Net_server starts services by invoking fork() followed by exec(). If the latter
fails (for instance because the service isn't installed), the forked child is
invoking exit(). This in turn unloads libbe, triggering static cleanup code in
BMessage, which deletes a couple of message ports that were inherited from the
parent during the fork. After that, net_server was desparately missing those
ports and no longer worked reliably.
* in InitTerminateLibBe, we now register an atfork-(child-)handler, which
takes care to re-initialize the static reply ports used by BMessage code
* added BMessage::Private::StaticReInitForkedChild wrapper and
BMessage::_StaticReInitForkedChild() implementation which overwrites the
inherited port IDs with a set of own ports
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@33050 a95241bf-73f2-0310-859d-f6bbb57e9c96
* The BView API can probably be regarded as good enough; the implementation
might need to be improved over time (also, some things as archivability
aren't fully implemented yet). The ToolTip.h header should get public once
finalized.
* Added new B_MOUSE_IDLE message that is sent to a BView after a certain
time has passed (BToolTipManager::ShowDelay()).
* Added small test app (ToolTipTest) that shows what is already working.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32078 a95241bf-73f2-0310-859d-f6bbb57e9c96
storage.
* Remove reduntant information from the header (field_count vs. fields_size).
* Remove checksums previously used to validate the message passing by area
mechanism.
* Move variables that are purely used by the instance out of the header and into
the BMessage object.
* Use more sensible types for the different message fields.
* Reduce some field sizes to realistic values.
* Make size_t values into uint32 values so the message format will not change
when later moving to 64 bits.
* Pack the structures used for flat message storage so it doesn't change
because of padding.
* Fix message passing by area. It never worked because the created area was
never actually filled with any data!
* Some more allocation checks with graceful fallbacks (should be all now).
* Some more checks for negative index values (should also be all now).
* Make printing more inline with how the rest of the class works and make some
of the output more consistent.
* Also add the new unsigned types to PrintToStream() output.
* Fix printing of unknown types and invalid BMessages, it would always have
printed only the first entry respectively the same error.
* Added some clarifying comments.
* Cleanup.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@32039 a95241bf-73f2-0310-859d-f6bbb57e9c96
window of the view into the application thread. This solves the
race condition with asynchronous SetViewCursor and deleting the
cursor immediately afterwards for real.
* The ServerApp now requires a reference to the current cursor,
just in case...
* Added TODOs for caching the BView token, it's currently resolved
for every single BView call that talks to the server... not good!
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@31133 a95241bf-73f2-0310-859d-f6bbb57e9c96
reference counting is maintained correctly in the app_server. While reviewing
this code, I have my doubts that my previous solution for handling pending
SetViewCursor() calls is always working as it is intended.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@31131 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Enable -Werror on GCC4 builds as well (limited to the same selected targets).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@29118 a95241bf-73f2-0310-859d-f6bbb57e9c96
bool return type. Did I mention this was an unstable Haiku extension?
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@28181 a95241bf-73f2-0310-859d-f6bbb57e9c96
the data of two BMessages and allows recursive comparison (BMessage inside
BMessage). Note that using this API might require you to recompile your app in
the path to R1.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@28174 a95241bf-73f2-0310-859d-f6bbb57e9c96
the message count doesn't really make much sense anyway (as opposed to
limiting the maximum size of waiting messages), we might still want to
enlarge this even more.
* This should further help with #2212.
* Minor cleanup.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@25709 a95241bf-73f2-0310-859d-f6bbb57e9c96
the new sematic of transfer_area so a message area is transfered into the right
teams' address space and it does not need to be cloned there anymore. Passing
by area is only used for messages bigger than a certain size (currently
hardcoded to 40KB) which should be somehow bound to the max port message size.
This makes passing large messages (i.e. > the port limit) possible, so for
example copy&paste of long text, image data, etc. should now work.
