Windows, Enchant and Aspell

2009-02-08 Thread Enrico Tröger
Hi,

I just noticed Tor has built enchant for Windows [1], great job. Cool.

Being curious whether it works, I installed aspell and two dictionaries
from their Windows installers, I used the default directory C:\Program
Files \Aspell.

Unfortunately and not really surprising, libenchant.dll or better the
provider libenchant_aspell.dll, doesn't find the aspell-15.dll.
Is there any way to tell enchant where to look for the installed aspell
installation?
Another problem is that libenchant_aspell.dll looks for
libaspell-15.dll but the Aspell Windows installers install a
aspell-15.dll (without the lib prefix).

So, after all, how to use enchant with Aspell at all on Windows?


In some other old code, where I used Aspell directly, I used some code
to find the Aspell installation directory from the Windows registry. I
put the code on http://www.uvena.de/tmp/win32_find_aspell.c, maybe it's
useful in some way. Could this maybe be reused for the Aspell enchant
provider or is there an easier way which I don't see?


[1]
http://ftp.gnome.org/pub/gnome/binaries/win32/dependencies/enchant-1.4.2-tml.zip

Regards,
Enrico

-- 
Get my GPG key from http://www.uvena.de/pub.asc
___
gtk-app-devel-list mailing list
gtk-app-devel-list@gnome.org
http://mail.gnome.org/mailman/listinfo/gtk-app-devel-list

Re: Windows, Enchant and Aspell

2009-02-08 Thread Enrico Tröger
On Sun, 8 Feb 2009 14:12:20 -0500, Matthew Talbert
ransom1...@gmail.com wrote:

2009/2/8 Enrico Tröger enrico.troe...@uvena.de:
 Hi,

 I just noticed Tor has built enchant for Windows [1], great job.
 Cool.

 Being curious whether it works, I installed aspell and two
 dictionaries from their Windows installers, I used the default
 directory C:\Program Files \Aspell.

 Unfortunately and not really surprising, libenchant.dll or better the
 provider libenchant_aspell.dll, doesn't find the aspell-15.dll.
 Is there any way to tell enchant where to look for the installed
 aspell installation?
 Another problem is that libenchant_aspell.dll looks for
 libaspell-15.dll but the Aspell Windows installers install a
 aspell-15.dll (without the lib prefix).

 So, after all, how to use enchant with Aspell at all on Windows?

I just ran into this. What I did was install Aspell, change the name
of aspell-15.dll to libaspell-15.dll, copy everything over to my app's
directory, and it all just works.

Hmm yes. I tried this as well and it works. But this is kind of hard to
explain to users and it's not very convenient at all.
Furthermore, I don't want to ship Aspell with my application and
especially not the Aspell dictionaries. 


Regards,
Enrico

-- 
Get my GPG key from http://www.uvena.de/pub.asc
___
gtk-app-devel-list mailing list
gtk-app-devel-list@gnome.org
http://mail.gnome.org/mailman/listinfo/gtk-app-devel-list

GtkOrientable: Can widgets now be flipped?

2009-02-08 Thread Murray Cumming
The new GtkOrientable's documentation
http://library.gnome.org/devel/gtk/unstable/gtk-Orientable.html#gtk-Orientable.description
says:
Historically, such widgets have been realized as subclasses of a common
base class (e.g GtkBox/GtkHBox/GtkVBox and
GtkScale/GtkHScale/GtkVScale). GtkOrientable is more flexible in that it
allows the orientation to be changed at runtime, allowing the widgets to
'flip'.

All those old widgets do now implement GtkOrientation:
http://library.gnome.org/devel/gtk/unstable/gtk-Orientable.html#gtk-Orientable.implementations

So can they actully be flipped at runtime now? The documentation seems
to suggest that they can't if they still use the old common base
classes.

