Re: Support for Windows 10 (CVSNT vs. SVN and CVS 1.x)

On Wed, Aug 30, 2017 at 8:18 AM, Arthur Barrett
<arthur.barrett-qn/[email protected]> wrote:
> Hi Michael,
>
> It sounds like you don't know much about CVSNT, so let me correct a
> couple of misconceptions.
>
>> If you're willing to pay, March Hare will sell you a client, but again
>> that's a command line tool.
>
> No - CVSNT (what we now call CVS Suite) is a single installed combining
> TortoiseCVS, WinCVS, WinMerge, Release Manager, Workspace Manager,
> Control Panel, CVSAgent and CVSNT combined.
>
> The March Hare solution has always been a 'combined' installer with all
> the components tested together into a single supported package.
>
> We supplied the TortoiseCVS and WinCVS projects with a simple 'command
> line only' installer for bundling in their community projects.

So that's better - if the OP has a budget you can give them exactly
what they need. Great!

> You've always been able to run CVS 1.x on Windows (or cygwin) - but
> CVSNT was created to address all the problems that doing that created
> (where you ended up with HELLO.cpp, Hello.cpp and hello.cpp in your
> repository).  Plus of course CVSNT added a LOT of extra features aimed
> at commercial software developers.
>
>> If you want to do some work yourself, you could try fixing TortiseCVS
>> (http://www.tortoisecvs.org/development.shtml) yourself. I don't know
>> where the Win10 problems lie, so I don't know if you'd need a new
>> CVSNT version or not. If you do, I suspect you could port the current
>> CVS client (http://www.nongnu.org/cvs/) without too much trouble, if
>> March Hare can't help with a new release of CVSNT.
>
> Because it took about 7 man years to create CVSNT 2.5 from the CVS 1.11
> sources, someone else can do it 'without too much trouble'?

I didn't say it would do everything your tool does. But if one isn't
using March Hare on the server side (and the OP didn't sound like he
was), then I don't think March Hare's CVS client adds THAT much. I've
managed to screw up plenty when using CVSNT, so your tools aren't
magic - just very nice.

I also didn't say it wouldn't take a while to put together such a
release. Time to completion and difficulty are not necessarily
related.

>> Basically: No one is doing free development on CVS
>
> Both CVS and CVSNT are under active development.  The CVS 1.x
> development may look stalled if you look at savannah for reasons it is
> not politik of me to go into - but I know quite a lot of work and
> discussion is going on behind the scenes.  There is a separate private
> mailing list for that.
>
> There are new releases of CVSNT every few weeks which we announce via
> twitter feed.
>
>> That's really down to the fact that it frankly
>> doesn't do everything that you want in a source control system (it
>> doesn't track history across file renames.
>
> CVSNT has tracked the history across file renames since somewhere in
> 2005 from memory.  Ditto 'automatic commit identifiers' aka 'atomic
> commit'.
>
> Many of the features of CVSNT are still fairly unique, eg: failsafe
> audit to sql database, user defined commit labels.  Full lists here:
> http://march-hare.com/cvsnt/newfeatures/
>

First, I said 'free' development. So work done on your commercial
version of CVS isn't what I'm talking about - I know you are
developing your version constantly.

What I'm talking about when I say 'free development' is the
open-source version of CVS. Which hasn't had a release in 9 years, and
the last commit to the 'src' directory was 6 years ago.

Given that the back end is still a directory full of RCS files,
where's it going to go? And why?

There's very little user-facing that it could add that isn't already
covered by the other popular open-source VCS and DVCS solutions out
there already.

CVS could probably add something new and interesting on the back end,
but most people don't admin a VCS - they just use it. And the people
who really care about the back ends probably have budgets, so they can
pick up a commercial tool like CVSNT.

Your commercial tools certainly have a future, but open-source CVS
really does not. It's too late. Had they been doing something
interesting when SVN was getting going, CVS could probably have won
the day. But it didn't, and now open-source CVS is dead.

I could be wrong, of course (wouldn't be the first time), but with no
release in 9 years, I'd be VERY surprised if there was ever another
open-source CVS release.

I'd be absolutely stunned if it garnered any significant userbase.

>
> Regards,
>
>
> Arthur Barrett
> Product Manager



-- 
Michael Kohne
[email protected]

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot