Deprecation

Yesterday I updated the deprecation.rst document such that I
think it 
contains all the items which are slated for deprecation (and
also 
contains full details of the upgrade suggestions, as Olly
suggested).

I've marked several things which have been deprecated for a
long period 
of time (ie, since 0.9.0 or earlier) for removal in 1.0.0,
in line with 
the policy outlined in the document.  If possible, I'd like
anyone 
interested to check over this list over the next few days,
so that we 
can get started on actually removing the functions to be
removed in 1.0.0.

I've also marked items in the bindings which should be
deprecated: most 
of these have been marked as "out-of-date" in
xapian.i for ages, but not 
clearly documented publically, so it's perhaps a bit more
controversial 
whether they should be removed now.  On the other hand, the
bindings 
have been less stable than the core until now, and it would
be easier to 
remove them at the same time as removing the equivalent from
the core.

Finally, I'm not convinced that the replacement for
Stem::stem_word() of 
using operator() will work for all the bindings - checking
whether it 
can be accessed from PHP, Ruby, Tcl8, and CSharp would be
useful 
(ideally, changing smoketest and the examples to demonstrate
how).  It 
works in Python, though. 

-- 
Richard

_______________________________________________
Xapian-devel mailing list
Xapian-devellists.xapian.org
http://lists.xapian.org/mailman/listinfo/xapian-devel

Oh - anyone interested might want to read and subscribe to
the relevant 
bug, too:
http://www.xapian.org/cgi-bin/bugzilla/show_bug.cgi?id
=126

-- 
Richard

_______________________________________________
Xapian-devel mailing list
Xapian-devellists.xapian.org
http://lists.xapian.org/mailman/listinfo/xapian-devel

On Tue, Apr 17, 2007 at 10:17:19AM +0100, Richard Boulton
wrote:
> Finally, I'm not convinced that the replacement for
Stem::stem_word() of 
> using operator() will work for all the bindings -
checking whether it 
> can be accessed from PHP, Ruby, Tcl8, and CSharp would
be useful 
> (ideally, changing smoketest and the examples to
demonstrate how).  It 
> works in Python, though. 

Except for Python, operator() is renamed to apply() (or
Apply() for C#).
See xapian.i around line 122 (in current SVN HEAD).

Cheers,
    Olly

_______________________________________________
Xapian-devel mailing list
Xapian-devellists.xapian.org
http://lists.xapian.org/mailman/listinfo/xapian-devel

Olly Betts wrote:
> On Tue, Apr 17, 2007 at 10:17:19AM +0100, Richard
Boulton wrote:
>> Finally, I'm not convinced that the replacement for
Stem::stem_word() of 
>> using operator() will work for all the bindings -
checking whether it 
>> can be accessed from PHP, Ruby, Tcl8, and CSharp
would be useful 
>> (ideally, changing smoketest and the examples to
demonstrate how).  It 
>> works in Python, though. 
> 
> Except for Python, operator() is renamed to apply() (or
Apply() for C#).
> See xapian.i around line 122 (in current SVN HEAD).

Interesting.  apply() works for PHP and TCL8, I haven't
tested with C#, 
but doesn't work for Ruby.

Looking at the generated code, Ruby provides access to
operator() with a 
call() method: ie, stem.call() works with Ruby, stem.apply() doesn't.

I assume that this is a bug in SWIG, or something like that.
 I've 
opened a bug in Xapian (#131) to make a note of it.

I've updated the smoketests and examples for PHP and TCL to
use 
stem.apply().

I've also updated the smoketest and examples for CSharp, but
haven't 
tested them because the CSharp bindings don't work on my
machine 
presently (see bug #132).

-- 
Richard

_______________________________________________
Xapian-devel mailing list
Xapian-devellists.xapian.org
http://lists.xapian.org/mailman/listinfo/xapian-devel

On Tue, Apr 17, 2007 at 10:17:19AM +0100, Richard Boulton
wrote:
> I've also marked items in the bindings which should be
deprecated: most 
> of these have been marked as "out-of-date" in
xapian.i for ages, but not 
> clearly documented publically, so it's perhaps a bit
more controversial 
> whether they should be removed now.  On the other hand,
the bindings 
> have been less stable than the core until now, and it
would be easier to 
> remove them at the same time as removing the equivalent
from the core.

I don't think it's sensible to go down the road of keeping
deprecated
aliases and compatibility code in the bindings when the
corresponding
code has been removed from the C++ library.  And the
bindings
documentation suggests referring to the C++ API, which does
mark them as
deprecated.

Cheers,
    Olly

_______________________________________________
Xapian-devel mailing list
Xapian-devellists.xapian.org
http://lists.xapian.org/mailman/listinfo/xapian-devel

Olly Betts wrote:
> I don't think it's sensible to go down the road of
keeping deprecated
> aliases and compatibility code in the bindings when the
corresponding
> code has been removed from the C++ library.  And the
bindings
> documentation suggests referring to the C++ API, which
does mark them as
> deprecated.

That makes things simple. Agreed. 

(Which is just as well, since you've just removed them.  )

-- 
Richard

_______________________________________________
Xapian-devel mailing list
Xapian-devellists.xapian.org
http://lists.xapian.org/mailman/listinfo/xapian-devel