Menu
▾
▴
Re: [Gtk2hs-devel] darcs patch: Also derive Eq for Modifier
|
From: Axel S. <Axe...@en...> - 2008-01-28 19:50:54
|
Yes, sorry, applied!
A.
On Jan 28, 2008, at 13:13, David Leuschner wrote:
> Mon Jan 28 13:11:08 CET 2008 David Leuschner <da...@lo...>
> * Also derive Eq for Modifier
> As the modifier list on Linux platforms now also includes "Alt2"
> when
> "Control" is pressed it is convenient to be able to write
> "Control `elem` modifiers" instead.
>
> New patches:
>
> [Also derive Eq for Modifier
> David Leuschner <da...@lo...>**20080128121108
> As the modifier list on Linux platforms now also includes "Alt2" when
> "Control" is pressed it is convenient to be able to write
> "Control `elem` modifiers" instead.
> ] {
> hunk ./gtk/Graphics/UI/Gtk/Gdk/Enums.chs.pp 119
> - } deriving(Bounded,Show) #}
> + } deriving(Bounded,Show,Eq) #}
> }
>
> Context:
>
> [Use the new Modifer type in Events generated by c2hs.
> A....@ke...**20080123133742
> This patch removes the manually declared definition of Modifier in
> Events
> which was broken since it lacked the enumTo method of enum. Since
> c2hs
> is now fixed to always generate these methods, I've defined
> Modifier in
> Gdk/Enums and merely import it. Before the fix to c2hs, c2hs would
> only
> generated the enumTo instance if the data type was deriving Eq since
> the previous implementation did an equality test. This complicated
> behaviour
> of c2hs probably led me back then to implement Modifier by hand in
> Events since
> it wouldn't work and I couldn't understand why.
>
> ]
> [Add Show instances to Enums.
> A....@ke...**20080123125608]
> [Revert the changes to the toFlags function since c2hs is now fixed.
> A....@ke...**20080123125159]
> [Had the wrong type in mind when I did the last patch.
> A....@ke...**20080122105341]
> [Fix the type that is read by widgetGetState.
> A....@ke...**20080122104814]
> [Fix Enum instance generation so enumFrom works.
> A....@ke...**20080120231618
> This patch changes the generation of enumFrom functions in Enum
> instances
> produced by c2hs. Before, the enumFrom functions would only be
> generated if
> Eq was in the deriving clause of the hook. However, there is a way to
> define the enumFrom function even without an Eq instance, which is
> what
> the patch implements.
> ]
> [Add succ and pred definitions to Modifier and fix toFlags function.
> A....@ke...**20080119201849]
> [Tutorial-Port: Correction of typing errors.
> hth...@zo...**20080102141654
> Eduardo Basterrechea has noted the typing errors he found
> during his translation to Spanish of the Gtk2Hs tutorial
> chapters 1 - 7. These and a few more are corrected here.
> ]
> [Tutorial-Port Spanish Chapters 5, 6 and 7
> hth...@zo...**20080101165758
> (also corrected a previous typo)
> ]
> [Tutorial-Port Spanish Translation Chapter 4 (1-7)
> hth...@zo...**20071215174442]
> [Fixed Enum class for Modifier and toModifier
> David Leuschner <da...@lo...>**20080118172605]
> [Do not show this internal module in the docs.
> A....@ke...**20080103210149]
> [Add a demo on the TreeModelSort proxy.
> A....@ke...**20080103192429]
> [Add the TreeSortable interface.
> A....@ke...**20080103190026
>
> This interface allows a view to sort the underlying model. Since our
> CustomStore does not implement this interface, the only object that
> actually does implement it is the TreeModelSort proxy which takes
> a child
> model and sorts its rows on-the-fly.
> ]
> [Update the TreeModelSort proxy to work with the new Haskell-land
> stores.
> A....@ke...**20080103185641
>
> This patch makes the TreeModelSort proxy fully functional.
> TreeModelSort
> encapsulates any model and provides a sorted view of it.
>
> ]
> [Add some bogus function that can be used to implement the sortable
> interface in our CustomStore.
> A....@ke...**20080103185545]
> [Allow a more general function to set CellRenderer attributes.
> A....@ke...**20080103185021
>
> The function cellLayoutSetAttributes was rather specific in that
> it only
> allowed setting attributes of the given CellRenderer from the
> given model.
> Allow a more general variant which is necessary when using proxy
> models
> like TreeModelSort or TreeModelFilter.
> ]
> [Add Show instances to enumerations.
> A....@ke...**20080103182105]
> [Make treeModelSortNewWithModel safe.
> A....@ke...**20071213142856]
> [Correct CellLayout to compiler with Gtk 2.4.
> A....@ke...**20071210141435]
> [Fix PixbufData's MArray instance for ghc-6.8
> Duncan Coutts <du...@ha...>**20071210033636
> The MArray class gained a getNumElements method in ghc 6.8
> Also take the opportunity to tidy things up a bit.
> ]
> [Forgot to add the file with helper functions for tree DND.
> A....@ke...**20071209081148]
> [Tutorial-Port-Spanish Chapters 1,2,3 and index
> hth...@zo...**20071208154317]
> [Export DialogFlags which is used as an arg of several
> MessageDialog functions
> Duncan Coutts <du...@ha...>**20071127021150]
> [Add DND functions to IconView.
> A....@ke...**20071204213515]
> [Docu fixes.
> A....@ke...**20071204213421]
> [Fix stamp problems in tree implementation.
