You are viewing a plain text version of this content. The canonical link for it is here.
Posted to common-issues@hadoop.apache.org by "Gabor Bota (JIRA)" <ji...@apache.org> on 2019/03/29 12:32:00 UTC

[jira] [Comment Edited] (HADOOP-16218) findbugs warning of null param to non-nullable method in Configuration with Guava update

    [ https://issues.apache.org/jira/browse/HADOOP-16218?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16804906#comment-16804906 ] 

Gabor Bota edited comment on HADOOP-16218 at 3/29/19 12:31 PM:
---------------------------------------------------------------

Here is my opinion and justification about the other findbugs issues:
 * *Null passed for non-null parameter of com.google.common.base.Preconditions.checkState(boolean, String, Object, Object, Object) in org.apache.hadoop.hdfs.qjournal.server.Journal.getPersistedPaxosData(long)*
In {{org/apache/hadoop/hdfs/qjournal/server/Journal.java:1064}} we call
{code:java}
Preconditions.checkState(ret != null &&
          ret.getSegmentState().getStartTxId() == segmentTxId,
          "Bad persisted data for segment %s: %s ; journal id: %s",
          segmentTxId, ret, journalId);
{code}
for this call findbug assumes that {{Argument 4 might be null but must not be null}}, but Guava 27.0's {{com.google.common.base.Preconditions#checkState(boolean, java.lang.String, java.lang.Object, java.lang.Object, java.lang.Object)}} is annotated like the following:
{code:java}
  public static void checkState(
      boolean b,
      @Nullable String errorMessageTemplate,
      @Nullable Object p1,
      @Nullable Object p2,
      @Nullable Object p3) {
{code}
so we have {{@Nullable}} on each parameter for the method. I don't see this warning as justified, or need to be fixed.

 * *Null passed for non-null parameter of com.google.common.base.Preconditions.checkArgument(boolean, String, Object) in org.apache.hadoop.hdfs.qjournal.server.JournalNode.getLogDir(String, String)*
In org/apache/hadoop/hdfs/qjournal/server/JournalNode.java:325 we call
{code:java}
    Preconditions.checkArgument(jid != null &&
        !jid.isEmpty(),
        "bad journal identifier: %s", jid);
{code}
for this call findbug assumes that {{Argument 3 might be null but must not be null}}, but Guava 27.0's {{com.google.common.base.Preconditions#checkArgument(boolean, java.lang.String, java.lang.Object)}} is annotated like the following:
{code:java}
  public static void checkArgument(
      boolean b, @Nullable String errorMessageTemplate, @Nullable Object p1) {
{code}
so we have {{@Nullable}} on argument 3, and that renders the assumption incorrect.

 * *Nullcheck of jid at line 346 of value previously dereferenced in org.apache.hadoop.hdfs.qjournal.server.JournalNode.getLogDir(String, String)*
This is about the {{assert jid != null;}} at JournalNode.java:346. IMHO that check is for devs to inform that the variable can't be null at that point - so just for visibility. I would leave it as is, it's not a redundant check, just additional information. (I'm not a fan of using {{assert}} in production code, but if it's there we can leave it).


was (Author: gabor.bota):
I'm checking these findbug issues, and I don't see why the first is breaking:
 * *Null passed for non-null parameter of com.google.common.base.Preconditions.checkState(boolean, String, Object, Object, Object) in org.apache.hadoop.hdfs.qjournal.server.Journal.getPersistedPaxosData(long)*
In {{org/apache/hadoop/hdfs/qjournal/server/Journal.java:1064}} we call
{code:java}
Preconditions.checkState(ret != null &&
          ret.getSegmentState().getStartTxId() == segmentTxId,
          "Bad persisted data for segment %s: %s ; journal id: %s",
          segmentTxId, ret, journalId);
{code}
for this call findbug assumes that {{Argument 4 might be null but must not be null}}, but Guava 27.0's {{com.google.common.base.Preconditions#checkState(boolean, java.lang.String, java.lang.Object, java.lang.Object, java.lang.Object)}} is annotated like the following:
{code:java}
  public static void checkState(
      boolean b,
      @Nullable String errorMessageTemplate,
      @Nullable Object p1,
      @Nullable Object p2,
      @Nullable Object p3) {
{code}
so we have {{@Nullable}} on each parameter for the method. I don't see this warning as justified, or need to be fixed.

 * *Null passed for non-null parameter of com.google.common.base.Preconditions.checkArgument(boolean, String, Object) in org.apache.hadoop.hdfs.qjournal.server.JournalNode.getLogDir(String, String)*
In org/apache/hadoop/hdfs/qjournal/server/JournalNode.java:325 we call
{code:java}
    Preconditions.checkArgument(jid != null &&
        !jid.isEmpty(),
        "bad journal identifier: %s", jid);
{code}
for this call findbug assumes that {{Argument 3 might be null but must not be null}}, but Guava 27.0's {{com.google.common.base.Preconditions#checkArgument(boolean, java.lang.String, java.lang.Object)}} is annotated like the following:
{code:java}
  public static void checkArgument(
      boolean b, @Nullable String errorMessageTemplate, @Nullable Object p1) {
{code}
so we have {{@Nullable}} on argument 3, and that renders the assumption incorrect.

 * *Nullcheck of jid at line 346 of value previously dereferenced in org.apache.hadoop.hdfs.qjournal.server.JournalNode.getLogDir(String, String)*
This is about the {{assert jid != null;}} at JournalNode.java:346. IMHO that check is for devs to inform that the variable can't be null at that point - so just for visibility. I would leave it as is, it's not a redundant check, just additional information. (I'm not a fan of using {{assert}} in production code, but if it's there we can leave it).

> findbugs warning of null param to non-nullable method in Configuration with Guava update
> ----------------------------------------------------------------------------------------
>
>                 Key: HADOOP-16218
>                 URL: https://issues.apache.org/jira/browse/HADOOP-16218
>             Project: Hadoop Common
>          Issue Type: Sub-task
>          Components: build
>    Affects Versions: 3.3.0
>            Reporter: Steve Loughran
>            Assignee: Steve Loughran
>            Priority: Minor
>
> Findbugs is fussing over some unchanged code in configuration
> {code}
> Null passed for non-null parameter of writeXml(String, Writer) in org.apache.hadoop.conf.Configuration.writeXml(Writer)
> Bug type NP_NULL_PARAM_DEREF_ALL_TARGETS_DANGEROUS (click for details) 
> In class org.apache.hadoop.conf.Configuration
> In method org.apache.hadoop.conf.Configuration.writeXml(Writer)
> Called method org.apache.hadoop.conf.Configuration.writeXml(String, Writer)
> At Configuration.java:[line 3490]
> Argument 1 is definitely null but must not be null
> Definite null passed to dangerous method call target org.apache.hadoop.conf.Configuration.writeXml(String, Writer)
> {code}
> Code looks fine; it's invoking Guava.Strings for the check for string == empty-or-null; maybe something changed there.
> Proposed: add @Nullable in Configuration writeXml



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

---------------------------------------------------------------------
To unsubscribe, e-mail: common-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: common-issues-help@hadoop.apache.org