This makes for much faster lookups, and avoids the need
to allocate an array.
sizeof(BWindow) is unchanged (on x86_64, 504; on x86, 376)
so this should not break ABI/API.
Quick benchmark in Tracker:
501 lookups, before: ~488 usec
501 lookups, after: ~284 usec
Change-Id: Ica2c728f0c5957380add53e14a826a4aade52df0
Reviewed-on: https://review.haiku-os.org/c/haiku/+/10720
Tested-by: Commit checker robot <[email protected]>
Reviewed-by: waddlesplash <[email protected]>
This paves the way for accelerants to map the framebuffer in
user memory, not kernel memory. It uses the ServerMemoryAllocator
in order to avoid mapping the framebuffer multiple times in
client applications.
While at it, clean up some names.
Tested with DirectWindowStars and GLTeapot, both work.
Change-Id: I4e0c003fae99891044af0a5e6e1d6a4506da6430
Reviewed-on: https://review.haiku-os.org/c/haiku/+/10571
Tested-by: Commit checker robot <[email protected]>
Reviewed-by: waddlesplash <[email protected]>
CreateConnection
Move expecting commands from RemoteDevice::Authenticate to where they
are called.In SSP we expect for the events after ConnectionComplete.
For the state machine, LINK_KEY_NOTIFY tells us the end of the pairing
in both pairings, so we clear all expected events there for the ones we added in
ConnectionComplete, In SimplePairingComplete we clear the ones added
in IOCapabilityRequest. Tested with both bluetooth 2.1 and older pairing.
Change-Id: Ia208506c9f289d9f44dc3b8b0e4cd8a726a5d726
Reviewed-on: https://review.haiku-os.org/c/haiku/+/10526
Reviewed-by: waddlesplash <[email protected]>
This parses only EIR_NAME_SHORT and EIR_NAME_COMPLETE, others are not
yet usefull, adds a way to fetch the cached friendly name and adds a
friendlyName field to RemoteDevice. Refactors DeviceListItem to work on
RemoteDevice as it's only being used for that.Also adds a check if the
cached name is complete or not. If not ask for the complete name.
Change-Id: I7e37fb9cf44cb5598ef348fdd4d781c5ae04e24e
Reviewed-on: https://review.haiku-os.org/c/haiku/+/10489
Reviewed-by: waddlesplash <[email protected]>
Tested-by: Commit checker robot <[email protected]>
Converting points array to `BPolygon` has significant cost, including
heap allocation, when drawing a lot of polygons.
Change-Id: I81e6b4031bac4beadcc79412c58eedddffc00a82
Reviewed-on: https://review.haiku-os.org/c/haiku/+/10337
Reviewed-by: waddlesplash <[email protected]>
This adds a new job to the Package Kit, and an invocation
in pkgman after changes are applied. The job takes a time_t before,
and an int32 minToKeep, and (after confirmation) deletes all state
directories which were last modified before that time, but also
keeps at least the minimum specified.
pkgman defaults to calling it with (now - 30 days, minimum 10).
SoftwareUpdater does the same, but without bothering the user
and just defaults to doing the cleanup.
Change-Id: I15f5232b11daba5955e7fa07f696ad7785507931
Reviewed-on: https://review.haiku-os.org/c/haiku/+/10404
Reviewed-by: waddlesplash <[email protected]>
It doesn't refer to some random old state, but the currently active
one. (If the currently active state is the default one, then it will
return nothing.) So, this should make things clearer.
Change-Id: Ib3fe842f5fb51eaf2ef2f31bad8b292b47a3fb36
Reviewed-on: https://review.haiku-os.org/c/haiku/+/10403
Tested-by: Commit checker robot <[email protected]>
Reviewed-by: waddlesplash <[email protected]>
The lighter "struct mutex" is in a private header, so we can't
inline it in a public one, so just use pthread_mutex here.
By adjusting padding, the class size stays the same (72 bytes
on 32-bit, 96 bytes on 64-bit; confirmed via static_assert.)
A quick benchmark running "new/delete BMessage" in a loop on 4
threads at once goes from 30-36 seconds before this commit
to around 13-17 seconds, sometimes as low as 3 seconds, afterwards,
so clearly this is a significant improvement.
This also eliminates another BLocker allocated on application startup.
... instead of GetIterator(). This allows us to avoid loading the
cache entirely in SetTo(), and allows consumers to read the cache
much more efficiently than loading it all into memory and then
freeing it again afterwards.