-- 
Murray Cumming
murr...@murrayc.com
www.murrayc.com
www.openismus.com


___
gtk-devel-list mailing list
gtk-devel-list@gnome.org
http://mail.gnome.org/mailman/listinfo/gtk-devel-list


Re: GtkOrientable: Can widgets now be flipped?

2009-02-08 Thread Hagen Schink
Hi!

On Sun, 08 Feb 2009 14:51:10 +0100
Sven Herzberg herzi...@gnome-de.org wrote:

 AFAIR, you can flip it of you create a GtkBox/GtkScale, but not if you
 create a Gtk[HV]{Box,Scale}... Mitch?

Gtk[HV]Box sets only the property in init. So it should be possible to change 
the orientation.

In the set_property function of GtkBox a resize is queued when you set the 
orientation which means that you can flip the orientation at runtime. I guess 
it is also possible in Gtk[HV]Box because it inherits the property.

In my animation demo my GtkAnimBox inherits from GtkBox and flipping is done 
only by setting the orientation property.

I hope this helps, have a nice Sunday!

Hagen
___
gtk-devel-list mailing list
gtk-devel-list@gnome.org
http://mail.gnome.org/mailman/listinfo/gtk-devel-list


Re: GtkOrientable: Can widgets now be flipped?

2009-02-08 Thread Christian Dywan
Am Sun, 08 Feb 2009 13:07:34 +0100
schrieb Murray Cumming murr...@murrayc.com:

 The new GtkOrientable's documentation
 http://library.gnome.org/devel/gtk/unstable/gtk-Orientable.html#gtk-Orientable.description
 says:
 Historically, such widgets have been realized as subclasses of a
 common base class (e.g GtkBox/GtkHBox/GtkVBox and
 GtkScale/GtkHScale/GtkVScale). GtkOrientable is more flexible in that
 it allows the orientation to be changed at runtime, allowing the
 widgets to 'flip'.
 
 All those old widgets do now implement GtkOrientation:
 http://library.gnome.org/devel/gtk/unstable/gtk-Orientable.html#gtk-Orientable.implementations
 
 So can they actully be flipped at runtime now? The documentation seems
 to suggest that they can't if they still use the old common base
 classes.

This is how the current state is to my awareness:

It is meant to work so that subclasses such a GtkHFoo and GtkVFoo
behave the same as they used to, and can't flip. However if you use the
originally abstract base class 'orientation' works as expected, at
construction as well as runtime.

Note that GtkBox is still abstract in trunk although it technically
supports 'orientation' because possible changes in default values of
properties has to be discussed. So you can not use it.

ciao,
Christian
___
gtk-devel-list mailing list
gtk-devel-list@gnome.org
http://mail.gnome.org/mailman/listinfo/gtk-devel-list


Re: GtkOrientable: Can widgets now be flipped?

2009-02-08 Thread Murray Cumming
On Sun, 2009-02-08 at 14:51 +0100, Sven Herzberg wrote:
 AFAIR, you can flip it of you create a GtkBox/GtkScale, but not if you
 create a Gtk[HV]{Box,Scale}... Mitch?

Ah. Were those base classes (GtkBox, etc) instantiable before?

If the base classes can now already do everything then shouldn't the V/H
classes be deprecated already?
 
-- 
Murray Cumming
murr...@murrayc.com
www.murrayc.com
www.openismus.com


___
gtk-devel-list mailing list
gtk-devel-list@gnome.org
http://mail.gnome.org/mailman/listinfo/gtk-devel-list


Re: GtkOrientable: Can widgets now be flipped?

2009-02-08 Thread Colomban Wendling
Christian Dywan a écrit :
 This is how the current state is to my awareness:

 It is meant to work so that subclasses such a GtkHFoo and GtkVFoo
 behave the same as they used to, and can't flip. However if you use the
 originally abstract base class 'orientation' works as expected, at
 construction as well as runtime.
Hi,

Since all of those widgets implements GtkOrientable, it seems obvious
for me that they can be flipped as any GtkOrientable can, and only
have a default orientation to behave as they used to; and the
documentation says, as I understand, only that the new design is more
flexible that the old now unused one, nothing more.
To be sure, I've tested to flip a GtkHBox with 2.15.4 from SVN, and
yes, it works perfectly.