> A....@ke...**20071204212616
>
> The story of TreeIter stamps is as follows: ListStore doesn't use
> them since
> the iterators remain valid throughout. TreeStore needs to
> invalidate the
> iterators each time a node is inserted since a new node might mean
> that it
> can't be indexed anymore with the available number of bits at that
> level.
> The function that inserts nodes correctly invalidated the tree
> iterators.
> However, this function also calls methods in TreeModel, passing in
> order
> to notify itself and the view. The passed in TreeIters have no
> stamp set,
> thus a warning is raised as soon as these tree iters get back to
> the model,
> usually in a form of ref'ing the node. This was slightly confusing as
> setting the stamp was so far automatically done in the C interface
> Gtk2HsStore.c. The stamp is now set explicity in the TreeIters
> before they
> are passed to the functions in the model.
>
> ]
> [Tutorial-Port Introduction (Chapter 1)
> hth...@zo...**20071130175603]
> [Tutorial-Port Popup, Radio and Toggle Actions (Chapter7.2)
> hth...@zo...**20071128165052]
> [Update TreeView and TreeViewColumn with new functions.
> A....@ke...**20071202213657]
> [Complete the default drag and drop interface for ListStore.
> A....@ke...**20071202211958
>
> This patch completes the former patch which adds the C functions
> to the C custom model. With this patch it is possible to reorder
> the rows of a list store using drag and drop. The drag and drop for
> tree stores doesn't quite work yet since I'm not sure what the
> exact semantics are (do we move a row at a time, or a full tree? is
> the last index of the drop path the position where we insert the
> new row?). This patch also enables the reordering in the two
> listtest and treetest demo programs.
> ]
> [Add drag and drop interfaces to the CustomStore.
> A....@ke...**20070720123434
>
> Both, ListStore and TreeStore should implement the drag and drop
> interfaces because the Gtk ListStore and TreeStore do. However, I'm
> not sure yet what they exactly implement; I assume they only accept
> rows from the same store. Figure out how to implement this in
> Haskell and provide a way to add the possiblity of accepting other
> dnd sources and destinations as well.
> ]
> [Tutorial-Port Menus and Toolbars (chapter 7.1)
> hth...@zo...**20071124130324]
> [Tutorial Port Paned Windows and Aspect Frames (Chapt.4)
> hth...@zo...**20071119134955]
> [Add to General Gdk: screen size querying, pointer grabbing,
> keyboard grabbing
> Bit Connor <bi...@mu...>**20071118123425]
> [Add Gdk Cursor module
> Bit Connor <bi...@mu...>**20071118123231]
> [Add Gdk Event related 'currentTime' function (binding to
> GDK_CURRENT_TIME)
> Bit Connor <bi...@mu...>**20071118122630]
> [Tutorial_Port: The Layout Container (Chapter 6.3)
> hth...@zo...**20071109152940]
> [gstreamer: M.S.G.Bus: improve API, add documentation
> Peter Gavin <pg...@gm...>**20071110052545]
> [glib: S.G.MainLoop: remove g_ prefix in some {#call#} tags for
> consistency (and to fix build error)
> Peter Gavin <pg...@gm...>**20071110042515]
> [glib: change Word to HandlerId in a few places
> Peter Gavin <pg...@gm...>**20071110042128]
> [glib: S.G.MainLoop: new API: sourceRemove
> Peter Gavin <pg...@gm...>**20071110041737]
> [glib: S.G.MainLoop: new API: mainContextFindSourceById,
> sourceDestroy, sourceIsDestroyed
> Peter Gavin <pg...@gm...>**20071110040450]
> [gstreamer: M.S.G.Core.Types code cleanups
> Peter Gavin <pg...@gm...>**20071109001145]
> [gnomevfs: use c2hs {#enum#} for FilePermissions instead of hand
> coding it
> Peter Gavin <pg...@gm...>**20071108034949]
> [gstreamer: M.S.G.Core.Buffer: add documentation, API improvements
> Peter Gavin <pg...@gm...>**20071103135504]
> [gstreamer: document M.S.G.Core.Constants; add version checks,
> remove MessageStateDirty (deprecated)
> Peter Gavin <pg...@gm...>**20071102003142]
> [gstreamer: add docs & fix formatting in M/S/G/Core/Bin.chs; rename
> to Bin.chs.pp
> Peter Gavin <pg...@gm...>**20071031200139]
> [gstreamer: move M.S.G.Core.Buffer.bufferOffsetNone to
> M.S.G.Core.Constants and use #{const ...}
> Peter Gavin <pg...@gm...>**20071030001137]
> [configure.ac: add GSTREAMER_CHECK_VERSION macro
> Peter Gavin <pg...@gm...>**20071029222529]
> [Makefile.am: filter CFLAGS/LDFLAGS for gstreamer & gnomevfs
> Peter Gavin <pg...@gm...>**20071029222356]
> [gstreamer: M.S.G.Core.Buffer/M.S.G.Base.Adapter: change ghc
> version check to OLD_BYTESTRING macro
> Peter Gavin <pg...@gm...>**20071029202836]
> [gstreamer: remove M.S.G.Core.Buffer.bufferWithDataM
> Peter Gavin <pg...@gm...>**20071029160749
> I need to rethink this one
> ]
> [remove old ghc-6.2 package config files
> Duncan Coutts <du...@ha...>**20071108220833]
> [Drop support for ghc < 6.4 and clean up
> Duncan Coutts <du...@ha...>**20071108212328
> There was quite a bit of configuritis needed to support ghc-6.0
> and 6.2
> ]
> [Remove the variant of toFlags.