Got rid of the fClonedArea member as it is not necessary with the new design,
renamed shared_area to message_area in the private message_header, avoid
an unnecessary allocation of the header for the copy constructors, check
allocations in a few more places and some minor cleanup.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@24321 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Removed storage/Alias.h as that stuff isn't even available on BeOS.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@21576 a95241bf-73f2-0310-859d-f6bbb57e9c96
BDirectory without declaring it first. Since that's an actual bug, I made the
changes in the Haiku headers, and copied them back to the build headers.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@21562 a95241bf-73f2-0310-859d-f6bbb57e9c96
* _WindowAt() and _CountWindows() now have an individual version of that
code which should be magnitudes faster.
* _WindowQuitLoop() no longer handles hidden windows specially - instead,
it now walks the window list in the correct direction which should fix
the issues.
* Also, it now uses WindowAt() and thus has an up-to-date view of the
window list (it will no longer ignore new windows).
* And finally, it will no longer dereference an unsafe pointer (for
BWindow::IsFilePanel()).
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@21505 a95241bf-73f2-0310-859d-f6bbb57e9c96
part of the view, you can add a BRect "be:area" to that message - very handy.
* Cleaned up AppDefs.h a bit.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20683 a95241bf-73f2-0310-859d-f6bbb57e9c96
this fixes bug #1005. As a result, the Disks icon will now appear in file panels
when you change that setting with a panel open.
* _ObserverList is now in the BPrivate namespace (and renamed to ObserverList).
* its BHandler map is now only temporarily used for handlers that do not belong to
a looper yet; when BHandler::SendNotices() is called, they will be transferred
into the BMessenger map.
* Invalid messengers are now removed from the map when encountered.
* Added TODO comments about a possible reference counting if a handler is added
twice to a list (right now, this will be ignored).
* All {Start|Stop}Watching() methods are now more or less safe to be used in low
memory situations (adding some items to the map can still throw an exception...).
* Renamed BHandler::InitData() to _InitData().
* Some refactoring and cleanup.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@20029 a95241bf-73f2-0310-859d-f6bbb57e9c96
with PostMessage() in case the message queue is full.
Some notes:
* for synchronous replies, we don't use this mechanism yet, but it could be
extended to do that as well.
* the code looks so complicated because we need a way to access the looper's
queue without locking it (to prevent deadlocks); like Dano's solution, I've
abused BTokenSpace to store a BDirectMessageTarget with a BHandler.
* we also need to decouple the lifetime of a looper's queue from its target,
as we cannot lock the looper, and therefore, can't guarantee it stays valid
as long as we're accessing it outside of BLooper.
* init_clipboard() now needs to be done after the global constructors have
been called - since sending messages now needs gDefaultTokens to be initialized.
Since this is done per image, it shouldn't cause any troubles, though.
* some minor cleanup, removed unused _msg_cache_cleanup_() and friends.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19968 a95241bf-73f2-0310-859d-f6bbb57e9c96
* Got rid of unused BLooper members
* renamed fTaskID to fThread
* Removed private and deprecated AddLooper()/RemoveLooper()/... stuff; BLooper is now
directly calling BLooperList methods.
* Got rid of extensive and useless comments
* Made a few TODOs more clear
* Merged InitData() and InitData(...) to _InitData(...)
* BLooper::Team() now uses BPrivate::current_team(), sTeamID is gone now.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19966 a95241bf-73f2-0310-859d-f6bbb57e9c96
* BLooper::AddMessage() is now using this method to determine wether or not to
update its looper thread.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19956 a95241bf-73f2-0310-859d-f6bbb57e9c96
Switcher now works as expected.
* Renamed TRoster::ActivateApp() to UpdateActiveApp(), as the app is already activated
at that point (the registrar only keeps track of it).
* BWindow::DispatchMessage() now calls the new BRoster::Private::UpdateActiveApp()
method when it receives a B_WINDOW_ACTIVATED message.
* Added BRoster::_UpdateActiveApp() which calls the new B_REG_UPDATE_ACTIVE_APP.
* Removed now unused B_REG_ACTIVATE_APP.
