Re: Lucene/Solr 8.5.1 bugfix release

2020-04-02 Thread Adrien Grand
My general take on this is that it's ok to upgrade a dependency in a patch
release if the dependency upgrade itself is a new patch release of the same
minor version. The changelog of Tika 1.24 seems to include not only bug
fixes but also some enhancements[1], so I'd rather do a 8.6 release in the
near future than backport this dependency upgrade to 8.5.

[1] https://tika.apache.org/1.24/index.html

On Thu, Apr 2, 2020 at 9:33 PM Cassandra Targett 
wrote:

> Should we consider backporting SOLR-14367 (the most recent Tika upgrade)?
> It addresses a CVE in Tika, and while I think we usually avoid changing 3rd
> party component versions in patch releases, but maybe we should in this
> case? The upgrade also looks like it was pretty straightforward (drop-in
> replacement).
>
> Cassandra
> On Apr 2, 2020, 12:47 PM -0500, Ignacio Vera , wrote:
>
> Hi,
>
> I propose a quick 8.5.1 bugfix release and I volunteer as RM. The main
> motivation for this release is LUCENE-9300 where Jim addressed a serious
> bug that can lead to data corruption when merging indices via IW#addIndices.
>
> If there are no objections I am planning to create a RC early next week.
>
> Best regards,
>
> Ignacio
>
>
>
>

-- 
Adrien


Re: Lucene/Solr 8.5.1 bugfix release

2020-04-02 Thread Cassandra Targett
Should we consider backporting SOLR-14367 (the most recent Tika upgrade)? It 
addresses a CVE in Tika, and while I think we usually avoid changing 3rd party 
component versions in patch releases, but maybe we should in this case? The 
upgrade also looks like it was pretty straightforward (drop-in replacement).

Cassandra
On Apr 2, 2020, 12:47 PM -0500, Ignacio Vera , wrote:
> Hi,
>
> I propose a quick 8.5.1 bugfix release and I volunteer as RM. The main 
> motivation for this release is LUCENE-9300 where Jim addressed a serious bug 
> that can lead to data corruption when merging indices via IW#addIndices.
>
> If there are no objections I am planning to create a RC early next week.
>
> Best regards,
>
> Ignacio
>
>


Re: Lucene/Solr 8.5.1 bugfix release

2020-04-02 Thread Adrien Grand
+1

On Thu, Apr 2, 2020 at 7:47 PM Ignacio Vera  wrote:

> Hi,
>
> I propose a quick 8.5.1 bugfix release and I volunteer as RM. The main
> motivation for this release is LUCENE-9300 where Jim addressed a serious
> bug that can lead to data corruption when merging indices via IW#addIndices.
>
> If there are no objections I am planning to create a RC early next week.
>
> Best regards,
>
> Ignacio
>
>
>


-- 
Adrien


Re: Gradle Wrapper Changes

2020-04-02 Thread Mike Drob
Thanks, Uwe! I knew about that for javac but didn't make the connection to
running java in single source mode!

On Thu, Apr 2, 2020 at 11:45 AM Uwe Schindler  wrote:

> I made a comment as review.
>
>
>
> The Java 8 issue can be solved very elegant. Problem is that Java 8 tries
> to load the file as a class, because it cannot directly run source files.
> This is some semi-automatism by Java 11, where java compiles the file if
> its not a class. This is not supported by Java 8.
>
>
>
> This can be bypassed by specifiying the “--source 11” parameter before the
> java file. This is an unknown option in Java 8 and produces a more useful
> error:
>
>
>
> >java --source 11 test\foo.java
>
> Unrecognized option: --source
>
> Error: Could not create the Java Virtual Machine.
>
> Error: A fatal exception has occurred. Program will exit.
>
>
>
> In addition, adding the source parameter makes the java source file format
> of the parameter explicit, like with Javac you should always supply the
> target java version when compiling something..
>
>
>
> Uwe
>
>
>
> -
>
> Uwe Schindler
>
> Achterdiek 19, D-28357 Bremen
>
> https://www.thetaphi.de
>
> eMail: u...@thetaphi.de
>
>
>
> *From:* Mike Drob 
> *Sent:* Thursday, April 2, 2020 6:29 PM
> *To:* dev@lucene.apache.org
> *Subject:* Gradle Wrapper Changes
>
>
>
> Devs,
>
>
>
> I'll be committing a set of changes for LUCENE-9266 shortly, PR can be
> found at https://github.com/apache/lucene-solr/pull/1390
>
>
>
> We've tested this on Windows, Mac OS, and Ubuntu, so I'm fairly confident
> in what we're doing here, but because this has the potential to be an
> intrusive change I'm sending out an extra note to the mailing list.
>
>
>
> If you happen to see an error like the following:
>
>
>
> Error: Could not find or load main class
> buildSrc.src.main.java.org.apache.lucene.gradle.WrapperDownloader.java
>
>
>
> Then check to make sure that your currently selected java executable is at
> least version 11. This is already a requirement for the master branch, but
> folks that work on multiple branches may get tripped up on this.
>
>
>
> The cosmetics of this will likely improve over time, but the current goal
> is to make forward progress on nightly smoke testing.
>
>
>
> Thanks,
>
> Mike
>


