Accelerant interface:
Introduce new hooks B_SET_BRIGHTNESS and B_GET_BRIGHTNESS. Brightness is
a float in the 0..1 range.
App_server:
Forward brightness things between BScreen and the accelerant.
intel_extreme:
Implement the hooks. Note that this only works for laptop panels, but
the driver will pretend to support it in other cases as well.
Screen preferences:
If the accelerant supports the B_GET_BRIGHTNESS hook, allow to set
brightness with a slider. Otherwise, the slidere is hidden and these
changes aren't visible.
The libvterm implementation was sending characters to app_server one by
one. This made it very slow to update the view and SerialConnect would
have noticeable flickering.
We now look through each line and try to find ranges of characters with
the same attributes. We can then draw each of these with a single
FillRect+DrawString. In the normal case (no special attributes or colors
anywhere), this means just one call per line, plus two for the cursor.
We just want to disable known broken ones, rather than opt-in for
working ones. This keeps the list of authentications shorter and is more
future proof.
Make TitleView a few pixels shorter at the default font size so that
it lines up with the scroll bar arrow to its immediate right.
Note the font size used in the title view before and still is 9px
(12px * 3/4)
Make this the minimum TitleView height even for smaller font sizes
so that the title view never is shorter than the height of a
scrollbar button. Make the minimum font size for TitleView 8px
because 6px (8px * 3/4) was just too small to read.
As font size is made larger it will increase the height of the
titlebar to accomodate the increased font size and consequently no
longer line up with the scrollbar arrow.
The code to scale the bar height via the font height was taken from
BColumnListView. The code to position the title in the middle of the
title bar was also taken from BColumnListView.
Some included style fixes:
* Rename height to fontHeight
* Rename loc to titleLocation
* 80 column limit
* https://github.com/raspberrypi/firmware/issues/643
* Mimics what the RaspberryPi guys adjusted it to in
the firmware.
* Solves corrupted UART serial on the RaspberryPi 2
* I think linux live-probes this frequency via dt-bindings.
Revision of last commit, everything I said about how we needed to
use 8px instead of 12px for the drag width and replicant tray
padding was correct, BUT, because we are subtracting from the
window width setting we need to account for the width of BOTH
draggers, so it becomes 2 * 8px or 16px that needs to be
subtracted -- sorry about that.
The minimum Deskbar width was calculated to fit 6 replicant icons,
however, this calculation doesn't take into account the extra width
taken up by the recently added second dragger. Consequently only 5
16x16 replicants currently fit.
However you can expand Deskbar to fit 6 (or more) icons now so it
is not a big deal.
Would it be desirable to adjust the minimum Deskbar width to fix 6
replicant icons again?
In this commit:
a44504a168
Deskbar: Refactor TRelicantTray::LocationForReplicant()
I made the following assumption that:
* if index == 0 return right away, no calculation required.
But this assumes that the first replicant will always fit on the
same line as the clock which doesn't work if you have a wide clock.
Update to always calculate the position even of the first replicant
in vertical (multi-row) mode.
In horizontal mode we still skip some calculations for the index 0
case.
Also I was using kClockMargin (12px) for the right margin, the
right margin should be only 8px: 5px for the dragger plus 3px more
for kTrayPadding. Finally we add in the width of the clock and
extra clock margin iif we're on first row and clock is not hidden.
Fixes (hopefully the last of) #8641
Damage in latter portions of the downloaded package data can cause the
meta-data for the caching to load OK, but the actual data is unable to
be parsed. In this case, the application can fail to refresh the
broken data. This change will mean that the data is "moved aside" into
another file so that on the next run, the application will load fresh
data in place.
This should have been done along with the time_t change, but I forgot
to check this then.
Technically this breaks ABI against BeOS, but:
1. BeOS used an int32, so we'd already slightly broken ABI here
2. Only one thing at HaikuArchives (VMwareAddons) and one recipe at HaikuPorts
(samba) uses this function at all.
If it turns out some critical BeOS app uses this, then I guess we can enclose
GCC2 guards around it, but since I can't find any evidence of that, I'm
pushing it without them for now.
The only thing of any real use in this directory was the Be Newsletter
article, and the objdump from BeOS R5 (at least I guess that's what it is...)
so I kept those.
As far as I can tell, these functions were "added" because they appear
in a symbols dump from R5. Nothing calls them, they don't do anything,
and they do not appear in R5 headers.
This reverts commit 2960780faa.
Sorry for the noise; I thought I'd run a compile but apparently it was
only a partial one.
It seems that the netboot core is linked into the main bootloader
no matter the build configuration, which means that since this one
function was called, the linker tried to resolve all the other symbols
from that static library. Which of course failed.
