* This also fixes the shutdown process, as the registrar no longer
kills the app_server.
* Removed SERVER_PORT_NAME definition as it has no use anymore.
* When creating the port of the registrar's authentication manager, we
now set it manually, so that the user/group functions work.
* This allows LaunchDaemon::_StartSession() to set up the user, and
groups as needed.
* Instead, the caller should have done this already. This is really
outside of the scope of the launch_daemon.
* Fixed Login with empty passwords; removed the (unused) test login
feature along the way.
* You can now add arbitrarily named fields to the message as well,
without having them all specified in the template.
* Also added a missing converter method that is called when there
are no values to add, and implemented all methods in the base
class, so that you only have to implement the methods you actually
need.
* BRoster now allows settings a "no-registrar" mode that is currently
only honored in _LaunchApp(), though.
* Job::Launch() is now using this, which also allows launching
applications by signature (ie. if the job name matches the
signature, you can omit the "launch" option).
* You can now put jobs/services into a target.
* Instead of having Login started as part of the normal boot process,
it's now in the "login" target.
* The app_server now launches the login target when a login becomes
available (ie. during startup, but that could be improved later on).
* Instead of launching Tracker/Deskbar directly, we now launch the
Login application.
* This will now start a new session for the selected user (the password
is currently ignored).
* When a user session is started, the launch_daemon forks, and the
child then restarts the LaunchDaemon application in user mode.
* It then registers itself with its parent, in order to resolve user
dependent services.
* Added a user launch file that will cause Tracker, and Deskbar to
start in the new session.
* get_roster_port_name() is no longer needed.
* This also removes the app_server restart code from the debug
server -- this will be done by the launch_daemon in the future.
* Instead of letting the kernel search for the syslog port, the
daemon now registers itself with the kernel (which even solves
a TODO).
* A port is created for the actual log messages from the launch_daemon,
and used on start.
* However, the SyslogTest does not yet work, due to the BMessage <->
KMessage communication problems.
* These methods don't really work yet, as BMessage doesn't support
replying with a KMessage; the request is received, but the reply
never gets to the target.
* Dropped "create_port" -- this is now the default for services.
* Additionally (or alternatively, if you use the "legacy" mode), you can
now create named ports, and specify their capacity.
* Added convenience methods to BLaunchRoster that automatically use the
signature of the current be_app.
* BRoster::Launch() cannot be used (yet), as it pre-registers the
application we're launching, and that won't work for the registrar or
anything else until the registrar is up and running.
* Renamed B_GET_LAUNCH_CONNECTIONS to B_GET_LAUNCH_DATA.
* Add the team ID to the get-launch-data reply.
* Added BLaunchRoster::GetPort() for convenience.
* Removed some superfluous debug output, but temporarily dump all stdio
to /dev/dprintf (ie. the syslog).
* Made job matching case insensitive (as MIME types should be).
* This enables a mechanism to profile almost the complete boot process
(starting with main2()), if SYSTEM_PROFILER is defined to 1.
* You can access the profiling data using "profile -r".
* This will be heavily inspired by Apple's launchd, as well as
systemd -- for now it really doesn't do a whole lot, though.
* What works so far: the configuration files are read, parsed, and
the jobs created.
* The jobs are even initialized, and their message ports created.
* BApplication now retrieves a previously created port from the
launch_daemon for use with BServer.
* Only the registrar actually uses this for now.
BColumnListView:
- Add helper method for getting the visible rect of a given field.
Refactor SuggestTextPosition to use it.
{Tree,Table}:
- Add wrapper to retrieve table cell rect using the aforementioned
BCLV helper.
* Address TODO about setting fSelected when nothing is done.
* Pass a pointer to the tab view to the BTab so that it can call Invalidate().
(Checked against BeOS).
* Call Invalidate() from the BTab after SetView() & SetName().
Fixes#12108 & #12196.
