Someone on #haskell noticed this.
Fri Nov 27 23:14:28 GMT 2009 Duncan Coutts <dun...@haskell.org>
* Fix documentation of layoutGetExtents
Crucially it now says which is the ink and logical rectangles.
Also the variable names are now consistent with their uses.
No actual change in behaviour.
If you want to double check, see:
http://library.gnome.org/devel/pango/stable/pango-Layout-Objects.html#pango-layout-get-extents
Fri Nov 27 23:14:28 GMT 2009 Duncan Coutts <dun...@haskell.org>
* Fix documentation of layoutGetExtents
Crucially it now says which is the ink and logical rectangles.
Also the variable names are now consistent with their uses.
No actual change in behaviour.
New patches:
[Fix documentation of layoutGetExtents
Duncan Coutts <dun...@haskell.org>**20091127231428
Ignore-this: 4a1178d41efc4062377e2ac83d8f585f
Crucially it now says which is the ink and logical rectangles.
Also the variable names are now consistent with their uses.
No actual change in behaviour.
] {
hunk ./gtk/Graphics/UI/Gtk/Pango/Layout.chs.pp 699
if idx==maxBound then CursorPosNextPara else
CursorPos (ofsFromUTF (fromIntegral idx) uc) (fromIntegral trail))
--- | Compute the physical size of the layout.
+-- | Computes the logical and ink extents of the 'PangoLayout'.
--
hunk ./gtk/Graphics/UI/Gtk/Pango/Layout.chs.pp 701
--- * Computes the ink and the logical size of the 'Layout'. The
--- logical extend is used for positioning, the ink size is the smallest
--- bounding box that includes all character pixels. The ink size can be
--- smaller or larger that the logical layout.
+-- Logical extents are usually what you want for positioning things. Note that
+-- both extents may have non-zero x and y. You may want to use those to offset
+-- where you render the layout. Not doing that is a very typical bug that
+-- shows up as right-to-left layouts not being correctly positioned in a
+-- layout with a set width.
--
hunk ./gtk/Graphics/UI/Gtk/Pango/Layout.chs.pp 707
-layoutGetExtents :: PangoLayout -> IO (PangoRectangle, PangoRectangle)
+-- Layout coordinates begin at the top left corner of the layout.
+--
+layoutGetExtents :: PangoLayout
+ -> IO (PangoRectangle, PangoRectangle) -- ^ @(ink, logical)@
layoutGetExtents (PangoLayout _ pl) =
hunk ./gtk/Graphics/UI/Gtk/Pango/Layout.chs.pp 712
- alloca $ \logPtr -> alloca $ \inkPtr -> do
- {#call unsafe layout_get_extents#} pl (castPtr logPtr) (castPtr inkPtr)
- log <- peek logPtr
- ink <- peek inkPtr
- return (fromRect log, fromRect ink)
+ alloca $ \inkPtr ->
+ alloca $ \logPtr -> do
+ {#call unsafe layout_get_extents#} pl (castPtr inkPtr) (castPtr logPtr)
+ log <- peek inkPtr
+ ink <- peek logPtr
+ return (fromRect ink, fromRect log)
-- | Compute the physical size of the layout.
}
Context:
[Use simpler embbeded demo replace previous one.
Andy Stewart <lazycat.mana...@gmail.com>**20091126123239
Ignore-this: 51d2f4d8a44bc9a63ed3ce6e7f2f25e6
]
[Remove debugging croft from demo.
Axel Simon <axel.si...@ens.fr>**20091125102701]
[Apply renaming to gstreamer.
axel.si...@ens.fr**20091124223444]
[Get rid of mkFunPtrDestroyNotify.
Axel Simon <axel.si...@ens.fr>**20091124215744
This is a cleanup patch towards having all function that pass Haskell function
closures to C use cleanup functions that are mere C addresses. GObject now
no longer exports mkFunPtrDestroyNotify but two C function addresses that
free function and stable pointers, resp. This patch also fixes the
memory management in Clipboard which is a beast.
]
[Use Gdk.EventM replace Gdk.Events, and make demo simpler.
Andy Stewart <lazycat.mana...@gmail.com>**20091123031143
Ignore-this: b20fa6438a839a94aa5757eabe07e90c
]
[Remove the use of weak reference from gstreamer.
axel.si...@ens.fr**20091122222224
As a temporary hack I removed the code to clean up a function closure. The
old code relied on GWeakRef which has been removed from GObject due to
the GC being able to finalize GObjects directly which will make
GWeakRefs trigger callbacks into Haskell land form the unsafe GC.
