[ECOS] Eclipse support for git

[ECOS] Eclipse support for git

[Øyvind Harboe]

I think it is important when choosing the next version control
system to think about whether we're getting onto the right
train.

The support today is important, yes, but what happens
2-3 years from now?

I'm confident that git will receive *plenty* of attention from
GUI makers as the user base continues to expand due
to the killer references that git has.

Eclipse support anyone?

http://www.eclipse.org/proposals/egit/



-- 
Øyvind Harboe
http://www.zylin.com/zy1000.html
ARM7 ARM9 ARM11 XScale Cortex
JTAG debugger and flash programmer

--
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

Re: [ECOS] Eclipse support for git

[Alex Schuilenburg]

Øyvind Harboe wrote on 2009-09-22 09:33:
> I think it is important when choosing the next version control
> system to think about whether we're getting onto the right
> train.
>
> The support today is important, yes, but what happens
> 2-3 years from now?
>
> I'm confident that git will receive *plenty* of attention from
> GUI makers as the user base continues to expand due
> to the killer references that git has.
>
> Eclipse support anyone?
>
> http://www.eclipse.org/proposals/egit/
>   
but just as git closes the gap wrt GUI support, hg will close the gap
wrt functionality...

You cannot make assumptions based on a moving target or what development
is likely to occur in one camp without considering the same for the
others.  Also, just because git is right for linux does not mean it is
right for eCos.  As for killer app, exactly how much of that
functionality will be used by the average Joe? 5%, 10%?  And the average
full time user? 50%, 60%?  From what I can tell wrt git, if Linus wants
to do something that git cannot do with a couple of simple commands,
Linus just goes and writes another command for that. Not that this is a
bad thing, just that for the average Joe it is another command/option
that is thrown into the mix. I believe in the KISS principle when it
comes to considering the needs of the masses and am not a great fan of
feature-bloat.

Of course we could always make a political decision and choose bazaar
because it is a GNU project.  I know of at least one project that was
forced to do so by RMS for that very reason... 

Anyway, the great thing about all of these DRCS systems is that, if
there ever is a clear winner, their designs are so similar, migration
from one to the other in theory should be very simple.  All these
systems have export/import functionality for this specific purpose.

So if we want to move to a DRCS, I suggest we do it sooner rather than
later or wait another year or so to see if there is a new DRCS leader or
someone comes up with a MDDRCS  (multi-dimensional distributed ...) or
new flavour of the month.

FWIW, Mercurial already has Eclipse support for a couple of years now:
  http://www.vectrace.com/mercurialeclipse/

Finally, dont get me wrong.  From my perspective I would like to see the
right technical choice being made for a DRCS for eCos, or for anything
else added to eCos.  If that turns out to be git, then great, I will use
it with pleasure.  If the technical evaluation comes up with little
differences between the choices, the evaluation should then look to
usability, documentation, testing, support, usage, etc.  In my own
evaluation of DRCS, this is where I see differences in many directions. 
Some favour git, some favour hg.

The hard part is deciding which differences really matter, so I am glad
I am not the one actually making the final choice.  I personally agree
with the author of this evaluation's summarising comments:
 
https://ldn.linuxfoundation.org/article/dvcs-roundup-one-system-rule-them-all-part-2

Namely:
If you really want a direct recommendation, it would be thus: download
Git and Mercurial, then throw dice. Personally I like the “it just
works” approach of Mercurial a bit better, and the fact that it is
written in Python. Also for cross-platform projects needing Windows
support Mercurial is (at this moment) probably a tiny bit less
problematic. If you work in a completely UNIX-centered environment, Git
might be the slightly better choice. However, both systems work very
well, are stable and very, very fast. So it's a tie, more or less."