The media_server is now able to remember the timesource associated to
a certain registered_node and always remove it when the owner
application crash, Fixes Ticket #11852
* Drop lib/edit and matching bsd header
* Convert Debugger to libedit build package
* Should solve problems with libedit consumers
not defining _BSD_SOURCE
* Progress on #10267
netresolv (and libbind) won't cache DNS requests, which can result in a
lot of DNS requests being made for the same host. Implement a simple
cache in RAM (local to each application) which will keep the most
recently requested addresses cached. This can speed up loading of an
HTTP page a lot, by saving a DNS request for each resource stored on the
same server as the main page.
libbind development was transferred to the NetBSD project at
http://wiki.netbsd.org/individual-software-releases/netresolv/
There isn't an official release yet, but they provide a set of patches
against the latest libbind release.
* Remove all files we don't use
* Merge the changes to the remaining files
* Add some new files we need
* Move getifaddrs implementation to libnetwork (instead of libbnetapi)
so it can be used by netresolv.
Fixes#8293 : netresolv uses getifaddrs to determine if there is a local
IPv6 address. If there is not, it will not return AAAA records.
* Was leaking fQueuedJobs on destruction.
* fHaveRunnableJobSem implementation was not completed; it was never
released.
* Added Pop() variant that is a bit more flexible, and allows for a
timeout as well as waiting even when the queue is empty, and can
return a status code.
* Fixes sharing semantics, so non-shared semaphores in non-shared
memory do not become shared after a fork.
* Adds two new system calls: _user_mutex_sem_acquire/release(),
which reuse the user_mutex address-hashed wait mechanism.
* Named semaphores continue to use traditional sem_id semaphores.
* Put it in the BSupportKit namespace, following the style introduced
with the package kit for now.
* The BSupportKit::BJob class no longer knows about the package kit's
Context class. However, the BPackageKit::BJob class does.
* Due to the namespace juggling, a lot of files had to be touched.
* The JobQueue class remains private.
* Due to the way Haiku is built on itself, you cannot build this change
under Haiku with an older release.
* When you receive a message from a KMessage, and reply to it,
it will automatically reply as KMessage, too.
* This allows to communicate with BLoopers from within the kernel
or libroot.so.
"Renaming" means the icu namespace is suffixed with the version number,
atm icu_55. Using "renaming" allows to use two different versions of ICU,
thus easing upgrades. For instance haikuwebkit uses a current version of ICU,
while the system uses a newer one after an upgrade.
* Replace all uses of the icu namespace in our public headers, with a macro
defaulting to icu. As the namespace is only used for private fields pointers,
there should be no impact.
* Locale kit *.cpp have to import the macro from <unicode/uversion.h> *before*
including any locale headers. Ditto for a Time preferences cpp file. This way,
the correct current icu namespace is referenced.
* Fixes bug #12057.
Those return uintNN_t types instead of our own types,
but uint32 for example is long while uint32_t isn't,
giving some trouble with the PRI* macros for example on PPC.
It seems like glibc also has paths.h and m4 fails to bootstrap
without _PATH_BSHELL.
This file really needs some cleanup btw, since most is actually
irrelevant or incorrect for Haiku.
The AddOnManager was in the global namespace, clashing with application
classes with the same name.
The input_server has an AddOnManager of its own. When the
shortcut_catcher filter was loaded by said AddOnManager, it in turn
loaded libgame.so, which in turn loaded libmedia.so, where an
AddOnManager was created for the global AddOnManager instance in
libmedia.so. Unfortunately the wrong AddOnManager, the one from the
input_server, was created. This lead to two AddOnManagers being active
in the input_server which very well could be responsible for #11049
and #11280.
This was a regression since the move of the AddOnManager from the
media_server to libmedia.so in hrev47086. This also fits with the two
tickets.
I actually noticed the problem when debugging the shutdown process of
the input_server, where the destruction of the wrong AddOnManager
caused a deadlock with itself.
* Do not define the symbols by default, as they are not in the default
libraries.
* Adjust jamfiles of all code using BSD extensions to define
_BSD_SOURCE.
* This makes Haiku slightly more compliant to standard C/POSIX.
It works analoguous to BView::RemoveSelf(), i.e. it removes itself from
the parent (layout in this case) and returns whether or not it had and
was successfully removed from said parent.
The BNetworkRoute class manages a route_entry and the sockaddr's
associated with it. It replaces the direct use of route_entry in the
BNetworkInterface API.
