[ 
https://issues.apache.org/jira/browse/LUCENE-5442?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13900038#comment-13900038
 ] 

Mark Miller commented on LUCENE-5442:
-------------------------------------

We have similar issues with the morphlines stuff (which also heavily uses 
tika). It's tough because it's extremely hard to have tests that can exercise 
all of this or to know that you have exercised it all.

It would be great if you could run something that figured out all the maven 
transitive version dependencies and let us know what version we are behind on 
or missing.

You still have to harmonize, so it won't necessarily be foolproof, but a great 
check.

> Build system should sanity check transative 3rd party dependencies
> ------------------------------------------------------------------
>
>                 Key: LUCENE-5442
>                 URL: https://issues.apache.org/jira/browse/LUCENE-5442
>             Project: Lucene - Core
>          Issue Type: Improvement
>          Components: general/build
>            Reporter: Hoss Man
>            Assignee: Steve Rowe
>
> SOLR-5365 is an example of a bug that croped up because we upgraded a 3rd 
> party dep (tika) w/o realizing that the version we upgraded too depended on a 
> newer version of another 3rd party dep (commons-compress)
> in a comment in SOLR-5365, Jan suggested that it would be nice if there was 
> an easy way to spot problems like this ... i asked steve about it, thinking 
> maybe this is something the maven build could help with, and he mentioned 
> that there is already an ant task to inspect the ivy transative deps in order 
> to generate the maven deps and it could be used to help detect this sort of 
> problem.
> opening this issue per steve's request as a reminder to look into this 
> possibility.



--
This message was sent by Atlassian JIRA
(v6.1.5#6160)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to