[bch%methodlogic.net@localhost: Re: cvs->git?]

[bch%methodlogic.net@localhost]

re-injecting back into ML to fix Adam's "reply" vs "reply-all"

On Thu, Aug 05, 2010 at 12:05:19AM +0200, Adam Hamsik wrote:
> Hi,
> On Aug,Thursday 5 2010, at 1:35 AM, bch%methodlogic.net@localhost wrote:
> 
> > On Wed, Aug 04, 2010 at 11:13:40PM +0000, bch%methodlogic.net@localhost 
> > wrote:
> >> On Wed, Aug 04, 2010 at 10:11:26PM +0200, Jan ?mydke wrote:
> >>> So to sum up my impression of your replies:
> >>> 
> >>> - There is no strong desire to change the version control, mainly
> >>>  because the pkgsrc users do not see much benefit of doing it
> >>> 
> >>> - But the users are open to adopt it as long as NetBSD project switches
> >>>  from CVS to something else
> >>> 
> >>> Two additional notes:
> >>> 
> >>> - I would definitely benefit from any decentralized VCS. At work I
> >>>  cannot use 'cvs up' because of a firewall... so I have to stupidly
> >>>  download the pkgsrc-wip-snapshot and unpack... every time I want to
> >>>  update. Instead I could update from a git-repo from my other site very
> >>>  easily...
> >>> 
> >>> - Still thinking over the problem of the very limited pkgsrc
> >>>  (and NetBSD) community. I don't dare to claim a faster VCS would flood
> >>>  the project with new developers, but I believe it is one of the most
> >>>  significant obstacles. When a Linux-bored guy checks NetBSD and sees
> >>>  quite a messy web-site, he may hesitate, but finding out it is all
> >>>  CVS-based - what? Sorry, bye. (What a pity for him, and for us!)
> >>> 
> >>> Best,
> >>> 
> >>> JS
> >> 
> >> I was alerted to this thread in #netbsd, and am getting into thread
> >> at first opportunity after subscribing to ML, so excuse non-ideal
> >> placement :)
> >> 
> >> To put another candidate into the mix which -does-
> >> have a hope of inclusion in NetBSD, please see fossil
> >> (http://fossil-scm.org), and see tech-repository thread at
> >> http://mail-index.netbsd.org/tech-repository/2010/01/18/msg000345.html
> >> for the discussion that occurred there.
> >> 
> >> Looking forward to more exploration!
> > 
> > Sorry to reply to self, but other pertinent info: 
> > http://methodlogic.net/BSDFossil.html
> > 
> > ...and near the end of the tech-kern thread I mentioned previously,
> > note that drh hisself (principle dev of fossil) indicated he'd be happy
> > to assist as he's able.
> 
> I have some fossil questions for you :) 
> 
> 1) Can I migrate CVS repository with whole history to fossil-scm ? if yes how 
> ?

Heh... is a work-in-progress itself... I was working on it with another
dev (haven't touched lately) for the general case, but also, SQLite was
ported CVS-> fossil. (author of SQLite == author fossil).

> 2) Have you tried to use fossil with NetBSD src or pkgsrc ? do you have any 
> public available repositories ?

Yes I have tried, no I don't have public repos. I did some kernel hacking
getting iwn(4) supporting a new NIC, and used fossil to manage my local
changes, managing all of /src. Honestly, I found it could be sluggish
at times, but didn't investigate further.

I also manage my own pkgsrc, and my pkgsrc-wip projects in fossil.

> 3) Can I remove/disable wiki/ticket/ whatever system from fossil build ?

It's designed to be a whole. The tickets, wiki-entries, and
change sets are atomic elems (artifacts) that are described here:
http://www.fossil-scm.org/index.html/doc/tip/www/concepts.wiki

> 4) How efficient is fossil with disk space, are you using compression ?

Fossil certainly uses zip compression across wire making for efficient
on-the-wire protocol (indeed, I've personally used it across slow
modem connections), but I'm not sure if it's compressed/decompressed on
disk. It's a good question... I'll look into it.

> Thank you for answering my question, I would like to convert our repo to 
> fossil if there are any tools available. I already tried to achieve that with 
> (git, hg) tools and they was only partially successful.
> 
> If we really want to think about moving to different SCM we need to be able 
> to convert our repo to it with whole history.

No question history is critical. I'll look to find some cycles and play w/
the pkgsrc-wip repo w/ fossil to see if I can convert.

> Thanks for your time and responses.
> 
> Regards
> 
> Adam.
> 

-- 
Brad Harder
Method Logic Digital Consulting
http://methodlogic.net
http://twitter.com/bcharder


----- End forwarded message -----

-- 
Brad Harder
Method Logic Digital Consulting
http://methodlogic.net
http://twitter.com/bcharder


------------------------------------------------------------------------------
The Palm PDK Hot Apps Program offers developers who use the
Plug-In Development Kit to bring their C/C++ apps to Palm for a share
of $1 Million in cash or HP Products. Visit us here for more details:
http://p.sf.net/sfu/dev2dev-palm
_______________________________________________
pkgsrc-wip-discuss mailing list
pkgsrc-wip-discuss%lists.sourceforge.net@localhost
https://lists.sourceforge.net/lists/listinfo/pkgsrc-wip-discuss