Using route_entry is fragile and inconvenient as it only holds pointers
to the sockaddr's. When getting a list of routes from the kernel, each
route_entry is set up so that its pointers point into the single flat
buffer that is passed around. Creating a copy of the route_entry and
then deleting the flat buffer makes the pointers in the copy stale.
Returning these route entries therefore always lead to a use-after-free
when they were eventually used.
BNetworkRoute also takes over the code and functionallity of getting
routes from RouteSupport. The corresponding method in BNetworkRoster is
replaced by a static method in BNetworkRoute.
Also distinguish between the default route and gateway of an interface.
GetDefaultRoute() now gets the default BNetworkRoute for the interface
while GetDefaultGateway() gets the associated gateway address within
that default route. Adjust network preferences panel to this change.
Note that we currently only seem to have per interface default routes
and not an actual global default route. This was already the case before
these changes and I did not further investigate what this means.
It can be used to get a stack trace of the current thread. Note that
this works by walking frame pointers and will not produce anything
useful if an application is compiled with the frame pointers omitted.
The stack base and end addresses have to be provided as arguments and
are used to check that the frame pointers fall within that range. These
values are thread specific and can be retrieved with get_thread_info().
No other sanity checks (like checking for loops in the linked list) are
done.
This is a simplified rewrite of the stack trace code from the kernel
debugger.
As this code is common to x86 and x86_64 but is not generic across
architectures I introduced x86_common as a directory to put such
sources.
Extend the get_nearest_symbol_at_address() private runtime_loader
export to include imageName and exactMatch arguments.
The imageName holds the SONAME of the image, if available, so cannot
neccessarily be extracted from the image path.
Whether or not there was an exact match, i.e. the symbol with its size
contains the address, is now returned in exactMatch.
When enabled (using heap_debug_dump_allocations_on_exit(true) or
MALLOC_DEBUG=e) this causes a dump of all remaining allocations when
libroot_debug is unloaded. It uses terminate_after to be called as
late as possible.
When combined with alloc stack traces this makes for a nice if a bit
crude leak checker. Note that a lot of allocations usually remain
even at that stage due to statically, lazyly and globally allocated
stuff from the various system libraries where it isn't necessarily
worth the overhead to free them when the program terminates anyway.
When configured to do so (using heap_debug_set_stack_trace_depth(depth)
or MALLOC_DEBUG=s<depth>) the guarded heap now captures stack traces on
alloc and free.
A crash due to hitting a guard page or an already freed page now dumps
these stack traces. In the case of use-after-free one can therefore see
both where the allocation was done and where it was freed.
Note that there is a hardcoded maximum stack trace depth of 50 and that
the alloc stack trace takes away space from the free stack trace which
uses up the rest of that maximum.
The get_stack_trace syscall generates a stack trace using the kernel
debugging facilities and copies the resulting return address array to
the preallocated buffer from userland. It is only possible to get a
stack trace of the current thread.
The lookup_symbol syscall can be used to look up the symbol and image
name corresponding to an address. It can be used to resolve symbols
from a stack trace generated by the get_stack_trace syscall. Only
symbols of the current team can be looked up. Note that this uses
the symbol lookup of the kernel debugger which does not support lookup
of all symbols (static functions are missing for example).
This is meant to be used in situations where more elaborate stack trace
generation, like done in the userland debugging helpers, is not possible
due to constraints.
* Add a BRow default constructor that use font size to compute height.
* Min height size for Title and Row are decoupled.
* The font ratio for Title and Row are decoupled.
* For small font use min height (set to usual 16.0).
* Better baseline formula.
* Fixes#11944.
This allows for something similar as was implemented in 217f090 but
makes it optional and configurable.
The MALLOC_DEBUG environment variable now can take "a<size>" to set
the default alignment to the specified size. Note that not all
alignments may be supported depending on the heap implementation.
* So that you know how much already was, and still has to be downloaded.
* Automatic whitespace cleanup.
* The link in FetchFileJob.h did not fetch the correct header under
Haiku anymore (since the addition of the private headers to the
image).
* drop my fdt tests
* we have to call fdt parsing code *after* cpu_init (why?)
