EFI boot needs -fpic but all boot code was built with -fno-pic.
This is now set accordingly in HAIKU_BOOT_CCFLAGS and
HAIKU_BOOT_C++FLAGS.
Also setup compile flags for EFI platform.
* Only set HAIKU_BOOT_PLATFORM to bios_ia32 if not defined
* Add gnuefi build feature
* Introduce BOOT_LDFLAGS, and move options for passing to linker
into ArchitectureSetup
* x86_64 compile fixes for warnings in boot loader
* loader/elf.cpp: don't include ELF32 support when targeting EFI
* relocation_func.cpp: copy of the relocation code from gnuefi
to make _relocate extern "C", and avoid including <efilib.h>
* boot_loader_efi.ld: copy of gnuefi's elf_x86_64_efi.lds,
modified to include support for C++ constructors, etc. Keep in
sync with the gnuefi package
Signed-off-by: Jessica Hamilton <jessica.l.hamilton@gmail.com>
There was some mixup with the interrupt registers, still:
- The driver uses 16-bit read/write, but on SandyBridge the register is
32 bits
- There is a global interrupt enable bit, which must be set to unmask
everything else
- The bits for vblank interrupt are not the same on SNB and later PCH
based devices, and the code mixed the two.
Move the computation of the interrupt bits to an helper function, and
use it everywhere to make sure we always use the right bits.
hrev50506 extended the range of acceptable X and Y movement for PS/2
mouse. However, the value is sent on 1 byte (+ sign) so the check would
now always be valid.
Remove the check completely, we now rely only on bit 3 of the first byte
to be set (and not missing packets) to remain in sync with the mouse.
Thanks to Marcus Overhagen for reviewing the changes.
Modesetting
===========
My previous hack was setting the transcoder registers, instead of the
display ones. Do that the way it is designed in the driver instead:
- If there is a transcoder, set its registers, but do not set the
display timings. The display will remain set at its native (and only)
resolution, and panel fitting will adjust the output of the transcoder
to match.
- If there is no transcoder, set the display registers directly to the
native resolution, as it was done on previous generation devices.
- fPipeOffset hacks no longer needed
DPMS
====
It seems the panel control register is not readable on PCH? Anyway, the
code would loop forever waiting for the bit to become unset when turning
the display off. Waiting seems to not be needed, so just remove it as
well as the "unlock" bit, which does not work for me and results in a
black screen.
Remaining hacks
===============
I still need to force HEAD_MODE_A_ANALOG to get output on pipe B (LVDS
display) working. I suspect something is common to the two pipes or not
allocated to the right one.
This version will have less side effects on other generations and help
with getting things to work on SandyBridge and possibly later devices.
Please test and report.
- Missing "dontAsk" when renaming some files,
- Missing exceptions in the warning alert construction to automatically
allow modifying some files (anything inside the settings folder)
- Missing translations
Fixes#12925.
... sorry I overlooked this case in my last commit. Need to add apps using
the locale collator Compare() method instead of strcasecmp() as well.
Fixes#7712 better.
Fix using a global BCollator defined in BarApp.
Called static CompareByName() method in TTeamMenu instead of
duplicating code, made method public.
This comes from a patch by 0xffea for GSoC 2014.
A locale aware version of NaturalCompare in WindowMenuItem.cpp
is still needed.
Fixes#7712
Signed-off-by: John Scipione <jscipione@gmail.com>
According to korli "Valid values seem to be between 255 and -256" which
is verified by syslog running on virtual box with a PS/2 mouse on Haiku
hrev47470.
Fixes#7802
Signed-off-by: John Scipione <jscipione@gmail.com>
Programatically generating a string usually doesn't work well with
localization. Better unfold all the possiblities.
Moreover, the callers of the functions had a long if condition with
3 B_TRANSLATE calls embedded in, and there was an erroneous ; at the end
of those, so the supposedly protected statement was executed anyway.
Fixes#12747.
This file uses a mix of tab styles, which gcc6 will warn about because
it sees "misleading indentation". Fix the function where this happens.
Fixes#12759.
- This has been bothering me for quite a while. My screen is wide enough
to accomodate more, and even if they don't fit the BTabView, it's still
possible to navigate many more tabs with the keyboard shortcuts.
- We should still remember to fix BTabView to accomodate more tabs than
it can fit with some scrolling support or similar, however.
- Catching an exception in the normal flow is not a good thing, instead,
test before calling the function (I've been doing too much python
lately)
- Avoid using uninitialized nodeRef if there is no system root and also
no specified root in the message.
Thanks to Ingo or the code review.
We only compared the height, width, and depth of modes to decide if the
mode was already set. However, at boot, we may have these settings right
and other things not properly configured, preventing VESA to set a
working mode unless forced from the boot menu.
Fixes#12918
- Avoid use of fixed size buffer without checks, use BString.SetToFormat
instead.
- Use the same format for the initial setting of the values, so there
are less strings to translate, and locales using a different
representation for numbers get the right thing done as well.
- Fix initial label for the outline slider, which did not show the value
unlike others (problem spotted by Diver, thanks!)
* Truncate() would cut Unicode characters in the middle.
* This commit removes arbitrary message length limit as well as workaround
it was needed for and fixes#5289.
* Status view now expands to maximum possible width and cuts the status
message if necessary using TruncateString().
Signed-off-by: Adrien Destugues <pulkomandy@pulkomandy.tk>
* Minor whitespace fix.
The package kit needs to provide the package daemon with the node_ref of
the root directory, so the package daemon gives the correct results when
the request comes from inside a chroot.
Moreover, the package kit must be able to handle a root with a system
but no home package dir, as is the case inside the chroot environment
created by haikuporter.
Fixes#12602.
get_ref_for_path() doesn't resolve symlinks.
Construct a BEntry from the path passing 'true' for the traverse
argument. That BEntry is then be passed to BDirectory's constructor.
The error was hidden by the fact that the code didn't bother to error
check the result of get_ref_for_path(), which would have indicated that
the passed in path was a link.
Thanks Rene for your help.
if the entry fails to init, we are probably screwed, but, try to get
the path using get_ref_for_path in that case. Worst case scenario it
fails as well and the FilePanel points to your home directory.
In the case that the Keymap directory does not exist and has to
be created, reinitialize ref to the newly created directory.
Thanks Stippi for your help
There is still a bug present where if the Keymap directory is a
symlink to another directory, the File Panel does not follow that
link and instead opens to your home directory. More on this later.
...trying to get/make Keymaps settings dir if not able to create
the Keymaps directory under settings.
Resolve symlink, if it is a directory, use that directory instead,
otherwise use home directory as a fallback.