> A....@ke...**20071108085631]
> [glib: replace toFlags with toFlags'
> Peter Gavin <pg...@gm...>**20071108001032]
> [c2hs: added definitions for succ, pred, enumFrom{,To,ThenTo}; also
> add support for bitwise and, or, xor, and complement
> Peter Gavin <pg...@gm...>**20071108000451]
> [Replace toFlags with Peter Galvin's version.
> A....@ke...**20071105155303]
> [glib: add toFlags': like toFlags but works doesn't fail when we
> haven't defined toEnum for a bit
> Peter Gavin <pg...@gm...>**20071105150631
> This is needed for a few reasons, the most important of which is
> for forward compatibility.
> If a newer version of the upstream package defines a new flag for
> a flag type, we shouldn't
> fail just because we don't recognize it. This version allows
> overlapping flags, for example, given:
>
> data OpenFlags = Read | Write | ReadWrite
> instance Flags OpenFlags
> instance Enum OpenFlags where ...
>
> If ReadWrite is equivalent to [ Read, Write ] in C land, then it
> will be in Haskell land as well.
>
> ]
> [Fix ball soe demo to work with newer soe api
> Duncan Coutts <du...@ha...>**20071107171349]
> [Add statusicon demo to tarball
> Duncan Coutts <du...@ha...>**20071107170756]
> [Use precise package versions when building and in package
> registration
> Duncan Coutts <du...@ha...>**20071107133455
> Previously we were being sloppy and allowing any version in the
> registration
> so there was no guarantee that the version we built against was
> the same as
> the version that would get used when the package was installed.
> Similarly
> we could build against packages that were registered per-user
> which would
> then fail if we did a global install. Now it only looks for global
> packages
> unless you ./configure --with-user-pkgconf
> This version of the patch for gtk2hs HEAD also covers gstreamer
> and gnomevfs
> ]
> [Follow renaming of Image.chs -> Image.chs.pp in Makefile
> Duncan Coutts <du...@ha...>**20071107115759]
> [compile clipboard functions conditionaly when gtk >= 2.2
> Duncan Coutts <du...@ha...>**20071105004442
> as the Clipboard type was only introduced in gtk-2.2
> ]
> [fix build with ghc 6.8.1
> Bertram Felgenhauer <in...@gm...>**20071103212034]
> [Make Adam's cairo patch build with cairo-1.0
> Duncan Coutts <du...@ha...>**20071102234514]
> [Add support for getting the stride and image data of an image surface
> ag...@im...**20071030183825
>
> This wraps the cairo functions:
> cairo_image_surface_get_stride
> cairo_image_surface_get_data
>
> Note that these functions were introduced in cairo 1.2, and so
> this sets a new
> high bar for gtk2hs cairo.
>
> ]
> [SVG: provide implementation for GObjectClass
> Bertram Felgenhauer <in...@gm...>**20071031135016]
> [GLDrawingArea: provide implementation for GObjectClass
> Bertram Felgenhauer <in...@gm...>**20071031134717
> Without this patch, trying to use a GLDrawingArea results in
> program: gtkglext/Graphics/UI/Gtk/OpenGL/DrawingArea.chs:73:0:
> No instance nor default method for class operation
> System.Glib.Types.toGObject
> ]
> [Remove duplicate item from EXTRA_DIST
> Duncan Coutts <du...@ha...>**20071030085915]
> [Tutorial Port Event and Button Boxes (Chapter 6.2)
> hth...@zo...**20071029173243
>
> Corrected a previous patch error in index.xhtml
> where .xhtml had been changed into .html.
>
> Corrected the chapter on font and color selection,
> which came after Alex Tarkovsky's patch, for un-
> necessary newlines after <pre>
>
> Corrected a few other missing or incorrect links in previous files
> ]
> [Makefile.am: in uninstall, unregister from user pkg conf when
> USERPKGCONF is set
> Peter Gavin <pg...@gm...>**20071021192610]
> [gstreamer: change M/S/G/DataProtocol/Constants.chs to
> Constants.hsc in makefile
> Peter Gavin <pg...@gm...>**20071028214736]
> [gstreamer: add M.S.G.Audio.AudioClock
> Peter Gavin <pg...@gm...>**20071025153404]
> [gstreamer: add a format argument to elementQueryDuration/Position
> Peter Gavin <pg...@gm...>**20071021210213
> this goes with the previous patch.
> also changes type of queryNewConvert to use Word64 rather than Int64.
> ]
> [gstreamer: change Int64 to Word64 in return types of
> elementQueryPosition/Duration
> Peter Gavin <pg...@gm...>**20071021210138]
> [gstreamer: improve vorbis-play demo
> Peter Gavin <pg...@gm...>**20071021205843]
> [gstreamer: add vorbis-play demo
> Peter Gavin <pg...@gm...>**20071021192836]
> [gstreamer: add mtl to packages dependencies
> Peter Gavin <pg...@gm...>**20071021192814]
> [gstreamer: add MessageAsyncStart & MessageAsyncDone to MessageType
> enum
> Peter Gavin <pg...@gm...>**20071021192726]
> [gstreamer: include C objects in profile lib
> Peter Gavin <pg...@gm...>**20071021192523]
> [gstreamer: export nsecond in M.S.G.Core.Constants
> Peter Gavin <pg...@gm...>**20071021180636]
> [gstreamer: make Adapter and Buffer build with ghc 6.8
> Peter Gavin <pg...@gm...>**20071021180323
>
> the problems actually come from the changes in bytestring. not
> sure if I should specifically check for that instead.