* Minor cleanup.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@19857 a95241bf-73f2-0310-859d-f6bbb57e9c96
in several steps:
1. early pre-registration with the registrar ("I wanna launch the
app, make sure noone interferes.")
2. load the app image
3. finish pre-registration with the registrar ("I have launched
the app, here is its team ID.")
4. start app main thread
5. send "on launch" messages to the app (argv, refs, others)
If the app is already running or being launched, 1. fails with a
conclusive error code and returns the team ID and the pre-registration
token of the app. Steps 2 - 4 are skipped and only the messages are
delivered using the team ID returned by 1.
This change fixes a race condition: The failed early pre-registration
request obviously cannot return the team ID, if the other thread
launching the app has not finished step 3 yet. Thus the argv/refs
message would not get delivered and Launch() would not return the
correct team ID.
Now we wait for the pre-registration to be finished in this case, using
the former _IsAppPreRegistered() mechanism, which already provided
such a waiting feature for one request. It has been extended to
accomodate an arbitrary number of waiting requests and renamed to
_IsAppRegistered().
This fixed bug #763.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@18728 a95241bf-73f2-0310-859d-f6bbb57e9c96
translators are installed via BTranslatorRoster::StartWatching().
* Fixed GetTranslatorData() - some hooks are only optional (this puts the
JPEG, GIF, and some other translators back into the game).
* Made BTranslatorRoster::Default() thread-safe.
* Improved documentation.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@17267 a95241bf-73f2-0310-859d-f6bbb57e9c96
read-only area shared between the Desktop and all applications.
* Right now, this area only contains the desktop colors, ie. B_PANEL_BACKGROUND_COLOR
etc.; ui_color() no longer needs to ask the server for these colors.
* The ui_colors are now maintained by DesktopSettings, though ColorSet is still there.
* The default colors are now hardcoded once and for everyone in InterfaceDefs.h, ie.
the app_server uses them as well.
* Desktop::Init() can now also return an error (but that is not yet accounted for).
* Cleaned up InterfaceDefs.h.
* Fixed wrong include in moreUTF8.h.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@17232 a95241bf-73f2-0310-859d-f6bbb57e9c96
have a static method for spawning one (BMessageRunner::StartSending()), as
suggested by Ingo.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@16966 a95241bf-73f2-0310-859d-f6bbb57e9c96
that is detached from its local object; ie. it will continue to send its
messages when you delete the object.
* Minor cleanup.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@16730 a95241bf-73f2-0310-859d-f6bbb57e9c96
This will allow to fix message related problems like drag and drop and scripting.
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@16708 a95241bf-73f2-0310-859d-f6bbb57e9c96
* all cursors owned by a team are visually different,
or (iaw) an already existing cursor is reused when
it is set by the client again
* changed various occurances of cursor data from "int8*"
to "uint8*"
* ServerCursors also remember the R5 data from which
they were created
* the reference counting and destruction of
ServerCursors changed: The cursor knows it is attached
to a CursorManager and one can simply use
ServerCursor::Acquire() and Release() and the reference
counting and everything is being taken care of
* destroying a ViewLayer will now correctly release a set
ServerCursor
* fixed a race condition when setting a cursor through
BView::SetViewCursor(): If the client code looks like this:
BCursor cursor(cursorData);
someView->SetViewCursor(&cursor, false);
there is a relatively high chance the BCursor destructor
told the ServerApp thread to destroy the cursor before
the ServerWindow thread got to "acquire" the cursor for
use by the view layer. The very same problem is likely the
reason that SetViewCursor works to unreliably on R5, even
when the "sync" flag is set to "true" (although it should
theoretically work in that case).
all these fixes make WonderBrush work fine again with the
new support of custom cursors.... coded by axeld and myself
(the joys of pair programming :-)
git-svn-id: file:///srv/svn/repos/haiku/haiku/trunk@16521 a95241bf-73f2-0310-859d-f6bbb57e9c96
