Re: Pascal bindings

[dmitry boyarintsev <sk...@gmail.com>]

Hello Subversion-dev,

I can see, that there's no much of interest in Pascal bindings.
Well, that's quite understandable because of Pascal language not being popular.

Anyway, I'll publish the headers on my site.

Thank you for your great product!

regards,
Dmitry

Re: Pascal bindings

[Stefan Sperling <st...@elego.de>]

On Thu, Sep 02, 2010 at 09:27:57PM +0100, Julian Foad wrote:
> C. Michael Pilato wrote:
> > On 09/02/2010 01:38 PM, Philip Martin wrote:
> > > dmitry boyarintsev <sk...@gmail.com> writes:
> > > 
> > >> Hello Subversion-dev,
> > >>
> > >> I can see, that there's no much of interest in Pascal bindings.
> > >> Well, that's quite understandable because of Pascal language not
> > >> being popular.
> > >>
> > >> Anyway, I'll publish the headers on my site.
> > >>
> > >> Thank you for your great product!
> > > 
> > > You are more likely to get some response if you send a patch against
> > > trunk with a log message, see
> > > 
> > > http://subversion.apache.org/docs/community-guide/general.html#patches
> > > 
> > > Even if you do that I don't know that there is any great demand for
> > > Pascal bindings.  Are they generated or written by hand?  Do they have
> > > regression tests?  Are they tied to any particular environment?
> > 
> > All great questions, Philip.  Another one that's on my mind is, "Is Dmitry
> > volunteering to stick around and maintain these bindings?"  We really try to
> > avoid drop-and-ignore contributions of this sort, where none of the active
> > committership appears to be interested (or perhaps even qualified) to
> > maintain the new code.
> 
> Mike, Dmitry said in his original mail that he would support them.
> Quoting it in full:
> 
> > Hello SVN people.
> > 
> > As suggested by Hyrum K. Wright I'm forwarding the mail here.
> > 
> > I'm wandering if there's an official Pascal (FreePascal or Delphi)
> > bindings for the Subversion. I've converted C-headers (1.5.6 tag) for
> > my own needs, and I guess it would be easy for me to catch up with the
> > trunk as well.
> > 
> > If they're acceptable, I would like to contribute them for the
> > project, as well as support and update the bindings with the new
> > product releases.
> > 
> > Any suggestions? Do you need to have Pascal bindings?
> > 
> > thanks,
> > Dmitry
> 
> - Julian

I'd say send a patch and we'll see about including them.
It cannot hurt to have them around.

Though I'm afraid that not many people here will be able to digest
the patches. But should we reject new bindings just because of that?
The other bindings we have historically hardly been maintained,
but they survive. Just recently we've been getting nice contributions
in the python bindings area.

Dmitry, one thing you might not have taken into account is that we try
to run the binding tests continuously as part of our buildbot regression
test runs. Sometimes, the bindings linger in a state where the tests
on the buildbots keep failing for ages. Ideally, you should be prepared
to fix the bindings when tests fail, instead of just when we release.
We update and tweak the APIs continuously during development, and it's
nice when the bindings hop along.

But in any case, I see no reason why you shouldn't send a patch.

Stefan

Re: Pascal bindings

[Julian Foad <ju...@wandisco.com>]

C. Michael Pilato wrote:
> On 09/02/2010 01:38 PM, Philip Martin wrote:
> > dmitry boyarintsev <sk...@gmail.com> writes:
> > 
> >> Hello Subversion-dev,
> >>
> >> I can see, that there's no much of interest in Pascal bindings.
> >> Well, that's quite understandable because of Pascal language not
> >> being popular.
> >>
> >> Anyway, I'll publish the headers on my site.
> >>
> >> Thank you for your great product!
> > 
> > You are more likely to get some response if you send a patch against
> > trunk with a log message, see
> > 
> > http://subversion.apache.org/docs/community-guide/general.html#patches
> > 
> > Even if you do that I don't know that there is any great demand for
> > Pascal bindings.  Are they generated or written by hand?  Do they have
> > regression tests?  Are they tied to any particular environment?
> 
> All great questions, Philip.  Another one that's on my mind is, "Is Dmitry
> volunteering to stick around and maintain these bindings?"  We really try to
> avoid drop-and-ignore contributions of this sort, where none of the active
> committership appears to be interested (or perhaps even qualified) to
> maintain the new code.

Mike, Dmitry said in his original mail that he would support them.
Quoting it in full:

> Hello SVN people.
> 
> As suggested by Hyrum K. Wright I'm forwarding the mail here.
> 
> I'm wandering if there's an official Pascal (FreePascal or Delphi)
> bindings for the Subversion. I've converted C-headers (1.5.6 tag) for
> my own needs, and I guess it would be easy for me to catch up with the
> trunk as well.
> 
> If they're acceptable, I would like to contribute them for the
> project, as well as support and update the bindings with the new
> product releases.
> 
> Any suggestions? Do you need to have Pascal bindings?
> 
> thanks,
> Dmitry

- Julian

RE: Pascal bindings

["Bolstridge, Andrew" <an...@intergraph.com>]