Regards,
Colomban


___
gtk-devel-list mailing list
gtk-devel-list@gnome.org
http://mail.gnome.org/mailman/listinfo/gtk-devel-list


Re: GtkOrientable: Can widgets now be flipped?

2009-02-08 Thread Mathias Hasselmann
Am Sonntag, den 08.02.2009, 18:36 +0100 schrieb Murray Cumming:
 On Sun, 2009-02-08 at 14:51 +0100, Sven Herzberg wrote:
  AFAIR, you can flip it of you create a GtkBox/GtkScale, but not if you
  create a Gtk[HV]{Box,Scale}... Mitch?
 
 Ah. Were those base classes (GtkBox, etc) instantiable before?
 
 If the base classes can now already do everything then shouldn't the V/H
 classes be deprecated already?

Oh, I hear you volunteer to update the gazillions lines of code in the
wild refering those classes? ;-)

I am pretty sure API users would rather switch to Qt, than to follow
this deprecation.

Well, actually those V/H classes even serve a purpose: When using them
in a derived widget you can be sure, that nobody will be that evil to
suddenly change the orientation of your derived widget.

Ciao,
Mathias
-- 
Mathias Hasselmann mathias.hasselm...@gmx.de
Personal Blog: http://taschenorakel.de/mathias/
Openismus GmbH: http://www.openismus.com/

___
gtk-devel-list mailing list
gtk-devel-list@gnome.org
http://mail.gnome.org/mailman/listinfo/gtk-devel-list


Re: GtkOrientable: Can widgets now be flipped?

2009-02-08 Thread Murray Cumming
On Sun, 2009-02-08 at 18:47 +0100, Mathias Hasselmann wrote:
 Am Sonntag, den 08.02.2009, 18:36 +0100 schrieb Murray Cumming:
  On Sun, 2009-02-08 at 14:51 +0100, Sven Herzberg wrote:
   AFAIR, you can flip it of you create a GtkBox/GtkScale, but not if you
   create a Gtk[HV]{Box,Scale}... Mitch?
  
  Ah. Were those base classes (GtkBox, etc) instantiable before?
  
  If the base classes can now already do everything then shouldn't the V/H
  classes be deprecated already?

A deprecation doesn't force people to change anything. It's not (yet) a
removal of the API, though I guess that is planned for the future. It is
a useful hint and it explains the apparent duplication.

 Oh, I hear you volunteer to update the gazillions lines of code in the
 wild refering those classes? ;-)
 
 I am pretty sure API users would rather switch to Qt, than to follow
 this deprecation.
 
 Well, actually those V/H classes even serve a purpose: When using them
 in a derived widget you can be sure, that nobody will be that evil to
 suddenly change the orientation of your derived widget.

Apparently that's not true though.

-- 
Murray Cumming
murr...@murrayc.com
www.murrayc.com
www.openismus.com


___
gtk-devel-list mailing list
gtk-devel-list@gnome.org
http://mail.gnome.org/mailman/listinfo/gtk-devel-list


Re: GtkOrientable: Can widgets now be flipped?

2009-02-08 Thread Mathias Hasselmann
Am Sonntag, den 08.02.2009, 21:04 +0100 schrieb Murray Cumming:
 On Sun, 2009-02-08 at 18:47 +0100, Mathias Hasselmann wrote:
  Am Sonntag, den 08.02.2009, 18:36 +0100 schrieb Murray Cumming:
   On Sun, 2009-02-08 at 14:51 +0100, Sven Herzberg wrote:
AFAIR, you can flip it of you create a GtkBox/GtkScale, but not if you
create a Gtk[HV]{Box,Scale}... Mitch?
   