-- Alex Schuilenburg

   >>>> Visit us at ESC-Boston  http://www.embedded.com/esc/boston <<<<
   >>>> Sep 22-23 on Stand 226  at Hynes Convention Center, Boston <<<<

       >>>> Visit us at ESC-UK  http://www.embedded.co.uk <<<<
       >>>> Oct 7-8 on Stand 433 at FIVE ISC, Farnborough <<<<


-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

Re: [ECOS] Eclipse support for git

[Øyvind Harboe]

From what I know about mercurial and git, I'm happy to get behind
whatever choice the eCos maintainers end up with. The only choice
I'm not happy with is CVS.

Personally I'm not excited about learning one more source control
system, but it's nice to know that the eCos maintainers are probably
not going to generalize from my preferences :-)


-- 
Øyvind Harboe
http://www.zylin.com/zy1000.html
ARM7 ARM9 ARM11 XScale Cortex
JTAG debugger and flash programmer

--
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

Re: [ECOS] Eclipse support for git

[Jonathan Larmour]

Alex Schuilenburg wrote:
> Øyvind Harboe wrote on 2009-09-22 09:33:
> 
>>I think it is important when choosing the next version control
>>system to think about whether we're getting onto the right
>>train.
>>
>>The support today is important, yes, but what happens
>>2-3 years from now?

I think in all cases, we should look at the state now. They are all 
deployed in high profile projects and aren't going away soon. Due to their 
better handling of metadata (compared to CVS certainly) it's also much 
more feasible to swap between them, as many people do. We absolutely want 
to make a good choice, but practically speaking this is not truthfully a 
"forever" decision. Oh oops, yes Alex says that too.

> https://ldn.linuxfoundation.org/article/dvcs-roundup-one-system-rule-them-all-part-2
> 
> Namely:
> If you really want a direct recommendation, it would be thus: download
> Git and Mercurial, then throw dice. Personally I like the “it just
> works” approach of Mercurial a bit better, and the fact that it is
> written in Python. Also for cross-platform projects needing Windows
> support Mercurial is (at this moment) probably a tiny bit less
> problematic. If you work in a completely UNIX-centered environment, Git
> might be the slightly better choice. However, both systems work very
> well, are stable and very, very fast. So it's a tie, more or less."

For what it's worth, despite being a linux fanboy myself obviously ;-), 
considering the wider eCos community is a priority. And as such, usability 
on Windows is a concern for me with git. And it's overall learning curve 
(the "it just works" above) inherent to its design. I would slightly 
prefer something that disenfranchised the small number of elite power 
users to something that disenfranchised a much larger number of less 
experienced users.

After all, one of the reasons for the popularity of CVS is its simplicity. 
Talking about octopus merges with an eCos n00b is not going to get people 
very far.

Of course one of the joys of DRCS's is the now-widespread ability to sync 
and mereg in from other repositories using a different DRCS. That's 
another reason why the power users will be able to get by. It also means 
that the real decision is indeed probably going to be with the maintainers 
as the ones who have to consider the wider eCos userbase as well as being 
the ones that definitely have to interact with that particular VCS.

The discussions so far seem have been a bit ad hoc. I'd like to bring them 
more on track. I personally agree with Alex's view that the choice of VCS 
is probably between Git, Hg, and Bzr. If anyone would like to suggest SVN, 
Arch or something more exotic, it's time to speak up.

Then we can do specific head-to-head comparisons between the three, which 
is tractable.

Jifl
-- 
--["No sense being pessimistic, it wouldn't work anyway"]-- Opinions==mine

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss

[ECOS] Re: Eclipse support for git

[Sergei Organov]

Jonathan Larmour <jifl@jifvik.org> writes:

[...]

> The discussions so far seem have been a bit ad hoc. I'd like to bring
> them more on track. I personally agree with Alex's view that the
> choice of VCS is probably between Git, Hg, and Bzr. If anyone would
> like to suggest SVN, Arch or something more exotic, it's time to speak
> up.

Even though I'm a git proponent, I'd suggest to stick to Hg for now as
the only person who is doing real work prefers Hg.

-- Sergei.


-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss