[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: bug report: tag is deleted by mistake
From: |
Todd Denniston |
Subject: |
Re: bug report: tag is deleted by mistake |
Date: |
Wed, 13 Aug 2008 15:25:03 -0400 |
User-agent: |
Thunderbird 2.0.0.16 (X11/20080707) |
Derek Price wrote, On 08/13/2008 03:11 PM:
Larry Jones wrote:
Derek Price writes:
Sam Louis wrote:
cvs up -r tagname a.txt
cvs rm -rf a.txt
cvs ci a.txt
I think that this operation should be forbidden, don't u think this ?
It most certainly shouldn't be forbidden, though an argument might be
made for marking the file "dead" on the branch and leaving the tag in
place.
I think that *is* what happens if the tag is a branch tag; deleting the
tag only happens when it's a revision tag.
Thanks for making what's happening here understandable.
There is an argument for
refusing to allow the commit of the delete the same way we refuse to
allow commit of changes in that case, but we've always been leary of
changing the way it works since someone somewhere might be depening on
it.
Ah, yes. My mistake, I misunderstood but recall this now. This is one
of those cases where I don't like CVS's behavior since it does seem
inconsistent. I wouldn't really be adverse to forbidding this, but I'm
not going to fight for it either.
I agree, now that I understand the scope of what Sam was seeing, that this is
inconsistent. I had been under the assumption that ALL commits against a
normal tag would fail.
Why is there an inconsistency here? And I don't mean why was it made this
way, I mean is there a special case in the code that allows this commit to
work while commits against changes to the text of the file will fail?
If there is a special case made just for this, I vote* to remove it or replace
it with printf("use `cvs tag -d %s %s` to remove the tag.", TagText, FileName).
*as a CVS user who is not making the code change. :)
--
Todd Denniston
Crane Division, Naval Surface Warfare Center (NSWC Crane)
Harnessing the Power of Technology for the Warfighter