Commit by ggibecka :: r595 /widget/trunk/ggiwidget/ (dropdownlist.c frame.c helper.c): (link)

Allow for the toplevel widget in the droplist to react to events.
remove debug output
minor coding style cleanups

Commit by ggibecka :: r594 /widget/trunk/ggiwidget/EXPSYMS: (link)

export ggiWidgetDrawFrame for use in user generated widgets

Commit by pekberg :: ggi-core/libggi/display/vnc/rfb.c:

Recognize the Call Control pseudo-encoding (from OLIVE).

Commit by ggibecka :: r593 /widget/trunk/ (3 files in 2 dirs): (link)

Add a name and a unique ID property to the widgets to allow for finding
widgets without having to keep pointer references.
Add generic Control call functions to do this.

Commit by ggibecka :: r592 /widget/trunk/ (3 files in 2 dirs): (link)

Replace the given widget, if the (up to now forbidden) place parameter
GWT_BY_WIDGETPTR is used in a link call.
This is useful for swapping out entries that have changed without first
having to find out where exactly they are attached which can be tricky,
e.g. for grids.
Note that this unlinks the widget and thus the given widget needs to be
destroyed properly unless it should be reused.

Commit by ggibecka :: r591 /widget/trunk/ggiwidget/EXPSYMS: (link)

Add SingleChild helper functions to exports as those help to create
simple container widgets (like own buttons or decoration frames)

Commit by pekberg :: ggi-core/libgg/gg/dl_win32.c:

 Don't fail completely if the error message requires inserted values, instead
show the error message without inserting any value (i.e. retain %1 etc).
There is no way of knowing what to insert short of making a list of all
the error messages and what inserts they take, for all versions of
Windows (who knows if the error messages differ?), so this is the best
that can be done, at least within reason.

Commit by pekberg :: ggi-core/libgii/gii/gii.c:

Fix bug introduced in r1.81 causing events to not be strictly monotonic.

Commit by pekberg :: ggi-core/libgii/input/directx/ (dxinput.h input.c):

 Add support for entering any unicode character by pressing the Alt key, then
typing + on the numeric keypad followed by the unicode character in hex.
The unicode character is sent after the Alt key is released.

Commit by pekberg :: ggi-core/libggi/default/linear_32/crossblit.c:

Remove some dead (i.e. impossible to reach) code.

Commit by pekberg :: ggi-core/libggi/display/vnc/rfb.c:

 Oops, the provided encoding descriptions didn't all fit in the given space.
Store the descriptions as pointers instead and constify while at it.

Commit by pekberg :: ggi-core/libggi/programs/util/inputdump.c:

 Don't look outside array bounds in the extra check to make sure an input
device was indeed found.

Commit by pekberg :: ggi-core/libgg/configure.in:

Fix typo in previous, reported by Olaf Buddenhagen.

Commit by pekberg :: ggi-core/libgg/configure.in:

 Mingwex version 3.17 fixes a killer bug in it's snprintf implementation. Stay
away from versions prior to that one (and fall back to _snprintf in msvcrt).

Commit by pekberg :: ggi-core/libgg/gg/lib/snprintf.c:

Print pointers as unsigned.

Commit by pekberg :: ggi-core/libgg/regress/gg/ (snprintf.c Makefile.am):

 Mingwex have had a snprintf implementation that doesn't match the sscanf
implementation in terms of %p, which we rely on when configuring with the
--disable-conffile option. Add a regression test, to make it easy to check if
it works as expected.
Our own replacement snprintf may also be incompatible with a sscanf
provided by the system, but that's another story...

Commit by cegger :: xggi/README.quickguide:

XGGI requires GGI 2.2.x. 'or newer' is wrong as XGGI hasn't been ported to GGI -current yet.

Commit by pekberg :: ggi-core/libgii/configure.in:

 The SIZEOF_* defines are correct for cross builds these days and system.h
is generated in libgg.

Commit by pekberg :: ggi-core/libgg/configure.in:

The SIZEOF_* defines are correct for cross builds these days...

Commit by pekberg :: wrappers/svgalib/m4/common.m4:

Newer autotools has less newlines causing syntax errors.