Re: CVS -> SVN -> Git

Markus Schiltknecht wrote:
> Michael Haggerty wrote:
>> If I understand the git-fast-import format
correctly, it is no problem
>> to add files from multiple points of time into a
branch.
> 
> Hm... reading the git-fast-import manpage it says about
the 'tag' command:

I was talking about "branch", not "tag".
 I understand the problem with
tagging.

In the absence of experience of my own, I had intended to
follow the
advice given in "Use Tag Fixup Branches" in the
git-fast-include
documentation page [1].

I hope to be able to work on output backends for other SCMs
after git;
then I'll need to know more about their specific
requirements.

Michael

[1] http://www.kernel.org/pub/software/scm/git/d
ocs/git-fast-import.html

------------------------------------------------------------
---------
To unsubscribe, e-mail: dev-unsubscribecvs2svn.tigris.org
For additional commands, e-mail: dev-helpcvs2svn.tigris.org

Hi,

Michael Haggerty wrote:
> I was talking about "branch", not
"tag".  I understand the problem with
> tagging.

AFAICT, the problem remains the same, no?

> In the absence of experience of my own, I had intended
to follow the
> advice given in "Use Tag Fixup Branches" in
the git-fast-include
> documentation page [1].

The way I'm reading it is, that they use branches instead of
tags, for 
tags (or branchpoints) which span multiple revisions. That's
what I 
think a "tag fixup branch" is.

Everything else I'm reading in that section seems to
confirm, that git 
can handle revisions with multiple parents. I.e. they can do
what I've 
outlined as Variant 2. b).  (The "one commit"
solution to "fix up the 
files" - what I'm referring to as an artificial
revision).

I can't find any difference between tags and branchpoints,
WRT importing 
them from CVS to git.

Regards

Markus

------------------------------------------------------------
---------
To unsubscribe, e-mail: dev-unsubscribecvs2svn.tigris.org
For additional commands, e-mail: dev-helpcvs2svn.tigris.org