[jira] [Commented] (SOLR-17246) releaseWizard.py remembers choices despite --dry-run

["Jan Høydahl (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/SOLR-17246?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17839849#comment-17839849 ] 

Jan Høydahl commented on SOLR-17246:
------------------------------------

Dry run is only designed to not execute commands. It is supposed to still persist state. So you can safely get a feel for the entire process, and everything feels and acts like a real release, except that it will never actually run the commands.

I'm sure it is possible to change this initial definition of "dry" and make the feature even more useful somehow, perhaps with a temp in-memory state only?

> releaseWizard.py remembers choices despite --dry-run
> ----------------------------------------------------
>
>                 Key: SOLR-17246
>                 URL: https://issues.apache.org/jira/browse/SOLR-17246
>             Project: Solr
>          Issue Type: Improvement
>      Security Level: Public(Default Security Level. Issues are Public) 
>            Reporter: Gus Heck
>            Priority: Major
>
> While perhaps creating the initial .solrrc file is inevitable, it definately doesn't seem like a "dry" run if the choices you made are persisted and show up when you later run without the --dry-run option.
> {*}Steps to reproduce{*}:
>  # run the release wizard in dry run mode
>  # go to checklist (option 1)
>  # prerequisites (option 1)
>  # read up on release process (option 1)
>  # Q: Mark task 'Read up on the release process' as completed? (y/n): y
>  # quit the release wizard script (options 5, 10 and 8)
>  # run the release wizard with out --dry-run
>  # go to checklist
> {*}Expected{*}: A fresh clean checklist
> {*}Actual{*}: The first checklist item is listed as 1/4 not 0 of 4 and if entered already has a checkmark
> The normal notion of a dry run is that it will not create any persistent changes. Yet we persist the acknowledgements of step completion, and this doesn't make sense because if the person executing the dry run hasn't run the command but only echoed it, we are potentially tricking them into thinking they've completed the step if they come back later and forget that they only did a dry run previously (or assume that since it is checked they must not have done a dry run...).



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

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