]
[Give correct name to the new Utils files in glib.
Axel Simon <axel.si...@ens.fr>**20091122215334]
[Add another demo that uses callbacks. This was mainly to test if -threaded now
works.
Axel Simon <axel.si...@ens.fr>**20091122192920]
[Add a convenience function to extract the index of an iterator.
Axel Simon <axel.si...@ens.fr>**20091122192837]
[Use the C function to finalize dynamic functions.
Axel Simon <axel.si...@ens.fr>**20091122192828
This patch changes mkDestoryNotifyPtr to only return the address of the
ghc C function freeHaskellPtr. Thus, any closures passed to Gtk+
functions are now directly freed by a C function rather than calling
back into the Haskell system. The downside is that this C function does not
know which object to finalize. It is, however, called with the data that
is supposed to be passed to the callback function. This used to be
nullPtr since this argument was not used. To make things work,
this argument now always has to be the address of the callback function
so that freeHaskellPtr will get this address as a an argument.
If I forgot to change any uses of mkDestroyNotifyPtr, then using this
function will lead to a SEGFAULT since freeHaskellPtr is called with NULL.
Gtk2Hs programs producing a SEGFAULT when closing are an indication of this.
]
[Move the weak reference function to disallow weak references from GObjects.
Axel Simon <axel.si...@ens.fr>**20091122185255]
[Add finialization from the Gtk+ main loop.
Axel Simon <axel.si...@ens.fr>**20091122120804
This patch adds a replaces the default destroy function for any GObject that
is created in Gtk+ or any libraries that use Gtk+. These objects are now
finalized using an idle handler that is executed by the Gtk+ main loop.
By being executed by the main loop, the finalizers are run from the same
thread as the main loop and, hence, the objects that hold Xlib or Win32
resources will now free these resources from the thread that normally
calls into Xlib/Win32. This fixes a problem with the -threaded runtime
of ghc in which finalization of objects could happen from other threads
which would cause Xlib errors.
]
[Add a wrapper around the finalizer of Gtk objects.
axel.si...@ens.fr**20091121132129]
[Add demon for cross-process embedded.
Andy Stewart <lazycat.mana...@gmail.com>**20091121171221
Ignore-this: c8d6723ea37ea41bf293ab835a64ab4c
]
[Fix the bug Duncan reported.
Axel Simon <axel.si...@ens.fr>**20091121142249]
[Comment fixes.
Axel Simon <axel.si...@ens.fr>**20091121142205]
[Add a demo for GtkBuilder.
John Millikin <jmilli...@gmail.com>**20091120183943
Ignore-this: ff9747ed9a9245fdf867058750d9f6a7
]
[Add support for GtkBuilder.
John Millikin <jmilli...@gmail.com>**20091120183812
Ignore-this: baf53e703b8b16a6e9ab888c2795e22e
]
[Add misssing property function and remove unnecessary export functions/comment
from Embedding modules.
Andy Stewart <lazycat.mana...@gmail.com>**20091117100846
Ignore-this: 9cb9f9de76b6a84f800c6dc52e6ce337
]
[Remove deprecated signals, turn Window in to Maybe Window.
Axel Simon <axel.si...@ens.fr>**20091117085016]
[Add support for get/set_prgname and get/set_application_name.
John Millikin <jmilli...@gmail.com>**20091116194529
Ignore-this: 76c96d55ae04408764dca1c887af08b4
]
[Add a forgotten dependency of Cairo on array.
Axel Simon <axel.si...@ens.fr>**20091117082531]
[Update Graphics.UI.Gtk.Embedding modules to Gtk+ 2.18.3
Andy Stewart <lazycat.mana...@gmail.com>**20091116145642
Ignore-this: d560619a2b3f3d5cb64ef2e794097fcb
]
[Do not export functions that don't exist.
Axel Simon <axel.si...@ens.fr>**20091116093910]
[Add new types for PixbufAnimation and Co.
Axel Simon <axel.si...@ens.fr>**20091115203505]
[Bring AboutDialog up to date.
Axel Simon <axel.si...@ens.fr>**20091115201408]
[AddGdkPixbufAnimation
arse...@rpi.edu**20091114073747
Ignore-this: d5b7b2b309eff252143efb00af09863e
]
[Add functions to Window.
