2005-09-09 Michael Natterer <mitch@gimp.org>
Added parent window API to the GimpProgress interface and to
the libgimp progress stuff. Might look strange, but does
the right thing in almost all cases (image window, file dialog,
script-fu dialog etc). Fixes bug #62988.
* app/core/gimpprogress.[ch]: added GimpProgress::get_window()
which should return a toplevel window ID if the progress is in a
window that wants to be the transient parent of plug-in dialogs.
* app/widgets/gimpwidgets-utils.[ch] (gimp_window_get_native): new
function which returns the window handle of a GtkWindow's GdkWindow.
* app/widgets/gimpfiledialog.c: implement ::get_window().
* app/display/gimpdisplay.[ch]: ditto. Removed window handle API.
* app/gui/gui-vtable.c: changed accordingly.
* libgimpbase/gimpbaseenums.[ch] (enum GimpProgressCommand):
added GIMP_PROGRESS_COMMAND_GET_WINDOW.
* app/plug-in/plug-in-progress.[ch] (plug_in_progress_get_window):
new function. Also renamed some functions to match the
GimpProgress interface, and not the legacy PDB procedure names.
* tools/pdbgen/pdb/progress.pdb
* app/core/gimppdbprogress.c: implement get_window() on both
sides of the wire, keeping backward compatibility (hopefully).
* libgimp/gimpprogress.[ch]: deprecated gimp_progress_install()
and added gimp_progress_install_vtable() which takes a vtable with
padding to be extensible. Added get_window() vtable entry and
dispatch it accordingly. Also added pulse() which was implemented
in a hackish way before. Everything is of course backward
compatible.
* libgimp/gimpprogressbar.c: inmplement the get_window() stuff
so a plug-in dialog containing a progress can be the transient
parent of another dialog in another plug-in.
* libgimp/gimpui.[ch] (gimp_ui_get_progress_window): new function
which returns a foreign GdkWindow of this plug-ins progress
window.
Renamed gimp_window_set_transient_for_default_display() to
gimp_window_set_transient() and make it use the progress' window
handle instead of the display's (which is the right thing to do in
almost all cases).
* libgimp/gimp.def
* libgimp/gimpui.def: add the new functions.
* tools/pdbgen/enums.pl
* app/pdb/internal_procs.c
* app/pdb/progress_cmds.c
* libgimp/gimpprogress_pdb.[ch]: regenerated.
* libgimp/gimpexport.c
* plug-ins/*/*.c: follow API change.
2005-09-09 Tor Lillqvist <tml@novell.com>
* libgimpbase/gimpdatafiles.c (is_script): Practice what I preach
and use g_ascii_strcasecmp() instead of g_strcasecmp().
2005-08-25 Sven Neumann <sven@gimp.org>
* libgimpbase/gimputils.h (gimp_canonicalize_identifier): sync
parameter name with implementation to please gtk-doc.
2005-08-24 Tor Lillqvist <tml@novell.com>
Make GIMP work on Win32 even if installed in a path containing
random non-ASCII characters.
* libgimpbase/gimpenv.c (gimp_toplevel_directory): [Win32] Use
wide-char API on NT-based Windowses. Keep the toplevel directory
name in UTF-8. The rest of GIMP assumes it is UTF-8 anyway, and
for instance passes file names constructed from it to g_open().
* app/plug-in/plug-in.c (plug_in_open): On Win32, if compiled
against GLib < 2.8.2, convert the pathname arguments passed to
g_spawn_async() to locale charset (system codepage). The
g_spawn*() functions in GLib < 2.8.2 take system codepage
arguments and not UTF-8.
2005-08-03 Michael Natterer <mitch@gimp.org>
Changed naming scheme for PDB procedure names from
random_crap_that_traditionally_has_underscores to
enforced-canonical-identifiers. I'm pretty sure some things are
broken after this commit. More changes to come...
* libgimpbase/gimpbase.def
* libgimpbase/gimputils.[ch]: added gimp_canonicalize_identifier().
* app/pdb/procedural_db.[ch] (struct ProcRecord): added
"gchar *original_name" to keep a procedure's original name as
reigstered by plug-ins (compat cruft).