Lucene/Solr 8.5.1 bugfix release

2020-04-02 Thread Ignacio Vera
Hi,

I propose a quick 8.5.1 bugfix release and I volunteer as RM. The main
motivation for this release is LUCENE-9300 where Jim addressed a serious
bug that can lead to data corruption when merging indices via IW#addIndices.

If there are no objections I am planning to create a RC early next week.

Best regards,

Ignacio


RE: Gradle Wrapper Changes

2020-04-02 Thread Uwe Schindler
I made a comment as review.

 

The Java 8 issue can be solved very elegant. Problem is that Java 8 tries to 
load the file as a class, because it cannot directly run source files. This is 
some semi-automatism by Java 11, where java compiles the file if its not a 
class. This is not supported by Java 8.

 

This can be bypassed by specifiying the “--source 11” parameter before the java 
file. This is an unknown option in Java 8 and produces a more useful error:

 

>java --source 11 test\foo.java

Unrecognized option: --source

Error: Could not create the Java Virtual Machine.

Error: A fatal exception has occurred. Program will exit.

 

In addition, adding the source parameter makes the java source file format of 
the parameter explicit, like with Javac you should always supply the target 
java version when compiling something..

 

Uwe

 

-

Uwe Schindler

Achterdiek 19, D-28357 Bremen

https://www.thetaphi.de

eMail: u...@thetaphi.de

 

From: Mike Drob  
Sent: Thursday, April 2, 2020 6:29 PM
To: dev@lucene.apache.org
Subject: Gradle Wrapper Changes

 

Devs,

 

I'll be committing a set of changes for LUCENE-9266 shortly, PR can be found at 
https://github.com/apache/lucene-solr/pull/1390

 

We've tested this on Windows, Mac OS, and Ubuntu, so I'm fairly confident in 
what we're doing here, but because this has the potential to be an intrusive 
change I'm sending out an extra note to the mailing list.

 

If you happen to see an error like the following:

 

Error: Could not find or load main class 
buildSrc.src.main.java.org.apache.lucene.gradle.WrapperDownloader.java

 

Then check to make sure that your currently selected java executable is at 
least version 11. This is already a requirement for the master branch, but 
folks that work on multiple branches may get tripped up on this.

 

The cosmetics of this will likely improve over time, but the current goal is to 
make forward progress on nightly smoke testing.

 

Thanks,

Mike



Gradle Wrapper Changes

2020-04-02 Thread Mike Drob
Devs,

I'll be committing a set of changes for LUCENE-9266 shortly, PR can be
found at https://github.com/apache/lucene-solr/pull/1390

We've tested this on Windows, Mac OS, and Ubuntu, so I'm fairly confident
in what we're doing here, but because this has the potential to be an
intrusive change I'm sending out an extra note to the mailing list.

If you happen to see an error like the following:

Error: Could not find or load main class
buildSrc.src.main.java.org.apache.lucene.gradle.WrapperDownloader.java

Then check to make sure that your currently selected java executable is at
least version 11. This is already a requirement for the master branch, but
folks that work on multiple branches may get tripped up on this.

The cosmetics of this will likely improve over time, but the current goal
is to make forward progress on nightly smoke testing.

Thanks,
Mike