Axel Simon <axel.si...@ens.fr>**20091115193909]
[Swap the width and height parameters of the compositing function.
axel.si...@ens.fr**20091115092525]
[Add the TreeModelFilter module and a demo.
axel.si...@ens.fr**20091113231456]
[Add a name field to Attributes, whenever this is possible. Add a function to
set the attributes of CellRenderer using ColumnIds, thereby allowing exactly
the same column-based interface as Gtk does.
Axel Simon <axel.si...@ens.fr>**20091113231319]
[Add yet another property for read access.
axel.si...@ens.fr**20091112224538]
[Fix build problems of Cairo (dep on array package) and documentation.
axel.si...@ens.fr**20091112224113]
[Make compile with Gtk+ 2.10
Axel Simon <axel.si...@ens.fr>**20091112223802]
[Remove the two broken functions that were meant to interface Pixbuf and Cairo
surfaces. Add an array access function for image surfaces so that pixel values
can be copied manually.
axel.si...@ens.fr**20091101181431]
[Fix C-side memory leaks in functions that take a GList
Duncan Coutts <dun...@haskell.org>**20091031135849
Ignore-this: 8703764bb7a8314d9d7dd0ec5f42836e
The C functions do not take ownership of the GList (they make a copy)
It is the responsibility of the caller to allocate and free the GList.
]
[Add Visual and Device to the type list.
axel.si...@ens.fr**20091030165445]
[Remove internal signals in Notebook.
axel.si...@ens.fr**20091026184338]
[Complete the Screen module and add the Display module.
axel.si...@ens.fr**20091026184323]
[More fixed to get the new mk... functions to work.
axel.si...@ens.fr**20091026184145]
[Fix gtkglext to use the new mk... functions.
axel.si...@ens.fr**20091026175006]
[Fix more modules to use the new mk... functions.
axel.si...@ens.fr**20091026174825]
[Import GObject into the File module.
axel.si...@ens.fr**20091026164154]
[Fix various usages of the mk... functions.
axel.si...@ens.fr**20091026162928]
[Fix construction of SVG objects to use new typles of constructor and
destructor.
axel.si...@ens.fr**20091026162810]
[Change the way objects are destructed.
axel.si...@ens.fr**20091026162357]
[Fix the Gst hierarchy.
axel.si...@ens.fr**20091026155328
The gstreamer modules used a different hierarchy for their objects which meant
that they were not interchangeable with the GObject-rooted hierarchy. There
still exists a MiniHierarchy for the light-weight objects of gstreamer but they
are now created using the same utility in tools/hierarchyGen.
]
[Streamline the types generated for gstreamer by removing the adapted type
generator and generalizing the exiting one.
axel.si...@ens.fr**20090512213627]
[Add an option to c2hs to wrap each call in a locking operation.
axel.si...@ens.fr**20090512213527]
[Extend the type hierarchy generator to allow for re-exporting of modules and
to emit mk... functions that contain Haskell constructor and a C destructor
function.
axel.si...@ens.fr**20090510075921]
[Complete the Screen object.
Axel Simon <axel.si...@ens.fr>**20091022175931]
[Bring apiGen up to scratch.
Axel Simon <axel.si...@ens.fr>**20091022094936
This patch tries to fix the apiGen utility to at least run with the latest
Gtk versions. There are still some flaws in the Makefile which is mainly due
to me not understanding how it is all supposed to work. I have changed the
paths from which files are downloaded which might have been a bit over-
zealous. Furthermore, I do not understand how Gdk modules were supposed
to be generated as this should be a side-effect of generating the Gtk
modules. The patch also addresses some changes in the structure of the
documentation xml files.
]
[Add notes that how to function `windowSetTransientFor` with fullscreen-window.
Andy Stewart <lazycat.mana...@gmail.com>**20091012123621
Ignore-this: 3ba0f2b28b10b027088697639198dc98
]
[UncommentGValueCharFunctions
arse...@rpi.edu**20091010214130
Ignore-this: 40983acf515caedccf561ebb50bd927a
]
[Create Windows/Invisible.chs.pp (GtkInvisible) and fix docs of Notebook.chs.pp.
Andy Stewart <lazycat.mana...@gmail.com>**20091008005744
Ignore-this: 953ca96f67b460d57b083d3409af0587
]
[Update modules (Dialog, Label, Notebook) to Gtk+ 2.16 and fix docs in previous
patch.