> -----Original Message-----
> From: C. Michael Pilato [mailto:cmpilato@collab.net]
> Sent: Thursday, September 02, 2010 7:47 PM
> To: Philip Martin
> Cc: dmitry boyarintsev; dev@subversion.apache.org
> Subject: Re: Pascal bindings
> 
> On 09/02/2010 01:38 PM, Philip Martin wrote:
> > dmitry boyarintsev <sk...@gmail.com> writes:
> >
> >> Hello Subversion-dev,
> >>
> >> I can see, that there's no much of interest in Pascal bindings.
> >> Well, that's quite understandable because of Pascal language not
> >> being popular.
> >>
> >> Anyway, I'll publish the headers on my site.
> >>
> >> Thank you for your great product!
> >
> > You are more likely to get some response if you send a patch against
> > trunk with a log message, see
> >
> > http://subversion.apache.org/docs/community-guide/general.html#patches
> >
> > Even if you do that I don't know that there is any great demand for
> > Pascal bindings.  Are they generated or written by hand?  Do they have
> > regression tests?  Are they tied to any particular environment?
> 
> All great questions, Philip.  Another one that's on my mind is, "Is Dmitry
> volunteering to stick around and maintain these bindings?"  We really try to
> avoid drop-and-ignore contributions of this sort, where none of the active
> committership appears to be interested (or perhaps even qualified) to
> maintain the new code.

While that's an ideal position to be in, isn't there room for a 'second class' addition to the main codebase where the code is present, hopefully working, yet comes with no warranty or is part of the automated test set. 

The alternative is that valuable code would end up being posted to some website and wouldn’t have the visibility in the future if someone did come along, decided that they wanted the feature, and then spent the effort to fix it up (assuming it was broken).

As was mentioned, the python bindings were not maintained until recently when someone did send contributions in. If the python bindings were not present at all, this would not have happened.

So I'd say a pragmatic approach would be to take the contribution, and place it off to the side slightly compared to the 'tested, maintained' codebase and have the tests available to run manually. Maybe if someone comes along to maintain the code, then it can be moved to being a 'full' member of the codebase. 


**** not that that is an issue in this case as Dmitry says he will actively maintain it, and other C bindings, although it appears he is already slightly alienated from contributing given his response.

Re: Pascal bindings

["C. Michael Pilato" <cm...@collab.net>]

On 09/02/2010 01:38 PM, Philip Martin wrote:
> dmitry boyarintsev <sk...@gmail.com> writes:
> 
>> Hello Subversion-dev,
>>
>> I can see, that there's no much of interest in Pascal bindings.
>> Well, that's quite understandable because of Pascal language not
>> being popular.
>>
>> Anyway, I'll publish the headers on my site.
>>
>> Thank you for your great product!
> 
> You are more likely to get some response if you send a patch against
> trunk with a log message, see
> 
> http://subversion.apache.org/docs/community-guide/general.html#patches
> 
> Even if you do that I don't know that there is any great demand for
> Pascal bindings.  Are they generated or written by hand?  Do they have
> regression tests?  Are they tied to any particular environment?

All great questions, Philip.  Another one that's on my mind is, "Is Dmitry
volunteering to stick around and maintain these bindings?"  We really try to
avoid drop-and-ignore contributions of this sort, where none of the active
committership appears to be interested (or perhaps even qualified) to
maintain the new code.

-- 
C. Michael Pilato <cm...@collab.net>
CollabNet   <>   www.collab.net   <>   Distributed Development On Demand

Pascal bindings

[dmitry boyarintsev <sk...@gmail.com>]

On Thu, Sep 2, 2010 at 5:38 PM, Philip Martin
<ph...@wandisco.com> wrote:
> You are more likely to get some response if you send a patch against
> trunk with a log message, see
> http://subversion.apache.org/docs/community-guide/general.html#patches
Thanks for the link. I'm not ready yet with the trunk version bindings.

> Even if you do that I don't know that there is any great demand for Pascal bindings.
I'm not sure if there's any demand on them. It feels like I'm the only
one who needs them. I've written a front-end for the subversion
command-line client with FPC, and I'm not satisfied with it. The
problem is parsing svn output. XML feature is good, but it cannot be
used for all cases. And sometimes it hits the performance. Using the
library seems to me the better solution.

The new Delphi (which is ObjectPascal IDE) is proud of its subversion
integration, but i'm sure it's written in C/C++, so they didn't have
to use Pascal bindings.

> Are they generated or written by hand?
They're 95% generated. Except for svn_error_code.h, since it uses
macros heavily, and the parser is not yet good enough to handle them
properly. Some macros are also hand translated:
i.e macro (svn_pool.h)
#define svn_pool_clear apr_pool_clear
is translated as inline function by hand, since Pascal tends not to use macros.

> Do they have regression tests?
How can this be applied? The Pascal bindings are not providing any
kind of high-level wrappers over the library. All Subversion functions
and structures are used directly.

> Are they tied to any particular environment?
No. The bindings should be used with Free Pascal Compiler or Delphi
and should work on any system these compiler support.
I've tested them on Windows (SlikSVN dlls) and MacOSX (Xcode .dylibs)
and the bindings did work fine.

> Is Dmitry volunteering to stick around and maintain these bindings?
Yes, I am volunteering. if the bindings are accepted (acceptable). I'm
also helping FPC project to maintain other C-library bindings (i.e.
OpenGL, OpenCL)

thanks,
dmitry

Re: Pascal bindings

[Philip Martin <ph...@wandisco.com>]

dmitry boyarintsev <sk...@gmail.com> writes:

> Hello Subversion-dev,
>
> I can see, that there's no much of interest in Pascal bindings.
> Well, that's quite understandable because of Pascal language not
> being popular.
>
> Anyway, I'll publish the headers on my site.
>
> Thank you for your great product!

You are more likely to get some response if you send a patch against
trunk with a log message, see

http://subversion.apache.org/docs/community-guide/general.html#patches

Even if you do that I don't know that there is any great demand for
Pascal bindings.  Are they generated or written by hand?  Do they have
regression tests?  Are they tied to any particular environment?

-- 
Philip