(my last mail was cut for some reason)
Keith,
> Without sending me off to some link to read some 3rd party marketing
> speak about why I would want to upgrade to Subversion (not that you
> would do that to me of course), can you please highlight from your own
> experience what such features of SVN you are desperately seeking that
> are not currently provided with the normal CVS.
> In other words, what nice features of Subversion are going to propel you
> and other contributors such as myself into more productivity because of
> features available in SVN that are not available in CVS.
> I'm quite happy with CVS and my WinCVS client. I haven't found myself
> saying, gee it would be really great if CVS had this feature or if CVS
> had that feature it would make my life really easy and the last thing I
> want to go and do is be forced to switch to some SVN client when I have
> something that works well already.
> [...]
I don't think there is one such compelling feature that would urge to switch
Castor's repository. But there are many small improvements, that would
justify the change in the result:
- Subversion works over HTTP. This makes it much more "firewall friendly"
than CVS. I for example don't have access to CVS on work. But I can
use Subversion repositories through the HTTP proxy.
- It stores not only a HEAD version, but also the BASE version of the
source in the local working copy. This results in the two following
advantages.
- SVN makes use of two way deltifying. So only file changes are transmitted
over the net to reduce the bandwith usage.
- You can work partly offline because the repository is accessed less often.
This is good if you only have dialin access to the internet.
- The "atomic commits" connect all related file changes. In CVS the history
information is stored only file by file.
This makes for example possible to get the list of recent changes in the
whole repository just by executing "svn log" in the root directory. No more
searching for change comments spread over all that different files...
- SVN supports renaming and removing of files while preserving the history
information to make easier refactoring of source code.
- With SVN there is no more the need of "branch root tags" to make effective
usage of branches.
- SVN doesn't exhibit CVS's time zone switch problems.
- If you are using TortoiseSVN, you can achieve integration with bug track
systems like Bugzilla.
There are other internal improvements, which are less visible to the user:
- better handling of binary files
- better branch and tag performance
- The repository is stored in a database like format to enhance performance.
(Currently there are available two options: Berkeley DB and SVN's own
plain file format. I would recommend the newer, plain format.)
- Keyword substitution takes place at checkout time, not at checking in changes,
so they are not stored in the repository version of the source, but only in
the
metadata.
- SVN doesn't exhibit the performance problems of CVS for large repositories.
What important features did I miss? ;-)
Regards,
Martin
-----------------------------------------------------------
If you wish to unsubscribe from this mailing, send mail to
[EMAIL PROTECTED] with a subject of:
unsubscribe castor-dev