> ]
> [gstreamer: hopefully fix takeObject & peekObject for real this time
> Peter Gavin <pg...@gm...>**20071020195000
>
> takeObject: to be used when a function returns an object that must
> be unreffed at GC.
> If the object has a floating reference, the float flag
> is removed.
>
> peekObject: to be used when an object must not be unreffed. A ref
> is added, and is
> removed at GC. The floating flag is not touched.
> ]
> [gstreamer: fix floating reference mess
> Peter Gavin <pg...@gm...>**20071009023334
>
> I figure, the safest and easiest thing to do is just remove the
> floating flag
> from any GstObject we see, without touching the refcount (except
> maybe to add one).
> ]
> [Fix mistake over #ifdev vs #if
> Duncan Coutts <du...@ha...>**20071029132150]
> [Fix compiling with gtk+-2.6
> Duncan Coutts <du...@ha...>**20071029132112]
> [gstreamer: lots of new code
> Peter Gavin <pg...@gm...>**20070906212129]
> [gstreamer: add new modules BaseSink, BaseTransform, PushSrc,
> Adapter; redid comment headers in source files
> Peter Gavin <pg...@gm...>**20070805140633]
> [gstreamer: update heirarchy stuff to match Duncan's recent changes
> Peter Gavin <pg...@gm...>**20070804160945]
> [gstreamer: better flag support for GstObject subtypes; replaced
> from* and to* for enums with cToEnum/cFromEnum; similarly for flags
> Peter Gavin <pg...@gm...>**20070726202800]
> [Make the StatusIcon stuff compile with gtk < 2.10
> Duncan Coutts <du...@ha...>**20071029113538]
> [Make GDateTime compile with old versions of glib
> Duncan Coutts <du...@ha...>**20071026124216]
> [gnomevfs: some documentation cleanups
> Peter Gavin <pg...@gm...>**20071009023250]
> [gstreamer: add bytestring dependency in Makefile.am if
> HAVE_SPLIT_BASE is set
> Peter Gavin <pg...@gm...>**20071020184923]
> [gtk: some source documentation cleanups
> Peter Gavin <pg...@gm...>**20071009023028]
> [gnomevfs: add function S.G.V.Volume.volumeUnmount
> Peter Gavin <pg...@gm...>**20070929170420]
> [gnomevfs: hide Marshal and Types from haddock
> Peter Gavin <pg...@gm...>**20070906205746]
> [make hierarchyGenGst not use splitobjs
> Peter Gavin <pg...@gm...>**20070805144738]
> [Move all deprecated exports to their own section in the export list
> Duncan Coutts <du...@ha...>**20071011130754]
> [Use qualified names for new signals
> Duncan Coutts <du...@ha...>**20071011130612]
> [Use the correct name for old signals
> Duncan Coutts <du...@ha...>**20071011130540]
> [Use the proper deprecated doc comment rather than a generic one.
> Duncan Coutts <du...@ha...>**20071011125201]
> [Deprecate getter/setter methods that duplicate attributes
> Duncan Coutts <du...@ha...>**20071011124907
> Add deprecated notes.
> Also combine docs for attributes synthesised out of getter/setter
> methods.
> ]
> [Disable generation hashes for the moment
> Duncan Coutts <du...@ha...>**20071011124525
> They don't quite work reliably yet for some reason.
> ]
> [Tutorial Port Scrolled Window (Chapter 6.1)
> hth...@zo...**20071020161937]
> [Tutorial Port Notebook (Chapter 5.4)
> hth...@zo...**20071014153504]
> [Tutorial_Port Font and File Selection (Chapter 5.3)
> hth...@zo...**20071009172147]
> [Add a comment on the return values of events.
> A....@ke...**20071015105656]
> [Gtk2Hs Tutorial: Fix vertical alignment in <pre> sections by
> removing unnecessary leading \n chars
> Alex Tarkovsky <ale...@gm...>**20071004171957]
> [binding textbuffer clipboard functions (v2)
> jn...@ar...**20071010172832
> Cut, copy and paste functions for the text buffer and a minimal
> clipboard
> module that just provides a type and a function to get the
> standard clipboards.
> ]
> [Add a catch-all constructor to the set of mouse buttons.
> A....@ke...**20071004144115]
> [Tutorial Port File Selection (Chapter 5-2)
> hth...@zo...**20070928162552]
> [Patch Chapter 5.1 into the ToC and header/footer navigation
> Alex Tarkovsky <ale...@gm...>**20070921021310]
> [Change Source File Names, Change Window Titles and Screenshots
> hth...@zo...**20070920132032
> Source file Names now match XHTMl chapters,
> Added and changed some window titles and screen shots for consistency
> Removed references to chapters from source code files, except in
> the very first (GtkChap2a.hs)
>
> ]
> [Tutorial Port Calendar (chapter 13)
> hth...@zo...**20070916161414
> Changed to XHTML. Links need to be added and
> there are 2 unordered lists with no <div> and id yet
> ]
> [Gtk2Hs Tutorial: XHTML-ize remaining files through chapter12;
> rename files according to ToC
> Alex Tarkovsky <ale...@gm...>**20070916192047]
> [Make gtk2hs build with ghc 6.8
> Bertram Felgenhauer <in...@gm...>**20070913160529
> - Add additional dependencies induced by the base package split.
> - fix one API breakage.
