931)
> at
> org.apache.jackrabbit.vault.fs.impl.io.DocViewImporter.startDocViewNode(DocViewImporter.java:410)
> at
> org.apache.jackrabbit.vault.fs.impl.io.DocViewSAXHandler.startElement(DocViewSAXHandler.java:353)
> {noformat}
> Alternatively I would also be happy about an optimized version of
> "DocViewImporter.isPropertyProtected", which does not need resolve every
> property all the time from scratch.
> Note: Optimizing {{isPropertyProtected}} might be the prefered way, as it is
> also invoked via the {{DocViewImporter.setUnprotectedProperties}} code path,
> which I also see very frequently in my dumps.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