   Ah. Were those base classes (GtkBox, etc) instantiable before?
   
   If the base classes can now already do everything then shouldn't the V/H
   classes be deprecated already?
 
 A deprecation doesn't force people to change anything. It's not (yet) a
 removal of the API, though I guess that is planned for the future. It is
 a useful hint and it explains the apparent duplication.

Still you have to react on deprecations better soon than later, if you
want to avoid drowning some day.

  Well, actually those V/H classes even serve a purpose: When using them
  in a derived widget you can be sure, that nobody will be that evil to
  suddenly change the orientation of your derived widget.
 
 Apparently that's not true though.

Oh, from reading this thread I thought this would be the case, but
apparently it is not. Seems you really can change the orientation of
some Gtk[HV]Box now. Shocking.

Maybe GtkBox and friends should get an internal flag for locking that
property? Whilst changing the orientation of widgets can be useful under
certain circumstances, there are others where you don't want to permit
this. Well, but maybe I think too restrictive right now, as you can do
nonsense to derived widgets already. Thinking about evil actions like
calling gtk_container_remove() on some essential child.

Ciao,
Mathias
-- 
Mathias Hasselmann mathias.hasselm...@gmx.de
Personal Blog: http://taschenorakel.de/mathias/
Openismus GmbH: http://www.openismus.com/

___
gtk-devel-list mailing list
gtk-devel-list@gnome.org
http://mail.gnome.org/mailman/listinfo/gtk-devel-list


Re: GtkOrientable: Can widgets now be flipped?

2009-02-08 Thread Murray Cumming
On Sun, 2009-02-08 at 21:36 +0100, Mathias Hasselmann wrote:
 Am Sonntag, den 08.02.2009, 21:04 +0100 schrieb Murray Cumming:
  On Sun, 2009-02-08 at 18:47 +0100, Mathias Hasselmann wrote:
   Am Sonntag, den 08.02.2009, 18:36 +0100 schrieb Murray Cumming:
On Sun, 2009-02-08 at 14:51 +0100, Sven Herzberg wrote:
 AFAIR, you can flip it of you create a GtkBox/GtkScale, but not if you
 create a Gtk[HV]{Box,Scale}... Mitch?

Ah. Were those base classes (GtkBox, etc) instantiable before?

If the base classes can now already do everything then shouldn't the V/H
classes be deprecated already?
  
  A deprecation doesn't force people to change anything. It's not (yet) a
  removal of the API, though I guess that is planned for the future. It is
  a useful hint and it explains the apparent duplication.
 
 Still you have to react on deprecations better soon than later, if you
 want to avoid drowning some day.

Yes, but surely that decision has been made already by the addition of
GtkOrientable?

 Oh, from reading this thread I thought this would be the case, but
 apparently it is not. Seems you really can change the orientation of
 some Gtk[HV]Box now. Shocking.

Actually Christian Dywan says they can't flip. But I still think they
are meant (by the developers who made the change) to be deprecated so
that isn't very relevant anyway.

-- 
Murray Cumming
murr...@murrayc.com
www.murrayc.com
www.openismus.com


___
gtk-devel-list mailing list
gtk-devel-list@gnome.org
http://mail.gnome.org/mailman/listinfo/gtk-devel-list


Re: GtkOrientable: Can widgets now be flipped?

2009-02-08 Thread Mathias Hasselmann
Am Sonntag, den 08.02.2009, 21:57 +0100 schrieb Murray Cumming:
 On Sun, 2009-02-08 at 21:36 +0100, Mathias Hasselmann wrote:
  Am Sonntag, den 08.02.2009, 21:04 +0100 schrieb Murray Cumming:
   On Sun, 2009-02-08 at 18:47 +0100, Mathias Hasselmann wrote:
Am Sonntag, den 08.02.2009, 18:36 +0100 schrieb Murray Cumming:
 On Sun, 2009-02-08 at 14:51 +0100, Sven Herzberg wrote:
  AFAIR, you can flip it of you create a GtkBox/GtkScale, but not if 
  you
  create a Gtk[HV]{Box,Scale}... Mitch?
 