> ]
> [GTK Tutorial Port Spin Boxes (Chapter 12)
> hth...@zo...**20070910131354
> Alex Tarkovsky will implement a change from HTML 4 Provisional to
> XHTML strict
> and a corresponding change in file names and extensions. New
> additions by me
> will match that, but until Alex's patches have been applied I'll
> carry on as before.
> However, this means that links will be broken.
> This chapter 12 patch is standalone, for the moment.
> ]
> [XHTML-ize and CSS-ify ToC and chapters 3 and 4; renumber ToC
> Alex Tarkovsky <ale...@gm...>**20070906210332]
> [Remove broken links from Glade tutorial
> Alex Tarkovsky <ale...@gm...>**20070904135020]
> [Fix gconf_MOSTLYCLEANFILES typo
> Duncan Coutts <du...@ha...>**20070903173646]
> [Tutorial_Port Text Entries and Statusbars (Chapter 11) + links update
> hth...@zo...**20070831132507]
> [Dialogs StockItems Progress Bars (chapter 10)
> hth...@zo...**20070831105357
> Chapter 10 plus crrected links in Chapter 1 and Chapter 8
> ]
> [Gtk2Hs tutorial: chapter 3 updates
> Alex Tarkovsky <ale...@gm...>**20070901094853]
> [Arrows and Tooltips Chapter 9
> hth...@zo...**20070830202333]
> [Undo accidentally-comitted debugging change to soe
> Duncan Coutts <du...@ha...>**20070830184107]
> [Remove a bit of soe debug cruft.
> Duncan Coutts <du...@ha...>**20070830131622]
> [Update Graphics.SOE.Gtk to new api and fix many bugs
> Duncan Coutts <du...@ha...>**20070830124427]
> [Add missing _stub to list in Makefile.am
> Duncan Coutts <du...@ha...>**20070830124410]
> [Fill in the GObjectClass methods for the ListStore and TreeStore
> Duncan Coutts <du...@ha...>**20070829020153
> We still seem to be getting some iter stamp errors :-(
> ]
> [Update Glade tutorial to focus on Glade 3
> Alex Tarkovsky <ale...@gm...>**20070820151705]
> [Gtk+2.0 Tutorial Port Chapter 8, Tables
> hth...@zo...**20070817141356]
> [Add _SPLIT variant for gstreamer to the CLEAN targets.
> A....@ke...**20070817124958]
> [gstreamer: use the new hierarchyGenGst; rename Object marshallers
> to better denote their semantics; start wrapping base libraries
> Peter Gavin <pg...@gm...>**20070723223441]
> [tools: added hierarchyGenGst, a modified version of
> hierarchyGenGst, needed for GStreamer
> Peter Gavin <pg...@gm...>**20070723223020]
> [gstreamer: Core/Types.chs: added isObject (etc.) functions
> Peter Gavin <pg...@gm...>**20070722015917
> Also renamed newObject/newMiniObject to mkNewObject/
> mkNewMiniObject & added real newObject/newMiniObject functions.
> ]
> [gstreamer: add attributes & properties to GObject subclasses;
> small code cleanups; add hierarchy for GStreamer.Base
> Peter Gavin <pg...@gm...>**20070722004623]
> [gstreamer: change ClockTime/ClockTimeDiff to Word64/Int64 instead
> of c2hs types
> Peter Gavin <pg...@gm...>**20070721214252]
> [gstreamer: move all current modules from Media.Streaming.GStreamer
> to Media.Streaming.GStreamer.Core
> Peter Gavin <pg...@gm...>**20070721185839
> I'm anticipating adding other submodules into GStreamer.
> ]
> [gstreamer: remove commented function messageParseAsyncStart in
> Message.chs
> Peter Gavin <pg...@gm...>**20070720201214]
> [gstreamer: fix types for event handlers in Element.chs
> Peter Gavin <pg...@gm...>**20070720201151]
> [gstreamer: fix export list in most modules
> Peter Gavin <pg...@gm...>**20070720201023]
> [gstreamer: add stub files to LIBADD in makefile
> Peter Gavin <pg...@gm...>**20070720200932]
> [gstreamer: initial import
> Peter Gavin <pg...@gm...>**20070719060228]
> [Add a demo on PangoLayout.
> A....@ke...**20070817112626]
> [Fix ref counting bug in sourceBufferCreateMarker
> Duncan Coutts <du...@ha...>**20070814182814
> (reported by bwwx in #haskell)
> The docs say that gtk_source_buffer_create_marker returns a new
> GtkSourceMarker, owned by the buffer. So we need to ref the
> object, so we
> should use makeGObject rather than constructNewGObject.
> ]
> [Gtk+2.0 Tutorial Port (part 1)
> hth...@zo...**20070811141048
>
> Start of a port to Gtk2Hs of the Gtk+2.0 tutorial
> included in the Gtk documentation.
>
> Note: Chapter 2 is supposed to be an introductory overview,
> as in the original, but for Gtk2Hs. I'm hoping someone
> who knows more will jump in here.
>
> The implemented chapters are:
> 1. Contents and copyright, as required by the original
> authors
> 2. To be written by an expert...
> 3. Getting Started
> 4. Packing Widgets
> 5.a. Packing Demonstration Program
> 5.b. Packing Using Tables
> 6. The Button Widget
> 7. Adjustments, Scale and Range
>
> The examples have been tested with Gtk2Hs-0.9.12
> (the first ones with .11) on Fedora 6.
>
> The html files have been taken from the original Gtk+2.0 tutorial,
> processed in Open Office Writer html mode, and then cleaned up
> with Screem and tidy. There are still some unreferred name tags left
> from the original.