Andy Stewart <lazycat.mana...@gmail.com>**20091007153405
Ignore-this: 9b7858f831226301fbc0493a8de4c820
]
[Update Multiline modules signals and Update Window/WindowGroup modules to Gtk+
2.16
Andy Stewart <lazycat.mana...@gmail.com>**20091006120659
Ignore-this: 31d7c20ba557bed65a92d6c9adb2d883
]
[Update all modules under Graphics.UI.Gtk.Multiline to Gtk+ 2.16.
Andy Stewart <lazycat.mana...@gmail.com>**20091001073008
Ignore-this: c9bd81bdac9f9c0f6d7ecad79dda5857
]
[Contain stub.o file in Makefile.am for generate Vte.chs.pp dynamic callback.
Andy Stewart <lazycat.mana...@gmail.com>**20090929121858
Ignore-this: d8db4a67742a1fe4dd965b38aaffc970
]
[Add signal connection function for VTE.
axel.si...@ens.fr**20090929070027]
[Add the dragFailed signal.
axel.si...@ens.fr**20090928193224]
[Fix documentation and some memory management stuff in VTE.
axel.si...@ens.fr**20090928193201]
[Adjust VTE binding and finish miss functions in GdkCursor
Andy Stewart <lazycat.mana...@gmail.com>**20090928013208
Ignore-this: 782feb53e623e3af75564e6533e0e0e0
* Adjust configure.ac make VTE binding 0.20.5 or higher
* Finish missing functions in Vte.chs.pp
** terminalGetTextIncludeTrailingSpaces
** terminalGetTextRange
** terminalMatchSetCursorType (this function need GdkCursorType that miss in
current Cursor.chs)
* Remove all deprecated functions in VTE 0.20.5
* Adjust export list of Vte.chs.pp
* Adjust comment of Vte.chs.pp, make it more Haskell style
* Finish miss functions of Cursor.chs
]
[Finish two more functions in VTE.
axel.si...@ens.fr**20090926135654]
[Binding for VTE 2.0 library
Andy Stewart <lazycat.mana...@gmail.com>**20090925003444
Ignore-this: 4d04c0c8aeddbc984358b55dd54843ec
]
[Add some more new signals.
axel.si...@ens.fr**20090924173347]
[Add the dragStatus function and more comments.
axel.si...@ens.fr**20090910211908]
[Make path conversion function safe as it may call back.
axel.si...@ens.fr**20090828073418]
[Update status in AUTHORS file to reflect reality
Duncan Coutts <dun...@haskell.org>**20090822005908
Ignore-this: e8d4122f02b53c63f8cf22acbd9826ce
]
[Jhbuild modules file for use on OSX
Hamish Mackenzie <ham...@firestream.co.uk>**20090719093602
Ignore-this: 19f10ea07de49ab9fae269c0a1fcd899
]
[Fix reference counting for GObjects.
axel.si...@ens.fr**20090720190441
This bug is only present in Glib libraries >2.10 which introduced floating
references. The problem is that even GObjects have always a floating
reference which has to be turned into a permanent reference. So far,
we omitted this call which had the effect that we would keep the object
around with a floating reference. Some objects in Gtk will call ref_sink
on an object that is passed to them, rather than ref. This causes the
floating reference to be turned into a proper one, rather than adding a new
reference. Thus, when the Haskell GC comes along, the object is unref'd and
destroyed while it's still being held in the Gtk object. This patch always
ref_sinks the object, resulting in a proper refence being held by Haskell
and forcing every ref_sink in Gtk code to have the semantics of ref.
]
[Add a comment to Notebook's setCurrentPage function.
Axel Simon <axel.si...@ens.fr>**20090611074238]
[Makefile.am: remove slashes after DESTDIR (fixes trac #1160)
pga...@gmail.com**20090510180614
Ignore-this: 5c57aa6375618d627e7568303a36abdd
]
[TAG 0.10.1
pga...@gmail.com**20090510160214]
[configure.ac: bump version to 0.10.1
pga...@gmail.com**20090429032110
Ignore-this: 682c3f529251a07b3de0761729af5eeb
]
[Bind GtkObject's destroy signal
m.ni...@gmail.com**20090430152124
Ignore-this: b67612a0e13a5cabc2c3392cb78fadaf
]
[Add missing export of Click(..) data type, repoted by Ben Franksen.
axel.si...@ens.fr**20090509184414]
[Input method bindings
m.ni...@gmail.com**20090504191302
Ignore-this: e8cbd55f0519ab413b24cae7b99d7b3d
]
[Improve apiGen names, for example hsv instead of hSV
m.ni...@gmail.com**20090430152250
Ignore-this: 89e23bb8809006e40a3708049de9c031
]
[gtk/Graphics/UI/Gtk/General/Structs.hsc: Accept responses 0 and smaller than
-11.
