>> Well. First of all you are using a special SVN usage philosophy where all
>> development is done in branches as the base for your argument.
> Yes... but who are WE to decide what philosophy the end users are to choose?
:-D You tell me! You were the one who started to involving different philosophies in the discussion. I still say that using HEAD for tagging is a bad idea no matter what philosophy you are using. Some philosophies can reduce the risks, but not eliminate it. Hence it's always better to use a specific revision instead of HEAD for tagging.
> While I have one view on it, you have another view on it. Just because you
> think it should be done one way doesn't make my way invalid.
I did not suggest to remove the feature. Only to make another option the default.
The SVN-usage philosophy has nothing to do with it. (And I still have not told you what method I prefer. I have only said that not everyone uses the method you used in your argument.)
>> The point is that when you tag (or branch for that matter) you generally
>> desire control over WHAT you are tagging.
> Unless you follow my method in which the trunk contains a pristine version
> at all times.
No. That will only guarantee absolute knowledge of the revision you are tagging if there is only one person who have commit access to trunk, and that is the same and only person that has access to create tags. In a multi user environment you cannot guarantee it. You just cant.
Simply shrugging and saying "Trunk is pristine, I don't care what I'm tagging." is not good enough when you strive for traceability and reproducibility.
> This is a philosophical debate to which there is no final answer. My main
> point, however, is that TSVN should follow the overall SVN convention in
> which operations work on HEAD by default.
I disagree. I think we should follow SVN convention as close as possible when it makes sense, but not when the SVN convention is a result of a compromise that's not applicable to a TSVN dialog.
Hans-Emil
Received on Wed Jun 21 15:05:36 2006