>
> ]
> [small makefile changes to clean up documentation files that were
> not being removed
> Peter Gavin <pg...@gm...>**20070805144521]
> [gnomevfs: updated license to LGPLv3, made all source file headers
> consistent
> Peter Gavin <pg...@gm...>**20070722024618]
> [glib: add support for 64-bit int object properties
> Peter Gavin <pg...@gm...>**20070721212009]
> [glib: export contructors for Source datatype in MainLoop.chs
> Peter Gavin <pg...@gm...>**20070719055702]
> [Import CUInt in the Types.chs modules as GType seems to be CUInt now
> Duncan Coutts <du...@ha...>**20070808171654
> Reported in Fedora 8 that c2hs reports GType as CUInt whereas it
> previously
> was CULong, so we now import both in the Hierarchy.chs.template.
> ]
> [Fix doc typo, "the the"
> Duncan Coutts <du...@ha...>**20070802031049]
> [Merge changes in cleaning split-objs with gnomefvs changes
> Duncan Coutts <du...@ha...>**20070728192403]
> [Be more careful about cleaning split-objs and stub files
> Duncan Coutts <du...@ha...>**20070728173305
> And build the code gen tools without using split-objs at all.
> ]
> [Add clean script to win32 scripts bundle
> Duncan Coutts <du...@ha...>**20070728163800]
> [Update win32 scripts to gtk2hs-0.9.12 + gtk+-2.12.14
> Duncan Coutts <du...@ha...>**20070727132759
> And add checkbox to override dll problem to gtk2hs installer script
> ]
> [Follow change in name of local cairo header file
> Duncan Coutts <du...@ha...>**20070727131745]
> [Document the 'on' function
> Duncan Coutts <du...@ha...>**20070727155728]
> [Change the class heirarch implementation to avoid coerce
> Duncan Coutts <du...@ha...>**20070727155053
> GObjectClass now has two real mothods, toGObject and
> unsafeCastGObject
> (renamed from fromGObject to better reflect the fact that it performs
> unchecked down-casts). Sub classes of GObjectClass do not have any
> additional methods, so we only carry one dictionary per object.
> So it should now be possible to make custom widgets in Haskell and
> make
> them instances of the existing widget classes etc. So this should
> give
> greater consistenncy and better data hiding.
> ]
> [Fix comboBoxGetModel, it previously did an unchecked downcast
> Duncan Coutts <du...@ha...>**20070727135408]
> [Add authors name to author's file.
> A....@ke...**20070726130748]
> [Add stock items to reference documentation.
> A....@ke...**20070726130215]
> [Make treeList demos build again
> Duncan Coutts <du...@ha...>**20070725145318]
> [Add Bulat Ziganshin to AUTHORS for dirlist contributions
> Duncan Coutts <du...@ha...>**20070725145245]
> [Add HAVE_GTK_VERSION_2_12 makefile conditional
> Duncan Coutts <du...@ha...>**20070725145202]
> [Change gtkmozembed provider preference to put xulrunner first
> Duncan Coutts <du...@ha...>**20070725144951
> Apparently xulrunner is now in vogue again so put it first
> So ther order of prefernece has changed from/to:
> -seamonkey, firefox, xulrunner, mozilla
> +xulrunner, seamonkey, firefox, mozilla
> ]
> [Build the statusicon demo
> Duncan Coutts <du...@ha...>**20070725144924]
> [StatusIcon demo
> Andrea Vezzosi <san...@gm...>**20070725094936]
> [Add GtkStatusIcon bindings
> Andrea Vezzosi <san...@gm...>**20070725033433]
> [Must not use _static version of pango_font_description_set_family
> Duncan Coutts <du...@ha...>**20070723145641
> Since the string buffer is only kept around temporarily, not
> forever like
> pango_font_description_set_family_static requires.
> ]
> [Fix bug in TreeStore on empty stores.
> A....@ke...**20070716123708
> This fixes a bug when running e.g. treeModelIterNChildren store
> Nothing when
> store is an empty TreeStore.
> ]
> [Rename the local cairo.h to something less ambiguous.
> A....@ke...**20070716123509
> I can't recall where this became a problem, but I guess it can't
> hurt to have this name distinct from the real header file.
> ]
> [glib: import GDateTime instead of GDate
> Peter Gavin <pg...@gm...>**20070716104909
>
> Brainfart.
>
> ]
> [glib: add GSource support
> Peter Gavin <pg...@gm...>**20070713205416]
> [glib: add GDate to exports in Glib.hs
> Peter Gavin <pg...@gm...>**20070713175625]
> [glib: make module name for GDateTime the same as the filename :)
> Peter Gavin <pg...@gm...>**20070712013200]
> [glib: added support for GDate/GTimeVal
> Peter Gavin <pg...@gm...>**20070711035813]
> [Make the new ModelView interface self contained.
> A....@ke...**20070716102722
> This patch moves a few model functions from CustomStore to
> TreeModel. Also,
> it duplicates the TreeIter and TreePath definitions by copying
> them into the
> ModelView.Types file. The TreeModel module is now complete.
> However, another
> step would be to re-export all functions from CustomStore via
> TreeModel and
> hide CustomStore. The user would then be able to define a new
> store by just
> importing TreeModel.
> ]
> [Add missing case in Drag. Fix documentation.
> A....@ke...**20070715163552]
> [Add a new function to Pixbuf. Correct documentation.
> A....@ke...**20070715163512]
> [Add columns to our Haskell tree model.
