[
https://issues.apache.org/jira/browse/XALANJ-2671?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17775910#comment-17775910
]
Joseph Kessselman edited comment on XALANJ-2671 at 10/16/23 8:27 PM:
---------------------------------------------------------------------
With all due respect – Could folks reach consensus on a solution? I'm perfectly
willing to attempt to apply either approach, or another approach, if I at least
vaguely understand it... but you two are further down into the Git weeds than I
am, and I don't have a valid opinion between the suggestions.
I honestly don't care if "bad practice" is needed this once to reverse a
decision that was itself bad practice. We are NOT planning on making a habit of
this. It just needs to work, reliably, once.
Vlad, I would love to see the process spelled out for your approach so I can
compare the two properly. If removing the bad commits without disrupting the
code history since then is trivial, please point out how. Is it just the two
mentioned above?
This issue is a blocker for my Maven proposal, assuming we do want the history
back (which everyone agrees is important). So I want it settled as soon as
reasonably possible.. If you two can't find your way to an agreement, I'll do
whatever I have the tools and knowledge to do, and worry about whether it was
the best solution another day.
was (Author: jkesselm):
With all due respect – Could folks reach consensus on a solution? I'm perfectly
willing to attempt to apply either approach, or another approach, if I at least
vaguely understand it... but you two are further down into the Git weeds than I
am, and I don't have a valid opinion between the suggestions.
I honestly don't care if "bad practice" is needed this once to reverse a
decision that was itself bad practice. We are NOT planning on making a habit of
this. It just needs to work, reliably, once.
Vlad, I would love to see the process spelled out for your approach so I can
compare the two properly. If removing the bad commits is trivial, please point
out which and how.
This issue is a blocker for my Maven proposal, assuming we do want the history
back (which everyone agrees is important). So I want it settled as soon as
reasonably possible.. If you two can't find your way to an agreement, I'll do
whatever I have the tools and knowledge to do, and worry about whether it was
the best solution another day.
> Source code history lost in "clearing all contents" + "updating master
> branch, to have same contents as..." commits
> -------------------------------------------------------------------------------------------------------------------
>
> Key: XALANJ-2671
> URL: https://issues.apache.org/jira/browse/XALANJ-2671
> Project: XalanJ2
> Issue Type: Bug
> Security Level: No security risk; visible to anyone(Ordinary problems in
> Xalan projects. Anybody can view the issue.)
> Reporter: Vladimir Sitnikov
> Assignee: Gary D. Gregory
> Priority: Critical
>
> I just noticed that the following commits virtually erased all the code
> history, and they made it way harder to use git for analysis of the changes.
> I would suggest stopping further commits to the branch that contains
> "clearing all + updating all" changes, and start over from a reasonable
> branch.
> Please, take this seriously. It is really crucial to keep git annotate in a
> good shape, otherwise bug analysis becomes extremely hard.
> 0affa12bb12e45bf681a6e6a918d68884d7e1877 "clearing all the project contents
> from the xalanj master branch"
> 356ce496faa80d97ae8a8406e37f6bd994b39233 "updating master branch, to have
> same contents as the branch xalan-j_2_7_1_maint"
> For instance, if you check "git blame" for DocumentCall.java, you'll find
> that "every line was added only 2 months ago":
> https://github.com/apache/xalan-java/blame/master/src/org/apache/xalan/xsltc/compiler/DocumentCall.java
> I am afraid the above commits are showstopper for further activity on the
> branch.
> git blame on xalan-j_2_7_1_maint branch looks much better:
> https://github.com/apache/xalan-java/blame/xalan-j_2_7_1_maint/src/org/apache/xalan/xsltc/compiler/DocumentCall.java
> Since there were no releases on the master branch since the bad commits, I
> would suggest removing the commits from the repository completely.
> It is fine from the ASF policies, and it would be great to kill the commits
> as having commits that modify 400'000 lines make Git history hard to browse
> in various tools.
> Please remove those commits as they make history browsing, and error analysis
> hard.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@xalan.apache.org
For additional commands, e-mail: dev-h...@xalan.apache.org
