You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@bigtop.apache.org by "Hari Krishna Dara (JIRA)" <ji...@apache.org> on 2018/03/08 14:45:00 UTC
[jira] [Comment Edited] (BIGTOP-2999) Formalize LOCAL_DOWNLOADS_DIR
environment variable as the local cache
[ https://issues.apache.org/jira/browse/BIGTOP-2999?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16368601#comment-16368601 ]
Hari Krishna Dara edited comment on BIGTOP-2999 at 3/8/18 2:44 PM:
-------------------------------------------------------------------
| why trying to reuse the contents, when the first action is to remove its contents ?
When FORREST_HOME is already set by user and is valid, it skips the entire logic, so no removal will happen. The support is useful for those environments where apache-forrest tar file is expanded in advance so that this logic can be skipped for additional gain in the build time.
Regarding the duplicate, thanks for another good catch. I seem to have somehow left the refactoring I was trying to do in the middle. I have now made further refactoring of the variables for the hadoop file on the lines of the pig file, however I haven't added any md5 check. If you think the md5 check adds significant value, then I can add that too by borrowing the logic from the pig file, but this would add some maintenance burden for updating the tomcat version in the future.
was (Author: haridsv):
| why trying to reuse the contents, when the first action is to remove its contents ?
When FORREST_HOME is already set by user and is valid, it skips the entire logic, so no removal will happen. The support is useful for those environments where apache-forrest tar file is expanded in advance so that this logic can be skipped for additional gain in the build time.
Regarding the duplicate, thanks for another good catch. I seem to have some left the refactoring I was trying to do in the middle. I have now made further refactoring of the variables for the hadoop file on the lines of the pig file, however I haven't added any md5 check. If you think the md5 check adds significant value, then I can add that too by borrowing the logic from the pig file, but this would add some maintenance burden for updating the tomcat version in the future.
> Formalize LOCAL_DOWNLOADS_DIR environment variable as the local cache
> ---------------------------------------------------------------------
>
> Key: BIGTOP-2999
> URL: https://issues.apache.org/jira/browse/BIGTOP-2999
> Project: Bigtop
> Issue Type: Improvement
> Components: build, hadoop, pig
> Reporter: Hari Krishna Dara
> Assignee: Hari Krishna Dara
> Priority: Major
> Attachments: BIGTOP-2999.patch, BIGTOP-2999_2.patch, BIGTOP-2999_3.patch, BIGTOP-2999_4.patch, BIGTOP-2999_5.patch
>
>
> BIGTOP-2989 and BIGTOP-2992 introduced support for LOCAL_DOWNLOADS_DIR environment variable as a location to check before downloading from Internet. This improvement aims to formalize this variable for the following:
> * For Hadoop, always download the file to this directory so that it can be reused in the build via a file:// URL
> * For Pig, use this directory even if a download is necessary.
> * Add details on this variable to confluence.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)