Really, you really need to:
1. Let the drag source window act as a drag target.
2. Let the drag shapped window re-enter the source window.
You have to fight with the shaped window to drag it
inside the origin window.
Miguel.
1. My icons suck.
2. Keyboard handling is not finished.
3. Save current-mode, load current-mode is not
implemented yet.
4. Listing mode is not working, as I have hardcoded
the icon view for now.
5. Will finish this tomorrow. Not much is missing.
6. No, this is not the final implementation.
7. Yes, later I will put some scheme for using
more icons as discussed in the ethernal "Re: Icons"
thread.
Miguel.
1998-05-07 Miguel de Icaza <miguel@nuclecu.unam.mx>
* gscreen.c (panel_icon_list_realized): Add drag and drop support
to the icon list view;
1998-05-06 Miguel de Icaza <miguel@nuclecu.unam.mx>
* gscreen.c (panel_fill_panel_icons): New function: Fill up the
icon list.
(x_create_panel): Change in the approach: we create both of the
widget views at creation time and we now carefully show every
component. I attach everything to the table, but not necesarly
show it.
- Fix for the gmc crashes when no window is open and you
start a program.
- Bad attempts at fixing the sigwinch propagation.
- Internationalization fixes from Alex (ru.po and
bunch of updates to the text mode edition to deal
wiht this).
Miguel.
1998-05-04 Miguel de Icaza <miguel@nuclecu.unam.mx>
* gwidget.c (x_update_input): Keep the cursor always visible.
* gview.c (view): Use a buttonbar widget.
* gwidget.c (buttonbar_clicked): Provide a trivial implementation
of the buttonbar that only responds to key sequences.
* gmain.c (dialog_panel_callback): Call midnight_callback to get
the main bindings working.
* gdesktop.c (perform_drop_manually): Use p to walk the dropped
data. This is an important bug fix.
* gmain.c (idle_destroy_window), main.c: Session management now
supports not showing a toplevel window (in case all of the windows
were closed).
1998-05-04 Miguel de Icaza <miguel@nuclecu.unam.mx>
* view.c (view_quit_cmd): Use dlg_stop.
* main.c (ctl_x_cmd): Implement ctl-x handling as a state of the
key press events. Drop the usage of mi_getch to acomplish this
task.
(midnight_callback): Deal with the current map depending on the
c-x state.
* background.c (message_1s1d): the unprocessed string with %d was
passed to message_1s; fixed small memory leak
* widget.c (button_callback): WIDGET_CURSOR: make cursor position
dependend from button type (fixes the off by one bug in advanced
changeown).
* Makefile.in: Links to mad.c and mad.h were created wrong when
srcdir != builddir and directory depth to srcdir and builddir were
different (e.g. $srcdir=/src/mc and $builddir=/src/compile/mc)
O_EXCL in retrieve_file_start2 then I will find a different fix.
Mon May 4 10:27:49 1998 Norbert Warmuth <k3190@fh-sw.de>
* ftpfs.c (retrieve_file_start2): Don't create target file O_EXCL,
in copy_file_file we check existance of the target file and know
that we want to truncate it.
Mon May 4 10:21:31 1998 Norbert Warmuth <k3190@fh-sw.de>
* file.c (copy_file_file): Schedule deletion of short target file
only when we created or truncated the target file and not already
when we decided to overwrite an existing file.
panel, the old way of using it was an historic relic.
- This now works properly with the GNOME edition.
- Finished/fixed the implementation of the desktop-remove-icon
functionality.
- Raster's new multi-ok window. For some reason this is not getting
loaded. I am compiling gdb 4.17 now, as it is better for egcs
output.
Miguel.
Now in wrap mode the End key and cursor up key behave much better
(e.g. when viewing binary files with lots of wrapped lines).
It's not perfekt but it's better than it used to be.
Tue Apr 28 06:52:24 1998 Norbert Warmuth <k3190@fh-sw.de>
* gnome/gcmd.c (gnome_open_terminal): Changed my_system(1,...) to
my_system(EXECUTE_AS_SHELL,...)
Tue Apr 28 06:06:03 1998 Norbert Warmuth <k3190@fh-sw.de>
* vfs/extfs.c (extfs_open, extfs_close): Changed my_system(1,...)
to my_system(EXECUTE_AS_SHELL,...), this fixes the broken copyin and
copyout in 4.1.32.
Tue Apr 28 06:11:08 1998 Norbert Warmuth <k3190@fh-sw.de>
* view.c (toggle_wrap_mode, toggle_hex_mode): Force recalculation
of bottom_first (we mustn't use an already calculated and cached
value because it is invalid for the new mode and the End key would
not move to the end of the file).
* configure.in: Renamed the option `--with-our-slang' to
`--with-included-slang' (this one looks better because we also
have an `--with-included-gettext').
Make the option `--with-ext2undel' recognice a given path.
* cmd.c (view_file_at_line): In plain view (F13) set the default
magic flag to zero in order to view the file content unprocessed
(esp. don't uncompress files if they are compressed). The
view_simple_cmd got broken when the default magic flag in view.c
was changed from 0 to 1.
* view.c (do_view_init, goto_line): Set wrap mode temporary off
to make goto line number work, i.e. `line number' now always means
line number in file and not line number on screen (in wrap mode
one long line wrapped once is displayed in two lines on the screen).
That's important when the viewer is invoked from the find file
dialog to display even in wrap mode approxiamtly the part of the
file where we found the content we searched for.
(move_forward2): In wrap mode lines were sometimes counted wrong
causing cursor up to move more than one line.
(move_backward2): Fixed the movement in wrap mode.
(change_viewer): Always re-init viewer when we have a filename,
i. e. if the viewer is invoked with simple_view_cmd then we can switch
with the F8 key between unprocessed file content und uncompressed
file content.
(view_init): re-init view also when magic flag was altered
fix a bug in the panel-widget selection.
i18n changes from Andrej Borsenkow.
Menus renamed, better texts -- thanks Havoc.
lav's fixes for curses compilation.
Miguel
* slint.c: Applied the patch from Bill Nottingham <wen1@cec.wustl.edu>
to make it link against SLang >=1.0.
I reviewed the diffs to the part of SLang we use and there are no
further changes to MC necessary (one function's return value
changed it's meaning but we don't use this return value).
* configure.in: Undone the change which prevented linkage against
SLang >=1.0
Pavel Machek