* pass fdt pointer to all FDT support calls to avoid confusion
once we get into the kernel land
* look for PL011 compatible uart and use it
* Add some saftey checks to serial putc code to avoid null*
* fdt_node_check_compatible returns 0 on success not 1
* fdt_get_device_reg needs to add the SOC base to the result
* fdt_get_device_reg might need to add the second range cell
instead of reg?
* And call it from the DNS client -- this is the only add-on that does
not trigger a configuration or settings update, so we have to notify
the changes manually.
* Get{Interface|Network|Service}() methods are now const.
* Added variants of Interface(), and Service() that are const.
* Added new BNetworkServiceSettings::IsRunning() convenience method that
reflects the status quo rather than the settings.
* Added BNetworkInterfaceListItem that can be used by interface protocol
add-ons to represent their functionality in the list view.
* It will automatically update itself on changes, and will show the
specified label, and address, if any, as well as indicate whether or
not the family has been disabled (which doesn't work perfectly yet,
as IFF_AUTO_CONFIGURED is on the interface level).
* Therefore, the interface list item will no longer show the address,
but the type of the device instead.
* Introduced the BNetworkConfigurationListener interface that is used
to broadcast network updates to.
* You can now optionally specify a converter that does the actual
message to settings conversion, and vice versa.
* This allows for more sophisticated and custom conversions.
* Network now monitors all network, and network settings changes, and
will notify all add-ons about those changes.
* Removed the global apply button. Instead, the static IP configuration
now got that button. All other changes will be instant.
* Added (still incomplete) helper classes for the settings messages.
* The net_server now uses these classes for its interfaces, and services.
* Renamed service_address to service_connection, as that better matches
what it is used for.
* As experimental API class BNetworkSettings.
* Added add/remove methods for interfaces, and services, too.
* Moved the conversion of the wireless networks into the settings class,
too, so that it only gives out converted ones (but accepts both
variants).
* The interfaces logic is within the main application, so is the
new interface view.
* Disabling/renegotiating does not work yet.
* Reveals a bug in some interface code; when you click on an ethernet
device first, there is no place for the wireless menu. When you then
press on a wireless device, it crashes in BMenuField code.
In the other direction, there is then an empty space.
* Fixed list item width reporting.
* Added an extra looperName argument to the private BApplication
constructor that BServer is using.
* This is now used to fix the ugliness that gave the registrar
a different looper name, and even saves a string comparison.
* Move more code into fdt_support
* We now can query FDT registers based on name or alias
* Return addr_t where it makes sense
* Copyright change ok'ed by mmu_man
* Called via arm_mailbox_bcm2835 *and* arm_framebuffer_bcm2835
* This is a bit messy. We really should be getting these
chipset-centric bases from the provided FDT / DTB.
* I can't think of a way to redo this without undoing
work towards FDT.
* The Raspberry pi 2 uses a new SoC which differs slightly
from the Raspberry Pi 1.
* Someday these two board targets could go away when we get
FDT support.
* To while there was some compatibility between
BCM2708 and BCM2805, it makes the BCM2806 changes
more confusing. We don't have any valueable BCM2708
targets.
They are effectively the same thing, but int64 is more generic.
I used an off_t because that is what Tracker passed in, yet, this could be
used in other places where an off_t doesn't make as much sense.
Which doubles the maximum width fixing #11809. I wanted to convert this to
a variadic functionbut that would require updating all of the instance that use
MessageFormat or a clever hack to get around the ambiguity of an off_t
argument and an int argument count with 0 arguments. Either way seemed
out of the scope of fixing this bug.
Fixes#11809
Another Beta1 bug down!
Implement the todo in Open Target Folder to select the result by sending
Tracker a message with the refs to select.
Add a new kSelect message constant to tracker_private.h to hook into.
Tracker reads in the message and finds the window that the refs are located in,
activates it, then selects the corresponding poses scrolling the pose view
if necessary.
Bump version of Open Target Folder to 1.0.2
Intel changed the PCH interrupt bits between Sandy Bridge and Ivy Bridge
to make space for the 3rd display pipe. Take this into account and check
for the correct bits on the newer devices.
Fixes#11522.
and remove TExpandoMenuBar as a friend class to BMenu hack.
In Deskbar, call the newly added MenuPrivate::SetLayout() method instead.
