[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Converting ClearCase to CVS

From: Kaz Kylheku
Subject: Re: Converting ClearCase to CVS
Date: Fri, 15 Feb 2002 18:17:01 GMT
User-agent: slrn/ (Linux)

In article <address@hidden>, Ajmal Chaumun wrote:
>This message is in MIME format. Since your mail reader does not understand
>this format, some or all of this message may not be legible.
>Content-Type: text/plain;
>       charset="iso-8859-1"
>Amazing how these are very much similar to what happened here. I do not want
>to repeat the whole story, but would simply add this.
>When we realized that going from ClearCase to CVS was still an option, we
>were told of the lack of features in CVS compared to ClearCase, e.g., no
>directory versioning, etc. However, even with those "deficiencies", this

Meta-CVS has solid directory versioning, supporting parallel development
of your project's directory structure. You can do merges, conflict resolution,
etc. And as a CVS user, you will find that you basically 
*already know how* to do these things.

Now here is a list of Meta-CVS ``blocking issues'' to use

- portability: 
  1 currently it only runs on Linux.
  2 it requires an operating system with hard link support though
    it can be made to work, albeit disk space inefficiently, and somewhat
    time inefficiently.
- reduced features over CVS in some areas:
  3 modules file is meaningless (but with directory versioning, you largely
    don't need this silly hack anyway).
  4 vendor imports won't work properly (unless you do them at the CVS level
    over raw Meta-CVS archives, in which case they will not only work,
    but work with directory versioning).
  5 currently, there is nothing like .cvsignore and cvswrappers under
    Meta-CVS. (The import and add commands will soon be revised to incorporate
    file type management).

Problem 1 is user driven. If someone wants it badly enough, they can port
it; there are enough hoooks there to do the job.  Point 2 shouldn't be a
problem. If it is ported to windows, it can use NTFS hard links. Nobody
should be using development tools on any ``lesser'' Windows besides NT
and 2K anyway.  

Problem 5 will be history within weeks; Meta CVS will gather up file types
and keep metadata about how a files having a given suffix should be added.

Problem 4 can be solved with some intelligent code that will analyze
a tree of files, and then figure out how to import it, with some
interactive assistance from the user. A vendor import under Meta-CVS
must be capable of importing snapshots whose directory structure has
changed. This will require matching the new tree against the existing
one, and cross-referencing the elements that have been renamed or
moved. This could be largely automated, with some help from the user,
who could override some of the automatic decisions as well as provide
a final integrity check.

Meta-CVS: version control with directory structure versioning over top of CVS.

reply via email to

[Prev in Thread] Current Thread [Next in Thread]