This technically breaks API/ABI, however the Package Kit APIs
are not considered stable (I don't think.)
All consumers adjusted. As we build the host tools with modern GCC
only, I made update_package_requires use a C++11 lambda function.
All others use out-of-line static methods.
"time pkgman list-repos -v" (which has to read the whole cache)
is ~0.5s before this change, and ~0.25s after this change,
on my test VM.
Change-Id: I6976b4cf5eb846fc925ed199dc00eb227fc81344
Reviewed-on: https://review.haiku-os.org/c/haiku/+/10247
Tested-by: Commit checker robot <[email protected]>
Reviewed-by: waddlesplash <[email protected]>
Reviewed-by: Andrew Lindesay <[email protected]>
* Use LITTLE_ENDIAN instead of __LITTLE_ENDIAN (LITTLE_ENDIAN is what
POSIX now specifies and everyone seems to define - NetBSD also
defines _LITTLE_ENDIAN, but not __LITTLE_ENDIAN)
* file descriptors greater than 2 must be explicitly passed to a child
on NetBSD. POSIX leaves that unspecified, see
https://pubs.opengroup.org/onlinepubs/9799919799/utilities/V3_chap02.html#exec
* system library functions get declared with visibility "default" on
NetBSD, so a "hidden" visibility attribute is ignored on a
redeclaration. Work around by putting the functions in a namespace -
as they are extern "C", they still end up in the object file without
the namespace qualifier.
* remove the cast on the NULL pointer on the funopen call - it's not
needed and the types are different between FreeBSD (fpos_t) and
NetBSD/OpenBSD (off_t); also trying to change that upstream
https://github.com/openSUSE/libsolv/pull/603
Change-Id: I6ab7f3c74d18960d7589b403a1c219cdac85a453
Reviewed-on: https://review.haiku-os.org/c/haiku/+/10133
Tested-by: Commit checker robot <[email protected]>
Reviewed-by: Adrien Destugues <[email protected]>
In BeOS, the tab label was synchronized with the view name. This means
BTab::SetLabel would change the view name. This behavior is unexpected,
and also annoying, for example it prevents using localized tab labels
but fixed view names for scripting.
* Document this divergence from BeOS in the Haiku book
* Detect if we're running in a BeOS app to preserve the old behavior
then
Change-Id: I5b758a035fe8752cee2ea7789a30ea47a01467b9
Reviewed-on: https://review.haiku-os.org/c/haiku/+/122
Tested-by: Commit checker robot <[email protected]>
Reviewed-by: Jérôme Duval <[email protected]>
Like check_access_permissions() but for write_stat() hook instead
of access().
Basic logic copied from BFS, but with the UID/GID logic rewritten
to better conform to behavior on other OSes: notably, we allow
chown() requests that keep the UID the same to go through, same
as Linux seems to, which should fix#19666. We now also validate
that the GID is one of the current team's groups.
BView-related code removed or disabled, of course, but this now uses
the ColorConversions class, and adds some more ImportBits APIs from that.
Change-Id: I805cef159bac52173ef16030eae69e83db6f061b
Reviewed-on: https://review.haiku-os.org/c/haiku/+/9391
Reviewed-by: waddlesplash <[email protected]>
The content of a BUrl should always be in encoded form, to simplify
handling and validation.
Deprecate the UrlEncode member function and make it private. Instead aadd a new
way to handle URL encoding:
- All ways to set an URL (constructors, SetUrlString, and all setters)
now take an extra boolean parameter indicating if the string is already
encoded. The default value is to encode strings automatically.
- The static version of UrlEncode and UrlDecode, which operate on a
string, are preserved and used by other parts of the API.
All unit tests adjusted to handle this, and still passing.
Fixes#12983
Change-Id: I06f06978d0d35e56d7c92b67f001856bb7dcafc8
Reviewed-on: https://review.haiku-os.org/c/haiku/+/1193
Reviewed-by: nephele nephele <[email protected]>
Reviewed-by: Adrien Destugues <[email protected]>
* Introduce notify_query_entry_moved to the fs_interface API.
axeld added a hook to the query parser (then in BFS) in hrev34317
to directly handle moves/renames, but just sent notifications for
it as B_ENTRY_REMOVED + B_ENTRY_CREATED (which is in fact the same
exact thing BeOS R5 did, I tested). We need to preserve that ABI/API,
unless flags that didn't exist on BeOS R5 are set.
* Rework notify_query_attr_changed to notify_query_attribute_changed.