This is a much cleaner way to implement this, thanks Stefano Ceccherini
for the tip!
Add private BMenu::_SetMenuLayout() method. Set TExpandoMenuBar
as a friend class in BMenu to call this method. A little hacky,
but, this keeps SetMenuLayout() from being exposed as part of
the public API.
Don't destroy and rebuild the ExpandoMenuBar when switching from
horizontal to vertical mode. Instead build the TExpandoMenuBar
when the application starts and then switch it from B_ITEMS_IN_ROW
to B_ITEMS_IN_COLUMNS by using the newly added _SetMenuLayout()
method.
When we resize from vertical to horizontal, recalc the max
menu item widths, this resizes the application menu items so
that they take up the right amount of space.
Since we no longer destroy the menu bar we no longer have to
save whether menu items are expanded or not in a separate list.
Instead we can store that information in directly in
TExpandoMenuBar. This removes a lot of code.
Fixes#9350
* When RemoveSelf() is called, we do not own our own layout items, so
we must not delete them.
* However, we do own them when we still have layout items left when
we get deleted ourselves.
* This fixes removing/adding a child view to a view without deleting
it inbetween (like the new Network preferences will do).
* Optimized item removal -- not a good idea to always remove item 0.
"BToolBar" matches the current convention ("BStatusBar", "BMenuBar", etc.)
I've no idea what I was thinking when I renamed this before...
Also adjust all users of BToolBar (Tracker and ShowImage).
This reverts commit 31ea76548a.
Adrien, please try again without clobbering the otherwise nice
BNetworkInterface API!
Conflicts:
src/kits/network/getifaddrs.cpp
This is a BReference that allows only const access to the referenced
object. This was not easily possible with the existing BReference for
two reasons:
* BReference<const Type> would not work, as BReference needs to change
the reference count of the referenced object. Adding mutable and casting
where appropriate wouldwork but,
* It is now also possible to assign a BReference to a BConstReference
(to the same type, of course). The reverse is not allowed, making it
more difficult to "const cast" the referenced object (it's still
possible to "get" the object pointer and cast that).
BConstReference can be used to provide shared read-only access to an
object, for example this can be used to cache non-copiable or
expansive to create objects.
* Added new truncation mode B_NO_TRUNCATION.
* The Truncation()/SetTruncation() methods itself are Dano-compatible,
however, there was no B_NO_TRUNCATION.
* BNetworkInterfaceAddress is moved to libnetwork. It is modified to not
use BNetworkAddress (which is in libbnetapi) and instead use sockaddr
and sockaddr_storage directly. All callers are adjusted to this.
* Some support code is shared between BNetworkInterface and
BNetworkInterfaceAddress, move it to libnetwork but in the BPrivate
namespace.
The atomic inlines were not implemented in a C89 safe way:
* Use of "static inline" not allowed, but static __inline__ is
* __inline__ is a GCC extension, but these are already in a __GNUC__
block (other compilers use a non-inline version)
* also fix a C++ style comment
* The default notifier didn't always take the setting into account.
* The mail server was not using the setting from the settings file and
instead waiting for a message that wasn't sent anywhere.
Fixes#10852.
* Use the preferred time source (GetTimeSource) for the node
* Fix node releasing when creating the connection fails
* Add virtual slots and padding
* Refactor _Connect method
* offsetof is not allowed on non-POD types so we need to use
offset_of_member (gcc2 accepts offsetof, and C++11 relaxed the
constraints on where it is allowed so it should work there too)
* we have offset_of_member as a workaround until we switch to C++11,
move it from khash (which is soon to be removed) to list.h which is the
other place where it is used (for this one single call in our whole
codebase)
Also fix a typo in vfs.cpp.
* Changed the way the attributes are written to make sure that everything
that can be written once is in fact written just once.
* The rename code in BMailProtocol::_ProcessFetchedHeader() was broken,
and caused the hang of the last commit.
* BMailFilter::HeaderFetched() now only alters the entry_ref, and returns
B_MOVE_MAIL_ACTION to move a mail.
* Instead of potentially moving the file around several times, the
BMailProtocol now takes care of carrying out the filter action just once,
including trying to make the file name unique.
* This also allows the IMAP add-on to know the final location of the mail,
and thus downloading a message actually works.