> A....@ke...**20070715162232
>
> This patch adds the ability to lookup values from TreeStore and
> ListStore in terms of columns. Some information in e.g. ComboBox
> have properties that are not shown by CellRenderers directly. These
> properties can therefore not be set by using function in CellLayout
> and the only way to connect them to the model is by pretending they
> access a certain column in the model. Hence this patch adds the
> ability to access a Haskell model using a column number. The idea
> is that these column numbers are opaque to the user of Gtk2Hs.
> Functions that use columns are called widgetSetBlahSource where
> wiget is the widget and Blah is the property.
> ]
> [gnomevfs: fix demos to work again, new demo TestVolumeMonitor
> pg...@gm...**20070707212333]
> [gnomevfs: use actual Volume type instead of VolumeClass volume
> => ... for signal handler argument type
> pg...@gm...**20070707212228]
> [gnomevfs: more documentation, type fixes, etc. in Volume/Drive/
> VolumeMonitor
> pg...@gm...**20070706195920]
> [gnomevfs: small fixes in cabal/package.conf files
> pg...@gm...**20070706032048]
> [glib: add support for GMainLoop/GMainContext
> Peter Gavin <pg...@gm...>**20070707210848]
> [Use a local .h file that includes both gnomefvs headers
> Duncan Coutts <du...@ha...>**20070705191142
> So we don't have to locally include extra headers for
> particular .chs files.
> This is the same system we use for sourceview, cairo and svgcairo.
> ]
> [gnomevfs: initial import
> pg...@gm...**20070705094200]
> [Use .NOTPARALLEL make directive to prevent parallel builds
> Duncan Coutts <du...@ha...>**20070705164541
> So we don't have to give users special instructions not to use -jN
> ]
> [Note that parallel make does not work.
> Duncan Coutts <du...@ha...>*-20070705144700]
> [Update DirList demo to work on Windows and include file
> modification times
> Duncan Coutts <du...@ha...>**20070705155824
> Thanks to Bulat Ziganshin for the patch.
> I also updated it to use multiple columns with titles and to avoid
> the use
> of unsafePerformIO by using the :=> monadic attribute assignment
> operator.
> ]
> [Note that parallel make does not work.
> Duncan Coutts <du...@ha...>**20070705144700]
> [Really add the SVG backend, forgot to add the main file.
> Duncan Coutts <du...@ha...>**20070705012450
> I'm clearly going mad. The file got included in the tarball,
> so I didn't notice it's absence from the darcs repo.
> ]
> [Fix typo in win32 build script
> Duncan Coutts <du...@ha...>**20070704040641]
> [Update win32 build scripts to build sourceview but not svgcairo
> Duncan Coutts <du...@ha...>**20070704035213]
> [Update a demo to follow change in api for pixbufNewFromFile
> Duncan Coutts <du...@ha...>**20070704032750]
> [Add Cairo SVG backend
> Duncan Coutts <du...@ha...>**20070704032635
> I swear I added this before, the cairo demo certainly already
> tests it.
> Somehow I must have managed to not record it, and then loose the
> code.
> Fortunately it was only a few lines, so quick to replace. Silly me.
> ]
> [Reimplement the way we keep track of gthread initialisation state
> Duncan Coutts <du...@ha...>**20070704012435
> Sadly we cannot just call g_thread_supported() because this macro
> expands
> into a reference to a gloabl variable that lives in the gthread
> dynamic lib.
> On windows, GHCi's dynamic linker cannot cope with references to
> global vars
> from C dlls. So instead we keep a global var in the wrapper, so
> it's in the
> same package, rather than crossing a dll boundary. This allows us
> to track
> the state, and because the state is allocated in C land, it
> survives :reload
> in GHCi, which was the problem with the original scheme. Ugg.
> ]
> [Fix for the sed extra-ghci-libs hackery: some libs have '_' in
> their names
> Duncan Coutts <du...@ha...>**20070704005240]
> [Updated windows installer script
> Duncan Coutts <du...@ha...>**20070703203650
> updated for latest Gtk2Hs version
> use a build for only a single ghc version (ghc 6.6.1)
> we previously included builds for both ghc 6.4.2 and 6.6
> include the sourceview component
> ]
> [Update Win32 build scripts
> Duncan Coutts <du...@ha...>**20070703203305
> Both the scripts for building Gtk2Hs on Windows and the scripts
> used to
> build a Gtk+ SDK for Windows. Now using Gtk+ 2.10.13 for the
> Windows builds.
> ]
> [Add a few people who contributed to this release to the AUTHORS file
> Duncan Coutts <du...@ha...>**20070703202905]
> [sed hackery to generate extra-ghci-libraries in package files on
> win32
> Duncan Coutts <du...@ha...>**20070703201811
> The .lib names and .dll names on windows do not match up, so we
> have to
> specify different libraries for ghci than for ghc. eg for gtk-2.0.lib
> the corresponding dll is libgtk-2.0-0.dll. We do this conversion
> using
> a sed rule plus a short list of hard-coded exceptions.
> ]
> [Unbreak SOE in the single threaded case again.
> Duncan Coutts <du...@ha...>**20070703201648
> This is getting embarassing.
> ]
> [Ugg, unbreak SOE in the non-threaded rts case.
> Duncan Coutts <du...@ha...>**20070630162214]
> [Make the check for the GThread system being initialised more robust
> Duncan Coutts <du...@ha...>**20070630000326
> We previously used an evil top level IORef however in GHCi that does
> not work because that state gets reverted when we :reload, and so we
> end up calling g_thread_init again which aborts.
> So now we have a little C wrapper around the g_thread_supported macro
> and we use that to check if g_thread_init was already done.