Marco Túlio Gontijo e Silva <mar...@holoscopio.com>**20090507111848
Ignore-this: f64b69b52a743ac54b404452178b29c0
]
[Add a function to insert the current clipboard content at the current cursor.
Axel Simon <axel.si...@ens.fr>**20090424152430]
[configure.ac: add bz2 dist tarball
pga...@gmail.com**20090429034820
Ignore-this: 5dfb2e2e91ab8b991d6908c7c08a51fe
]
[gtk: OwnerChange event only since gtk 2.6
pga...@gmail.com**20090429032134
Ignore-this: 23e3be727b8d43711321accaf35a4bf6
]
[Fix drawWindowGetPointer, trac #802.
axel.si...@ens.fr**20090427194638]
[Add an Eq instance for all objects.
axel.si...@ens.fr**20090426153152
This patch makes it possible to use == to compare two objects or, rather,
their pointers. Since for one case an object is merely containing data
and has its own Eq instance, it was necessary to add a 'noEq' flag to
the type generator that, when given, prevents the generation of the
deriving Eq clause. This patch fixes trac #1158.
]
[New version of carsim demo
mauricio.antu...@gmail.com**20090410234420
It has been a couple years since I wrote what is now the
'carsim' gtk2hs demo. I realized many of gtk+, gtk2hs and ghc
user library have been deprecated and replaced since then.
As it's in a demo directory, I thought it would be better to
have an updated version using current recomended API. Except
for a 'Pause' button, nothing important has changed in the user
interface. Changes in code include:
- Deprecated code, like System.Time and a few of gtk2hs, has
been replaced.
- Bug fix in about dialog.
- Use of EventM.
- Code is now UTF-8, but that only affects author name.
- Hopefully, better code and documentation.
]
[gtk: allow zero as argument to toResponse in G.U.G.General.Structs
pga...@gmail.com**20090424152136
Ignore-this: 922761ff9383f47f6efc030377241316
]
[Unregister packages in reverse order to avoid dependency conflict.
Axel Simon <axel.si...@ens.fr>**20090424144710]
[Makefile.am: add package dependencies to haddock command line
pga...@gmail.com**20090424022757
Ignore-this: 4bbe2fde77816e01e7b8be81863cb234
]
[Correct arguments of a finally statement.
Axel Simon <axel.si...@ens.fr>**20090417143511
This code has obviously never been tested. Thanks to Bertram Felgenhauer to
spot this.
]
[Adapt demo to use new Exception module.
Axel Simon <axel.si...@ens.fr>**20090417142437]
[Add 'clean-all' option to tools/apiGen/Makefile
mauricio.antu...@gmail.com**20090409024158
This adds a 'clean-all' option to apiGen's make
file. It will erase everything not on the original
distribution. This may be usefull if we are
testing different versions of packages, and need
to remove tarballs between tests.
]
[gio: Avoid calling (freeHaskellFunPtr nullFunPtr).
Bertram Felgenhauer <in...@gmx.de>**20090407084617]
[Make explicit use of concurrent finalizers.
Axel Simon <axel.si...@ens.fr>**20090327110603
This patch acounts for the change of semantics that finalizers undergo with
the release of GHC 6.10.1. Specifcially, finalizers are now by default
run during garbage collection, making it impossible to have them call back
to Haskell land. This patch uses functions that specifically schedule any
finilizers to be run as new threads once the mutator springs back to life.
]
[Add eventClick to EventM
Hamish Mackenzie <ham...@firestream.co.uk>**20090307080530
Ignore-this: ff7b605b58d6021248f9951b4c641209
]
[Fix documentation.
axel.si...@ens.fr**20090217180909]
[TAG 0.10.0
Peter Gavin <pga...@gmail.com>**20090207054815]
Patch bundle hash:
7d6e7ad862ec988d4b677f138244d70b7d30b021
------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
trial. Simplify your report design, integration and deployment - and focus on
what you do best, core application coding. Discover what's new with
Crystal Reports now. http://p.sf.net/sfu/bobj-july
_______________________________________________
Gtk2hs-devel mailing list
Gtk2hs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/gtk2hs-devel