(procedural_db_init_procs): canonicalized list of deprecated
procedures.
* app/plug-in/plug-in-proc-def.c (plug_in_proc_def_free): free
original_name.
* app/plug-in/plug-in-message.c: canonicalize procedure names
which are received over the wire.
* app/plug-in/plug-in-rc.c: serialize the original_name and create
the canonicalized name on-the-fly when deserializing.
* app/plug-in/plug-in-run.c: pass the original_name to plug-ins
when running them because they strcmp() the passed procedure name.
* app/plug-in/plug-ins.c (plug_ins_add_to_db): pass
canonical procedure names to procedural_db_execute().
(plug_ins_file_proc_compare): special-case "gimp-xcf", not "gimp_xcf".
* app/xcf/xcf.c: changed static XCF procedures accordingly.
* tools/pdbgen/app.pl
* tools/pdbgen/lib.pl: do some trivial substitutions to generate
canonicalized names in app/, and C identifiers with underscores in
libgimp/.
* tools/pdbgen/pdb/brushes.pdb
* tools/pdbgen/pdb/fileops.pdb
* tools/pdbgen/pdb/gradients.pdb
* tools/pdbgen/pdb/image.pdb
* tools/pdbgen/pdb/palettes.pdb
* tools/pdbgen/pdb/patterns.pdb
* tools/pdbgen/pdb/plug_in.pdb
* tools/pdbgen/pdb/procedural_db.pdb
* tools/pdbgen/pdb/text_tool.pdb
* tools/pdbgen/pdb/transform_tools.pdb: canonicaloized procedure
names in calls to std_pdb_deprecated() and in procedure names in
generated C code.
* app/pdb/*_cmds.c
* libgimp/*_pdb.c: regenerated.
2005-07-29 Sven Neumann <sven@gimp.org>
* libgimpbase/gimpbaseenums.[ch]: added GimpForegroundExtractMode
in preparation for other implementations to be added in the future.
* libgimp/gimpenums.c.tail
* tools/pdbgen/enums.pl: regenerated.
* app/core/gimpdrawable-foreground-extract.[ch]
* app/tools/gimpforegroundselecttool.c
* tools/pdbgen/pdb/drawable.pdb: added the mode parameter.
* app/pdb/drawable_cmds.c
* libgimp/gimpdrawable_pdb.[ch]: regenerated.
2005-05-14 Michael Natterer <mitch@gimp.org>
* libgimpbase/gimpbaseenums.[ch]: register descriptions for the
GimpPDBProcType enum.
* plug-ins/dbbrowser/gimpprocview.c
* tools/pdbgen/pdb/procedural_db.pdb: get rid of all selfmade
enum->string mapping. Get the strings from the GType system instead.
* app/pdb/procedural_db_cmds.c
* libgimp/gimpproceduraldb_pdb.c: regenerated.
2005-04-22 Sven Neumann <sven@gimp.org>
* libgimpbase/gimpbaseenums.h (GimpTransformDirection): removed
(traditional) and (corrective) from the enum descriptions.
2005-04-16 Sven Neumann <sven@gimp.org>
Applied a patch (with minor modifications) from Karine Delvare
that makes it possible to choose between various ways to remove
color from a drawable (bug #155507):
* libgimpbase/gimpbaseenums.h: added new enum GimpDesaturateMode.
* libgimpbase/gimpbaseenums.c
* libgimp/gimpenums.c.tail
* plug-ins/pygimp/gimpenums.py
* tools/pdbgen/enums.pl: regenerated.
* app/core/gimpdrawable-desaturate.[ch]: implement other formulas
to remove colors from a drawable.
* tools/pdbgen/pdb/color.pdb: let gimp_desaturate() call
gimp_drawable_desaturate() with GIMP_DESATURATE_LIGHTNESS so that
it behaves like it always did.
* app/pdb/color_cmds.c: regenerated.
* app/dialogs/Makefile.am
* app/dialogs/desaturate-dialog.[ch]: new files that define a simple
dialog to choose a mode for desaturation.
* app/actions/drawable-actions.[ch]: use the new dialog.