> ]
> [More SOE fixes
> Duncan Coutts <du...@ha...>**20070629182128
> The main loop now uses a single call to mainGUI which makes
> threading simpler.
> Mouse button release events are now collected as well as button
> down ones.
> Don't specify position of the window for the BouncingBall demo;
> having it
> appear at (800,800) is not very helpful on a 1024x768 screen :-)
> ]
> [Bump version number to something bigger than 0.9.12
> Duncan Coutts <du...@ha...>**20070628172323
> This is now the main/head branch which will lead to release
> version 0.9.13
> The branch that will lead to the 0.9.12 release is at:
> http://darcs.haskell.org/gtk2hs-branches/gtk2hs-0.9.12/
> ]
> [Add several soe demos
> Duncan Coutts <du...@ha...>**20070628171553]
> [Comment out debugging output in soegtk
> Duncan Coutts <du...@ha...>**20070628164947]
> [Make soegtk work with the threaded RTS
> Duncan Coutts <du...@ha...>**20070628164122
> So it should work in GHCi
> ]
> [Fix all package files to use a dep on base and not haskell98
> Duncan Coutts <du...@ha...>**20070531110545]
> [The gtk package now always depends on mtl
> Duncan Coutts <du...@ha...>**20070531105251
> At least with ghc-6.4 and later, not that the .package.conf files are
> only used for ghc-6.4 and the other form is used for ghc-6.2 and
> before
> so we just add it to one file and not the other, rather than needing
> to add more conditional stuff to configure.ac
> ]
> [Change sed script to work with solaris sed
> Duncan Coutts <du...@ha...>**20070531105126
> The problem was with the optional "qualified" clause, but it turns
> out
> we don't need that anyway which allows it to work with the default
> solaris sed.
> ]
> [The ps/pdf_surface_set_size functions appeared in cairo 1.2
> Duncan Coutts <du...@ha...>**20070531104425
> I had thought they were there before but I was confusing them with
> ps/pdf_surface_set_dpi which only appeared in cairo 1.0 but were
> removed
> before the ps & pdf backends were declared stable in cairo 1.2.
> ]
> [Have the gtk package dep on mtl even when not building the cairo
> bindings
> Duncan Coutts <du...@ha...>**20070531094825
> We now always need mtl, previously it was only the cairo bits that
> needed it.
> ]
> [Remove CVS $Revision and $Date tags
> Duncan Coutts <du...@ha...>**20070528172216
> darcs does not use these so they were out of date and thus useless
> ]
> [Don't depend on haskell98 package
> Duncan Coutts <du...@ha...>**20070528154423
> Except in c2hs
> ]
> [Tidy up imports in the gtk package
> Duncan Coutts <du...@ha...>**20070528150937
> drop lots of unused imports
> ]
> [Tidy up imports in the gtkglext package
> Duncan Coutts <du...@ha...>**20070528142849]
> [Tidy up imports in the code produced by the type and signal code
> generators
> Duncan Coutts <du...@ha...>**20070528142715]
> [Tidy up imports in the cairo package
> Duncan Coutts <du...@ha...>**20070528142604]
> [Tidy up imports in the glade and mozembed packages
> Duncan Coutts <du...@ha...>**20070528142538]
> [Tidy up imports in the sourceview package
> Duncan Coutts <du...@ha...>**20070528142416]
> [Tidy up imports in the gconf package
> Duncan Coutts <du...@ha...>**20070528142335]
> [Tidy up imports in the glib package
> Duncan Coutts <du...@ha...>**20070524161155]
> [Tidy up imports in TypeGenerator and HookGenerator
> Duncan Coutts <du...@ha...>**20070524161007]
> [flags properties need the actual flag gtype code
> Duncan Coutts <du...@ha...>**20070528180817
> just like with enums.
> ]
> [Add property functions for Char type
> Duncan Coutts <du...@ha...>**20070528165852]
> [Use the _readonly versions of a couple pango functions when available
> Duncan Coutts <du...@ha...>**20070528143645
> We don't provide any way of modifying the lines so it's ok for us
> to always
> use the optimised read-only versions of these functions.
> ]
> [Use the glib property implementation of
> MenuComboToolbar.Toolbar.toolbarStyle
> Duncan Coutts <du...@ha...>**20070528143506
> rather than the getter/setter style
> ]
> [Use the glib property implementation of
> MenuComboToolbar.Menu.menuTitle
> Duncan Coutts <du...@ha...>**20070528143330
> rather than the getter/setter version.
> ]
> [Export targetListAdd from Selection module
> Duncan Coutts <du...@ha...>**20070528143216
> I presume this was supposed to be exported, it's not otherwise used.
> ]
> [Remove unused files in signal generator
> Duncan Coutts <du...@ha...>**20070528102729]
> [Use cpp without -P if we're building without docs
> Duncan Coutts <du...@ha...>**20070524160833
> So we get accurate ghc error message locations for .hs.pp
> and .chs.pp files.
> ]
> [TAG 0.9.11.1
> Duncan Coutts <du...@ha...>**20070524134504]
> Patch bundle hash:
> 9ea227e65b7c4888a85ab14581188523ec6f52e8
> ----------------------------------------------------------------------
> ---
> This SF.net email is sponsored by: Microsoft
> Defy all challenges. Microsoft(R) Visual Studio 2008.
> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
> _______________________________________________
> Gtk2hs-devel mailing list
> Gtk...@li...
> https://lists.sourceforge.net/lists/listinfo/gtk2hs-devel
|