Top was using the 'theorical' interval value to determine the percentages.
The measured thread times were taken in an always a bit larger interval than
that theorical value, hence the negative '%' occuring regularly.
Should fix#4589.
fBackgroundImage wasn't being cleared when removing the desktop
background image because the loading process wasn't ignoring
loading failures (as is the "no background for this workspace" setting).
fBackgroundImage not being NULL was causing crashes mainly in
ScreenChanged(), where it was trying to re-center a non-existant
background. Was causing #7376 and maybe others.
In List Mode, consider the extra margin added to the right of
the last column when computing the Extent of the PoseView. This
was causing the ResizeToFit to resize a bit too small.
Fixes#4318
* Avoid name collisions
* This uart stuff may work better as a class at
some point, however I didn't want to rock the
u-boot boat *too* much as I don't have the
hardware to test.
* Add nested function wrappers to allow usage of other
uart drivers depending on board. We may want to use this
on other platforms at some point (haha, maybe)
In previous, it was still working as I was more interested into != 0 than
exact count. Still, for clarity sake, I add -l to get the actual count.
Thanks Axel.
* Thanks go out to Simon Arlott for replacing
the first32k.bin blob with assembly removing
the need for first32k.bin hack.
* This assembly is a modified version removing
the Linux kernel boot args.
* haiku_loader renamed to kernel.img will boot
on Raspberry Pi directly.
* Called mvattr, move is not yet implemented which is why it's not part of
the image yet.
* Defaults to rename, moving attributes needs an extra option.
* Suggestions welcome :-)
* The ShapeListView also knows the PathContainer and StyleContainer
* When constructing the drag message, also include a complete archive
of the dragged shapes, bundled with archives for each included path
and the style.
* When handling the drop, and it came from another I-O-M window,
exract the Shape archive bundle from the drag message. For the
Style and the included Paths, try to find an existing Style and
existing Paths and reference those in the added Shape(s) instead
of adding duplicates.
The bash script tried to verify the existence of a file,
but there were several fils fitting the wildcard expression.
So, [ -e .../*gcc ]; was interpreted as if it had too many
arguments. Replaced that expression by a count of files.
Fixes#8493.
- Conditionally enable and disable event masks as needed when we detect
the mouse entering/exiting the deskbar area. This is necessary because
otherwise the app_server appears to eat mouse messages due to the ctrl+cmd+
resize shortcut. Also change some conditionals slightly to better deal with
differing combinations of autohide and/or autoraise.
- Rework Deskbar's handling for the autoraise/autohide cases to not
need event masks. This prevents various problems that would occur
if one happened to be working in another window that overlapped part
of Deskbar, as reported in #8497. Instead, we now use a message filter
to redirect the messages as needed.
* Put archived versions of the selected paths into the drag message.
* If the base class version HandleDropMessage() failed, it means the
drag message came from another window. Reconstruct the paths and
add them via the AddPathsCommand.
* Add archived versions of the selected Styles to the drag message.
* If the base class didn't handle the drop message, try to unarchive
Styles and add them via AddStylesCommand.
* Make Kernel ARM UART slightly more generic
through (BOARD_UART_CLOCK) configured per board
* Add initial Raspberry Pi serial code
* Still rough and non-working
Passing NULL or a blank string to SetToolTip(const char*) sets the
tooltip to NULL by calling SetToolTip(BToolTip*) with a NULL argument.
Calling SetToolTip(BToolTip*) with a NULL argument calls HideToolTip()
because sometimes the tool tip can change without the mouse moving,
for example because the user clicked.
Thanks Axeld and Stippi.
Replace the leaf bitmap with a non-cutoff version derived from the
'data/artwork/Haiku Leaf' file. So, when icon size is > 16 the full
deskbar leaf is shown without the bottom cut-off. It should look
pretty much identical at 16x16 and in vertical mode at any icon size.
Also add a deskbar_leaf.png file to the data/artwork directory
although I didn't use that file in Deskbar, instead I embed an
archive bitmap into icons.h. I got rid of icon-freelogo.rdef and
icons_logo.h.
The bug is that in horizontal mode the tooltip will remain set to
the last moused over team menu item even if the mouse is no longer
over a menu item. The bug can be seen in the following screenshot:
http://26.media.tumblr.com/tumblr_m3gze8s1xi1r0f0hfo1_400.png
To fix this bug, allow you to set the tooltip text to blank or NULL
in SetToolTip(const char* text). In ShowToolTip() check to see if
the tooltip text is blank or NULL and if so, don't show the tip.
Setting the tooltip to blank or NULL effectively unsets the tooltip
on a view.
This fixes a bug introduced in hrev44075 (I think) where the internal
tooltip window wouldn't resize itself after setting new tooltip text.
So if you set the tooltip to a short string then set it to a long
string the tooltip would get cut-off or if you set it to a long string
and then set it to a short string the tooltip window would be too wide.
The result can be seen in Deskbar in the following screenshot:
http://30.media.tumblr.com/tumblr_m3fg327NsV1r0f0hfo1_400.png
After invalidating the layout the internal tooltip window gets resized
correctly.
* Change ShowTip() point parameter name to where.
* Add a parameterless ResetWindowFrame() overload that get's the current
where and calls ResetWindowFrame(BPoint where) which does the actual
work. FrameResized() calls this parameterless ResetWindowFrame()
method instead of doing the work in that method. This is functionaly
the same but allows me to call the parameterless ResetWindowFrame()
elsewhere.
Only update the tooltip if the mouse goes over a new menu item making
sure to exit if the item is NULL, there is a visible label, or it is the
same item. If the mouse moves outside the view reset the last item to
NULL.
* Rename DrawLabel() and SetDrawLabel() to HasLabel() and SetHasLabel()
to make it clear what these methods are setting a getting and setting
a bool and not doing the actual work for drawing the label.
