[Denemo-devel] Re: Denemo-devel: Key bindings

[Jean-René Reinhard]

Le Tue, May 06, 2008 at 09:08:20AM +0100, Richard Shann écrivait:
> Hi,
> Denemo has moved to Savannah - and, worse still we have no control over
> the sourceforge site, which has become an orphan. So please use
> denemo <address@hidden>
> for emails.

I had a hard time to find a pointer on a denemo devel-list. Would it be possible
to better advertise this change on the website? Googling for denemo didn't get
me on Savannah.

> The keybindings you have written a patch for are being phased out - the
> file standard.accels holds keyboard accelerators which are the ones that
> are supposed to be used. They are defined by right-clicking on menu
> items (or checking edit->keyboard setup->allow quick shortcut edits and
> pressing a key when on a menu item).

If one lets the application handle the keymaprc, things will probably be
allright. However, isn't editing the keymaprc file by hand another valid option 
worth supporting?

If not a disclaimer should be added to keymaprc to direct the user to the proper
way to edit this file.

>  I think something is wrong with the CVS/distribution - perhaps that the
> standard.accels file that is distributed does not have the main
> keypresses defined, so that the old keybindings are kicking in. I'll
> look into this.

Something is probably wrong since I checked out a CVS repository on sourceforge,
following guidelines on the sourceforge denemo web page.
There is no standard.accels file there. Where is the reference repository
located? Savannah?

