vcs workdir abstraction lib

[RonnyPfannschmidt]

hi,
i recently started to extract anyvc from pida.

i put the development repo into http://www.bitbucket.org/RonnyPfannschmidt/anyvc/

current state:

simple workdir ops and status information tests for svn, bzr, hg,
darcs
status informations are completely broked for git
monotone is degraded to untested again

no cvs/rcs

please drop me some comments

Regards, Ronny

[Nicolas Chauvat]

Hi,

On Tue, Sep 16, 2008 at 12:57:02AM -0700, RonnyPfannschmidt wrote:
> i recently started to extract anyvc from pida.
>
> i put the development repo into http://www.bitbucket.org/RonnyPfannschmidt/anyvc/

For version control abstraction, see also vcslib in logilab-devtools.

http://www.logilab.org/project/logilab-devtools
http://www.logilab.org/cgi-bin/hgwebdir.cgi/logilab/devtools/file/695fae46619a/vcslib/

Any chance we can share code on this topic ?

--
Nicolas Chauvat

logilab.fr - services en informatique scientifique et gestion de connaissances

[RonnyPfannschmidt]

On 16 Sep., 10:09, Nicolas Chauvat <nicolas.chau...@logilab.fr> wrote:
> Hi,
>
> On Tue, Sep 16, 2008 at 12:57:02AM -0700, RonnyPfannschmidt wrote:
> > i recently started to extract anyvc from pida.
>

> > i put the development repo intohttp://www.bitbucket.org/RonnyPfannschmidt/anyvc/

>
> For version control abstraction, see also vcslib in logilab-devtools.
>

> http://www.logilab.org/project/logilab-devtoolshttp://www.logilab.org/cgi-bin/hgwebdir.cgi/logilab/devtools/file/695...

>
> Any chance we can share code on this topic ?

sure

as far as i can see your api's seem to be centered around subprocess
and cvs
i want to keep anyvc as agnostic as possible, so each vcs class is
pretty much a self-contained controller that does stuff to the workdir/
repo.

history abstractions are on my wishlist, i just didn't get to actually
start implementing them due to the other issues i need to address
first

>
> --
> Nicolas Chauvat
>
> logilab.fr - services en informatique scientifique et gestion de connaissances  

Regards Ronny

[Paul Boddie]

On 16 Sep, 10:09, Nicolas Chauvat <nicolas.chau...@logilab.fr> wrote:
>
> On Tue, Sep 16, 2008 at 12:57:02AM -0700, RonnyPfannschmidt wrote:
> > i recently started to extract anyvc from pida.
>

> > i put the development repo intohttp://www.bitbucket.org/RonnyPfannschmidt/anyvc/

>
> For version control abstraction, see also vcslib in logilab-devtools.
>

> http://www.logilab.org/project/logilab-devtoolshttp://www.logilab.org/cgi-bin/hgwebdir.cgi/logilab/devtools/file/695...

>
> Any chance we can share code on this topic ?

There's also the code in Tailor, although I can't comment on any APIs
that may be exposed.

http://progetti.arstecnica.it/tailor

My experiences with Tailor have been mostly positive, although it's
usually a struggle finding out what options need to be set in order to
get it to do its work.

I believe the Bazaar people also have libraries for interacting with
other repository formats, thinking back to presentations at EuroPython
2006, but I've never really looked at it.

Paul

[Cody Precord]

Hello,

On Sep 16, 2:57 am, RonnyPfannschmidt <ronny.pfannschm...@gmail.com>
wrote:

> hi,
> i recently started to extract anyvc from pida.
>

> i put the development repo intohttp://www.bitbucket.org/RonnyPfannschmidt/anyvc/

>
> current state:
>
> simple workdir ops and status information tests for svn, bzr, hg,
> darcs
> status informations are completely broked for git
> monotone is degraded to untested again
>
> no cvs/rcs
>
> please drop me some comments

In Editra the Projects plugin provides an abstraction layer that
currently has implementations for (cvs/svn/git/bazaar) and probably
will be adding hg support soon. It supports most common commands
(update/status/checkin/revert/history/diff/remove/add) for each.

The SourceControl library itself is separate from any dependancies but
the UI that uses it shares some code with Editra so without a little
work I don't think the ui will run standalone anymore. The
SourceControl only provides an interface but I have been currently
working to separate the command portion of the api from the gui code
so that it can be used more easily. In the current svn most of the
commands and subprocess/threading code has been moved out but still
has some ties to a few things that I hope to rework soon.

Anyway more information can be found at these links:
http://code.google.com/p/editra-plugins/wiki/ProjectsPlugin
http://code.google.com/p/editra-plugins/source/browse/#svn/trunk/Projects/projects

To see it in action you can download Editra and install it through the
Plugin Manager:
http://editra.org


Cody

[RonnyPfannschmidt]

i took a short look at the code and i really didn't like what i saw,
sorry but i dont't want code like that.

>
> Anyway more information can be found at these links:http://code.google.com/p/editra-plugins/wiki/ProjectsPluginhttp://code.google.com/p/editra-plugins/source/browse/#svn/trunk/Proj...

>
> To see it in action you can download Editra and install it through the
> Plugin Manager:http://editra.org
>
> Cody

Ronny

[Cody Precord]

Greetings,

On Sep 16, 2008, at 12:30 PM, RonnyPfannschmidt wrote:

>
> i took a short look at the code and i really didn't like what i saw,
> sorry but i dont't want code like that.

Politeness goes a long way in this world. I was being polite in
extending a reply to your message and providing you with another
source of reference that could possibly help with what you are doing,
and now I receive without reason what I can only perceive as an insult
in reply. Which is quite uncalled for and rather unnecessary. Simply
saying it is not really what your looking for would have been quite
sufficient and a far more effective means of communication.

Good day,

Cody

[RonnyPfannschmidt]

Sorry if that came over harsh,
it wasn't my intention to insult you
but its the usual way i communicate with people about code,

That model works usually fine and it helps for the evolution of code.
(the inital iterations if anyvc i hacked up within pida where just as
horrible).

But after a few "wtf, this is shit, do it more nice, change this and
that" from others, it got more pretty, better testable and more
simple. (still a long way to go tho)

Unfortunately i sometimes fail to realize when im outside of of my
usual communication channels
and people consider that way of talking about code as offensive and
rude.

So i'm sorry for missbehaving and i will try to be polite now.

> Good day,
>
> Cody

Regards, Ronny