This method is Haiku-specific (it was added in hrev39131), but it
wasn't very useful: it only sent the name of the file, not the name
of the attribute that changed.
It now sends the name of the attribute along with an "int32 cause",
same as notify_attribute_changed does.
* Rename the private flag B_ATTR_CHANGE_NOTIFICATION to B_QUERY_WATCH_ALL.
This better describes what it does (it's analogous to the B_WATCH_ALL
node_monitor flag). Probably we should make it public at some point...
(although probably at a lower value.)
* Overhaul QueryParser notifications logic to implement WATCH_ALL
properly, make use of the new and improved notification methods,
and to deduplify some logic.
This makes it possible to use queries with B_QUERY_WATCH_ALL
in a similar fashion to watching a directory node with B_WATCH_CHILDREN,
avoiding the need to watch every single node individually.
ElfW(x): to automatically use the relevant Elf32 or Elf64 type for the
current system
NT_GNU_BUILD_ID: constant defining the GNU "build ID" note in ELF files.
WebKit uses both of these to store info in javascript object caches, so
it can detect if the same version of WebKit is being used and the cache
can safely be reused, or if it should be reconstructed.
Change-Id: Ia9eccc1acc79ac4982b87609d2815c4a611e4176
Reviewed-on: https://review.haiku-os.org/c/haiku/+/9143
Reviewed-by: nephele nephele <[email protected]>
We can't rely on it not changing between calls, especially as
MaxUpdateTextWidth() changes the value and re-gets it, and
that's called at the end of this method always. So we need to
keep our own copy rather than using the one that's owned by
the subclass.
Fixes a use-after-free uncovered by the guarded heap in VolumeSlider,
reported in #19493.
_Uninstall() before _Hide() in BMenu because the window must
be available when we _Uninstall() especially for shortcuts.
Remove #define and always assume USE_CACHED_MENUWINDOW.
_Uninstall() whenever we _Hide() in BMenu.
Change-Id: I5dae85f6edf1f0b4ccf67a6d9d77470576671cee
Reviewed-on: https://review.haiku-os.org/c/haiku/+/9012
Tested-by: Commit checker robot <[email protected]>
Reviewed-by: waddlesplash <[email protected]>
The shortcut detection was working, just not the display of the
modifiers in the menu. I've added back the necessary code to fix
this in BMenuItem.
BWindow does the heavy lifting of preparing the keys and
modifiers. I have changes _FindShortcut() used by BMenuItem to
send the prepared modifiers back to BMenuItem.
Set the parameters raw in the constructor, they will get fixed
up in Install().
I also make sure to use the prepped version of the key and mods
in BWindow::AddShortcut() to remove the old one. This is a minor
update that eliminates an edge failure case of malformed input.
Fixes#19395 a regression from hrev58589.
Change-Id: I4333f89149ff843f92dbffbd53d58ffc2def6760
Reviewed-on: https://review.haiku-os.org/c/haiku/+/8943
Tested-by: Commit checker robot <[email protected]>
Reviewed-by: John Scipione <[email protected]>
Since BObjectList is a template class, this only breaks ABI where
BObjectList was exposed in public methods, and even then it's only
a name mangling break and we should be able to add compatibility
methods if necessary.
(The old "bool owning" member variable is left intact for ABI
compatibility, for the moment, though it's otherwise unused now.)
Tracker's PoseList is the only remaining type that has a "bool owning"
switch in the constructor rather than template parameters.
This should significantly improve the output of static code analysis
tools that previously detected list operations as causing use-after-frees
and double-frees, as well as make code maintenance easier by making it
easier to determine what list owns (or does not own) an object.
It should also be a minor performance optimization, since the branches
for calls to delete/free should now be optimized out altogether.
Still boots to desktop and Tracker, Deskbar, Debugger all tested
and verified as working.
Change-Id: If2a24a6f0d22e7a506ef554fcfdd328907279ed4
Reviewed-on: https://review.haiku-os.org/c/haiku/+/8915
Reviewed-by: waddlesplash <[email protected]>
Make it clear that this enum just defines ASCII control characters, and
there is no 'free space'. Also add comment showing which key combination
produces each control character.
Change-Id: I01884132842281cce6c13dd9398b1aeddadae58a
Reviewed-on: https://review.haiku-os.org/c/haiku/+/8899
Reviewed-by: nephele nephele <[email protected]>
Tested-by: Commit checker robot <[email protected]>
Reviewed-by: waddlesplash <[email protected]>
Jessica Hamilton