Jean-René Reinhard

> Richard
> 
> On Tue, 2008-05-06 at 00:25 -0700,
> address@hidden wrote:
> > ------------------------------
> > 
> > Message: 6
> > Date: Fri, 2 May 2008 21:57:04 +0200
> > From: Jean-Ren? Reinhard <address@hidden>
> > Subject: [Denemo-devel] Key bindings
> > To: address@hidden
> > Message-ID: <address@hidden>
> > Content-Type: text/plain; charset="iso-8859-1"
> > 
> > Hi,
> > 
> > I'm rather new to denemo. I had to set it up for my girlfriend. I have come
> > across several usability issues, which I would like to report.
> > 
> > I checked out the cvs today, but I just saw plans to move to git on the 
> > archives
> > of the devel list. Is the cvs still the reference? Is the git repository 
> > already
> > set up? If it is the case, where can it be fetched?
> > 
> > The issues I would like to tackle today are keybindings issues. As almost 
> > every
> > interaction with the score are made through keybindings, it's rather 
> > important.
> > When I first tried denemo, the default keymap was unusable, because notes 
> > could
> > not be entered. I managed to change the keymap to use denemo, but since the
> > keymap configuration is not documented, it's rather awkward.
> > 
> > After some exploration of the source code, I think I solved the issue. 
> > Denemo
> > stores the keybindings in a two level hashtable. When given a GdkKeyEvent, 
> > it
> > checks for a keybinding, using as the first key the modifiers of the event, 
> > and as second key the keyval of the event. The issue stems from that Gdk 
> > interprets Shift to modify the keyval, and denemo also uses Shift as a 
> > Modifier. For example, on my azerty keyboard, I need to press Shift+& to 
> > get a keyval for 1. However when pressing Shift+&, the GdkEvent contains 
> > (mod : Shift, keyval : 1). The keyval is ok, but not the modifier. If I 
> > want to use this binding, I have to use :
> > <bind>Shift+1</bind>, which is weird. One could document this behaviour, 
> > but I
> > think it should be avoided, since one should change the denemo keymap for 
> > every
> > system keymap. Another solution would be to stop using Shift as a modifier
> > inside denemo. This solves the issue. I attach a patch that does just that.
> > 
> > Another issue considering the keymap configuration is that the action names
> > depends on the locale of the system on which denemo is build. When I 
> > compiled denemo earlier today, this was rather misleading. If I give a 
> > binding to InsertQuarterNote, the binding will just be ignored. Once again, 
> > this is not documented. Furthermore, no keymap file containing the actions 
> > name can be found in the cvs tree. One has to  build it from the 
> > appropriate po file. I can provide a defaut keymap file for my langage and 
> > document this really weird behaviour. But I really wonder what is the 
> > rationale behind this behaviour, since the useful action names depends on 
> > the locale of the system where denemo is build, not of the locales of the 
> > system where it is used.
> > 
> > Hopes this helps,
> > 
> > Cheers,
> > Jean-Ren?
> > 
> > 
> > -------------- next part --------------
> > Index: kbd-custom.h
> > ===================================================================
> > RCS file: /cvsroot/denemo/denemo/src/kbd-custom.h,v
> > retrieving revision 1.6.2.3
> > diff -u -r1.6.2.3 kbd-custom.h
> > --- kbd-custom.h    14 May 2006 19:04:01 -0000      1.6.2.3
> > +++ kbd-custom.h    2 May 2008 18:55:11 -0000
> > @@ -11,9 +11,11 @@
> >  #include <limits.h>
> >  #include <gdk/gdk.h>
> >  #include <denemo/denemo.h>
> > -
> > +/*
> >  #define MASK_FILTER(state) state & (GDK_SHIFT_MASK | GDK_CONTROL_MASK \
> >                                 | GDK_MOD1_MASK)
> > +*/
> > +#define MASK_FILTER(state) state & (GDK_CONTROL_MASK | GDK_MOD1_MASK)
> >  
> >  struct name_and_function
> >  {
> > Index: keyboard.c
> > ===================================================================
> > RCS file: /cvsroot/denemo/denemo/src/Attic/keyboard.c,v
> > retrieving revision 1.1.2.12
> > diff -u -r1.1.2.12 keyboard.c
> > --- keyboard.c      5 Jan 2007 20:49:21 -0000       1.1.2.12
> > +++ keyboard.c      2 May 2008 18:55:12 -0000
> > @@ -30,17 +32,21 @@
> >    if (0 == strcmp (key, "Ctrl"))
> >      ret = 4;
> >    else if (0 == strcmp (key, "Shift"))
> > -    ret = 1;
> > +  //  ret = 1;
> > +    ret = 0;
> >    else if (0 == strcmp (key, "Alt"))
> >      ret = 8;
> >    else if (0 == strcmp (key, "Ctrl+Shift"))
> > -    ret = 5;
> > +  //  ret = 5;
> > +    ret = 4;
> >    else if (0 == strcmp (key, "Alt+Shift"))
> > -    ret = 9;
> > +  //  ret = 9;
> > +    ret = 8;
> >    else if (0 == strcmp (key, "Alt+Ctrl"))
> >      ret = 12;
> >    else if (0 == strcmp (key, "Alt+Ctrl+Shift"))
> > -    ret = 13;
> > +  //  ret = 13;
> > +    ret = 12;
> >  
> >    return ret;
> >  }
> > 
> > ------------------------------
> > 
> > Message: 7
> > Date: Sat, 3 May 2008 09:31:05 +0200
> > From: Jean-Ren? Reinhard <address@hidden>
> > Subject: Re: [Denemo-devel] Key bindings
> > To: address@hidden
> > Message-ID: <address@hidden>
> > Content-Type: text/plain; charset=iso-8859-1
> > 
> > > Another issue considering the keymap configuration is that the action 
> > > names
> > > depends on the locale of the system on which denemo is build. 