* However, with my test inbox, it currently hangs, and a current Debugger does
not work on my older system -- I guess I need to update.
* Replaced the duplicated space mechanism within the "HaikuMailFormatFilter"
that is substantially faster, and handles all whitespace, not just spaces.
It will also replace tabs with spaces.
* Instead of abusing BArchive::Archive() we now use a BMailSettingsView as
a base view for all filter/protocol settings that works with
BMailAddOnSettings.
* Cleanups in E-mail which is now completely layout friendly. But also still
crashes when changing the views.
* The path will now be relativized before storing it.
* On load, the add-on will be tried to load from the user, then common
and finally system add-on directory.
* Not everything compiles; all protocols, and inbound filters do, though.
* Renamed a few classes to give a better idea what they are for; prefixed
public classes with the 'B' prefix.
* Moved ProtocolConfigView's classes into the BPrivate namespace.
* Moved BMailFilter into its own file.
* Added BMailFilter::DescriptiveName(). This is now used by the RuleFilter
in order to give a description of what it's doing (ie. no more dozens of
"Rule filter" entries in the preferences).
* Removed no longer used MailAddon.h.
* Renamed Addon to AddOn where found, since that is more consistent with the
rest of the API.
* Merged the former MailProtocol with the former MailProtocolThread; the
differentiation between those two was pretty messy.
* All configuration views touched so far are now using the layout kit.
* The RuleFilter is currently broken functionality wise; I have not yet decided
how to solve the stuff it uses (TriggerFileMove() does not exist anymore,
for example).
* BMailAddOnSettings (formerly known as AddonSettings) now directly subclass
BMessage; there are no Settings() and EditSettings() method anymore. The
class uses a copy of itself to determine whether or not it has been changed.
* Lots of cleanup.
* Renamed to BMailNotifier, as it's part of the public API.
* Renamed Notifier.{cpp|h} to DefaultNotifier.{cpp|h} as that's the class it
implements.
* Made the mail counts uint32, and the byte counts uint64.
* Renamed imap_config.cpp to ConfigView.cpp, IMAPFolderConfig.(h|cpp)
to FolderConfigWindow.(h|cpp).
* Got the latter to build.
* Added Settings class to simplify and unify the BMessage based
settings access.
* Removed the InboundProtocol[Thread] implementation from the build for now.
* This was never implemented and no one noticed until now.
* A default value for a pointer doesn't make that much sense anyway, so
using the FindPointer method is fine.
int64_t is signed. Although it does not make a difference by itself, because
INT64_MAX is still a valid number for uint64_t (UL), the later INT64_MIN
declaration depends on INT64_MAX, and therefore got implicitly casted to
unsigned type.
This fixes the following program on a x86_64 system:
#include <stdint.h>
int main() {
int64_t test = 5;
if (test < INT64_MIN)
return 1;
return 0;
}
This is a regression since commit 1d13a609 ("stdint.h: define [U]INT64[MAX|MIN]
with [U]L on x86_64").
Signed-off-by: Jerome Duval <jerome.duval@gmail.com>
Created docs for NumberFormat, DateFormat, DateTimeFormat, and TimeFormat
and moved the docs from BLocale to the respective new file. Also
DurationFormat was updated as well.
doxygen once again compiles the docs without warnings.
There doesn't seem to be anything ini the implementation that would
cause a problem, as long as you don't try to change this while the
window is already open.
gcc2 was relying on the c99 functions being there, but they are not in
the std namespace.
* Disable the C99 functions and macros in C++ mode
* Redefine them as inline functions in cmath in the std namespace.
Fixes#7396.
* There is no need to delay this to AllAttached
* Apps may want to override the SetDivider, and doing it as late as
AllAttached can be annoying.
Fixes#10734.
I already made this patch in Heidi's clone of BToolbar. Adding it here
allows me to drop Heidi's version altogether.
I also relicensed the header file at the permission of Stephan. The .cpp
file still lists him as author and copyright holder.
Signed-off-by: Jessica Hamilton <jessica.l.hamilton@gmail.com>
* Make it possible to extract more useful data from the certificate
* Also get the OpenSSL error message when a certificate can't be
validated. Send it to the verification failure callback so it can be
shown to the user.