 Ah. Were those base classes (GtkBox, etc) instantiable before?
 
 If the base classes can now already do everything then shouldn't the 
 V/H
 classes be deprecated already?
   
   A deprecation doesn't force people to change anything. It's not (yet) a
   removal of the API, though I guess that is planned for the future. It is
   a useful hint and it explains the apparent duplication.
  
  Still you have to react on deprecations better soon than later, if you
  want to avoid drowning some day.
 
 Yes, but surely that decision has been made already by the addition of
 GtkOrientable?
 
  Oh, from reading this thread I thought this would be the case, but
  apparently it is not. Seems you really can change the orientation of
  some Gtk[HV]Box now. Shocking.
 
 Actually Christian Dywan says they can't flip. But I still think they
 are meant (by the developers who made the change) to be deprecated so
 that isn't very relevant anyway.

It is relevant, as I'd seriously question a decision to deprecate
classes as widely used as Gtk[HV]Box.

The community's effort to follow that little deprecation is close to
infiniteness. Keeping Gtk[HV]Box around costs exactly 54 lines of code
and four public symbols.

Ciao,
Mathias
-- 
Mathias Hasselmann mathias.hasselm...@gmx.de
Personal Blog: http://taschenorakel.de/mathias/
Openismus GmbH: http://www.openismus.com/

___
gtk-devel-list mailing list
gtk-devel-list@gnome.org
http://mail.gnome.org/mailman/listinfo/gtk-devel-list


Re: GtkOrientable: Can widgets now be flipped?

2009-02-08 Thread Tristan Van Berkom
[...]
 If the base classes can now already do everything then shouldn't the 
 V/H
 classes be deprecated already?

Gtk[VH]Foo deprecation++ from me, yes it should be deprecated if it doesnt
serve a purpose - hardcoding the orientation IMO does not serve a
purpose at all.

Somebody made a really rediculous comment about users switching to QT
because of a simple deprecation, MAYBE they have grounds now with
a stripped 3.0 in sight, my opinion ? deprecate everything useless you can
before that deadline, because we may be stuck with the same crap for
another 3+ years.

Im sorry if I sound harsh, I dont even personally maintain GTK+ and I have
personally HAD IT with outside pressure on the GTK+ team to be so backward
forward and 4th dimentionally compatible - the deprecations have been
comming for a while - if users switch to QT it will at least give us a couple
years peace so that we can at least develop some features and do some
real refactoring.

Have a little solidarity, we write GTK+, it says it even in every source
file, it is provided AS IS, take it or leave it - people use GTK+ because
it rocks - it wont rock for long if we cant refactor, do the math.

I am on a half hour lunch break... have to run

Laters folks,
 -Tristan
___
gtk-devel-list mailing list
gtk-devel-list@gnome.org
http://mail.gnome.org/mailman/listinfo/gtk-devel-list


Re: GtkOrientable: Can widgets now be flipped?

2009-02-08 Thread Matthias Clasen
On Sun, Feb 8, 2009 at 7:07 AM, Murray Cumming murr...@murrayc.com wrote:


 So can they actully be flipped at runtime now? The documentation seems
 to suggest that they can't if they still use the old common base
 classes.

The intention is to eventually make GtkScale and friends
instantiatable.  When that happens you can use these classes to get
runtime flippable containers.

This has been put on hold until we can reach an agreement on default
values of properties for those newly-instantiatable classes.
Therefore, GTK+ 2.16 will keep all those classes abstract. If you want
runtime flippable containers now, you can always just trivially
subclass e.g. GtkScale.

Matthias
___
gtk-devel-list mailing list
gtk-devel-list@gnome.org
http://mail.gnome.org/mailman/listinfo/gtk-devel-list