> > 
> > Apparently, I was wrong here. After setting up LC_MESSAGES appropriately,
> > everything went fine. I misunderstood gettext.
> > 
> > The keymaprc file still has to be changed according to locale, which should 
> > be
> > documented, and defaut keymaprc for the supported langages should be 
> > provided.
> > 
> > Cheers,
> > Jean-Ren?
> > 
> > 
> > 
> > 
> > ------------------------------
> > 
> > Message: 8
> > Date: Tue, 6 May 2008 09:28:48 +0200
> > From: Jean-Ren? Reinhard <address@hidden>
> > Subject: Re: [Denemo-devel] Key bindings
> > To: Jeremiah Benham <address@hidden>
> > Cc: address@hidden
> > Message-ID: <address@hidden>
> > Content-Type: text/plain; charset=iso-8859-1
> > 
> > That's a pity. Git is so much powerful and flexible than CVS...
> > 
> > Le Fri, May 02, 2008 at 04:36:47PM -0500, Jeremiah Benham ?crivait:
> > > don't use git yet. It is not up to speed with cvs. I was experimenting
> > > with it and it seems not everyone wants to switch. I may just get
> > > deleted soon.
> > > 
> > > Jeremiah
> > > 
> > > On Fri, 2008-05-02 at 21:57 +0200, Jean-Ren? Reinhard wrote:
> > > > Hi,
> > > > 
> > > > I'm rather new to denemo. I had to set it up for my girlfriend. I have 
> > > > come
> > > > across several usability issues, which I would like to report.
> > > > 
> > > > I checked out the cvs today, but I just saw plans to move to git on the 
> > > > archives
> > > > of the devel list. Is the cvs still the reference? Is the git 
> > > > repository already
> > > > set up? If it is the case, where can it be fetched?
> > > > 
> > > > The issues I would like to tackle today are keybindings issues. As 
> > > > almost every
> > > > interaction with the score are made through keybindings, it's rather 
> > > > important.
> > > > When I first tried denemo, the default keymap was unusable, because 
> > > > notes could
> > > > not be entered. I managed to change the keymap to use denemo, but since 
> > > > the
> > > > keymap configuration is not documented, it's rather awkward.
> > > > 
> > > > After some exploration of the source code, I think I solved the issue. 
> > > > Denemo
> > > > stores the keybindings in a two level hashtable. When given a 
> > > > GdkKeyEvent, it
> > > > checks for a keybinding, using as the first key the modifiers of the 
> > > > event, and as second key the keyval of the event. The issue stems from 
> > > > that Gdk interprets Shift to modify the keyval, and denemo also uses 
> > > > Shift as a Modifier. For example, on my azerty keyboard, I need to 
> > > > press Shift+& to get a keyval for 1. However when pressing Shift+&, the 
> > > > GdkEvent contains (mod : Shift, keyval : 1). The keyval is ok, but not 
> > > > the modifier. If I want to use this binding, I have to use :
> > > > <bind>Shift+1</bind>, which is weird. One could document this 
> > > > behaviour, but I
> > > > think it should be avoided, since one should change the denemo keymap 
> > > > for every
> > > > system keymap. Another solution would be to stop using Shift as a 
> > > > modifier
> > > > inside denemo. This solves the issue. I attach a patch that does just 
> > > > that.
> > > > 
> > > > Another issue considering the keymap configuration is that the action 
> > > > names
> > > > depends on the locale of the system on which denemo is build. When I 
> > > > compiled denemo earlier today, this was rather misleading. If I give a 
> > > > binding to InsertQuarterNote, the binding will just be ignored. Once 
> > > > again, this is not documented. Furthermore, no keymap file containing 
> > > > the actions name can be found in the cvs tree. One has to  build it 
> > > > from the appropriate po file. I can provide a defaut keymap file for my 
> > > > langage and document this really weird behaviour. But I really wonder 
> > > > what is the rationale behind this behaviour, since the useful action 
> > > > names depends on the locale of the system where denemo is build, not of 
> > > > the locales of the system where it is used.
> > > > 
> > > > Hopes this helps,
> > > > 
> > > > Cheers,
> > > > Jean-Ren?
> > > > 
> > > > 
> > > > -------------------------------------------------------------------------
> > > > This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
> > > > Don't miss this year's exciting event. There's still time to save $100. 
> > > > Use priority code J8TL2D2. 
> > > > http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
> > > > _______________________________________________ Denemo-devel mailing 
> > > > list address@hidden 
> > > > https://lists.sourceforge.net/lists/listinfo/denemo-devel
> > > 
> > 
> > 
> > 
> > ------------------------------
> > 
> > -------------------------------------------------------------------------
> > This SF.net email is sponsored by the 2008 JavaOne(SM) Conference 
> > Don't miss this year's exciting event. There's still time to save $100. 
> > Use priority code J8TL2D2. 
> > http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
> > 
> > ------------------------------
> > 
> > _______________________________________________
> > Denemo-devel mailing list
> > address@hidden
> > https://lists.sourceforge.net/lists/listinfo/denemo-devel
> > 
> > 
> > End of Denemo-devel Digest, Vol 14, Issue 1
> > *******************************************
>