CreateThreadEvent::DoDPC() missed a reference release to balance the
acquired reference before queuing the DPC, resulting in the
CreateThreadEvent objects being leaked.
This also removes the destructor that tried to cancel the DPC. Since
the class is reference counted and only destroyed when the DPC has
run and released the last reference, this didn't make much sense.
The signal to the team/thread is only actually sent in a deferred
procedure. To ensure that the team/thread stays valid between the DPC
being queued and it actually running, we need to acquire a reference.
Fixes#11390, where the DPC was run after the team was already
destroyed.
This isn't really a bus_manager yet, but just minimal support so
we can get rid of hardcoded ARM SoC support from the core kernel
code.
Needs lots of work, like proper handling of #address-cells and
the like. Also, generic attribute handling, device_manager
integration, and I could go on for hours ;)
* Partly reverting hrev47655, as the moved declarations are expected
by many ports to be accessable via string.h.
Following standards is a good thing in general, but not if it causes
more problems than it helps ...
* VMArea::AddWaiterIfWired(): Replace the ignoreRange argument by a
flags argument and introduce (currently only) flag
IGNORE_WRITE_WIRED_RANGES. If specified, ranges wired for writing
are ignored. Ignoring just a single specified range doesn't cut it
in vm_soft_fault(), and there aren't any other users of that feature.
* vm_soft_fault(): When having to unmap a page of a lower cache, this
page cannot be wired for writing. So we can safely ignore all
writed-wired ranges, instead of just our own. We even have to do that
in case there's another thread that concurrently tries to write-wire
the same page, since otherwise we'd deadlock waiting for each other.
The thread that is being [un]scheduled already has its time_lock locked
in {stop|continue}_cpu_timers(). When updating the TeamTimeUserTimer,
the team is asked for its cpu time. Team::CPUTime() then iterates the
threads of the team and locks the time_lock of the thread again.
This workaround passes a possibly locked thread through the relevant
functions so Team::CPUTime() can decide whether or not a thread it
iterates needs to be locked or not.
This works around #11032 and its duplicates #11314 and #11344.
when uninitializing a partition or a disk (removing the partition
table), check that all partitions from that table are unmounted, as they
are about to become invalid.
Fixes#8827.
The language and formatting conventions can now only be set when
creating the objects. This removed the needs for locking them when
formatting to avoid some other thread changing the format while it's
being used.
Adjust tests and DeskBar TimeView to the API changes.
Although done with the best intentions, the usage of #warning in the
ARM build makes it hard to see what's going on, or see any "real"
warnings.
Remove at least this particularly often triggered one, so we can
have a relatively "quiet" build again...
* Added VFS helper function check_access_permissions() that combines
several partially correct versions to the one true version (tm).
* All but BFS (since recently) missed the S_IXOTH for root on directories,
and all but packagefs missed proper group handling.
The BOOT_GDT_SEGMENT_COUNT was based on USER_DATA_SEGMENT on both
x86 and x86_64. However, on x86_64 the order of the segments is
different, leading to a too small gBootGDT array. Move the define to
the arch specific headers so they can be setup correctly in either case.
Also add a STATIC_ASSERT() to check that the descriptors fit into the
array.
Pointed out by CID 1210898.
There were 3 setting files, each an archived BMessage. Now there is only
one with all the data inside.
* Rework the SettingsPane class to save and load settings from a
message, rather than having each panel pick a file path on its own
* Move saving the app filters to the preference app, rather than the
server (so it's done at the same place as other settings)
* Rework loading prefs in the server so the settings message is read
from the file once and all settings are loaded from it.
This turns out to be more changes than I anticipated.
Fixes#9424.
* Since DNS are normally restricted to ASCII, the use of UTF-8 in domain
names is implemented using a "punycode" encoding.
* The request to the DNS server must be sent with the ASCII
representation of the domain name, however the Unicode one should be
used for user-visible parts.
* ICU provides an implementation of the conversion, which we use here.
* Conversion is currently done in-place and modifies the BUrl object
(this is similar to UrlEncode/UrlDecode).
* Adjust existing IDN test to make use of these methods. It's passing
now.
Axel Dörfler