2005-04-05 Sven Neumann <sven@gimp.org>
* configure.in: renamed localedir to gimplocaledir.
* libgimpbase/Makefile.am (AM_CPPFLAGS)
* tools/Makefile.am (AM_CPPFLAGS): changed accordingly.
* gimp.pc.in: added gimplocaledir to the pkg-config file to give
plug-ins a chance to get gimp_locale_directory() at compile time.
2003-03-06 Manish Singh <yosh@gimp.org>
* configure.in: revised tests for AltiVec. Define ALTIVEC_EXTRA_CFLAGS
for the extra compiler options needed. Also, support runtime checking
for AltiVec through the sysctl on Mac OS X, instead of SIGILL assembly
instruction checking, which needs GNU as.
* app/base/cpu_accel.c: use the sysctl if available.
* app/composite/Makefile.am: use ALTIVEC_EXTRA_CFLAGS.
* app/composite/gimp-composite-altivec.c: conditionally #include
altivec.h
2005-02-19 Hans Breuer <hans@breuer.org>
* app/base/pixel-processor.c : TILE_WIDTH is used unconditionally
so always include "tile.h"
* app/base/tile-swap.c : WIN32 needs <process.h> for _getpid()
* app/dialogs/user-install-dialog.c : include gimpwin32-io.h
* libgimpbase/gimpwin32-io.h : there are no group or other
flags in msvcrt, define S_IGRP etc in terms of _S_IREAD etc
* plug-ins/script-fu/script-fu.c plug-ins/script-fu/siod-wrapper.c :
no script-fu server on win32, make respective function calls conditional
* libgimpconfig/makefile.msc : new file
* **/makefile.msc app/gimpcore.def : updated, gimp builds
and runs once more with ms toolchain
2005-02-12 Michael Natterer <mitch@gimp.org>
* libgimpbase/gimpbaseenums.h (enum GimpProgressCommand):
added GIMP_PROGRESS_COMMAND_PULSE.
* libgimpbase/gimpbaseenums.c
* plug-ins/pygimp/gimpenums.py
* tools/pdbgen/enums.pl: regenerated.
* app/core/gimppdbprogress.c: implement GimpProgress::pulse()
and send a PULSE command to the callback.
* libgimp/gimpprogress.c: handle PULSE by calling the set_value()
callback with a value of -1 and document that hack in the API docs.
* libgimp/gimpprogressbar.c: interpret -1 as request to pulse.
2005-01-23 Sven Neumann <sven@gimp.org>
* libgimpbase/Makefile.am
* libgimpbase/gimpbaseparams.[ch]: removed this file again.
* libgimpbase/gimpmemsize.[ch]
* libgimpbase/gimppath.[ch]
* libgimpbase/gimpunit.[ch]: moved the paramspec definitions to
the types they belong to.
* libgimpbase/gimpbase.h: changed accordingly.
* libgimpwidgets/Makefile.am (libgimpwidgetsinclude_HEADERS):
install the new header files.
2005-01-23 Sven Neumann <sven@gimp.org>
* app/paint/paint-enums.h
* libgimpbase/gimpbaseenums.h: moved enums to libgimpbase (from
app/paint and libgimp). The remaining enums in
app/paint/paint-enums.h need special treatment.
* app/paint/paint-enums.c:
* libgimp/gimpenums.h
* libgimpbase/gimpbaseenums.c
* tools/pdbgen/enums.pl: regenerated.
* libgimpbase/gimpbase.def
* libgimp/gimp.def: updated.
2005-01-23 Sven Neumann <sven@gimp.org>
* app/core/core-enums.h
* libgimpbase/gimpbaseenums.h: moved enums to libgimpbase (from
app/core and libgimp). The remaining enums in app/core/core-enums.h
need special treatment.
* app/core/core-enums.c:
* libgimp/gimpenums.h
* libgimpbase/gimpbaseenums.c
* tools/pdbgen/enums.pl: regenerated.
* libgimpbase/gimpbase.def
* libgimp/gimp.def: updated.
2005-01-23 Sven Neumann <sven@gimp.org>
* app/base/base-enums.h
* libgimpbase/gimpbaseenums.h: moved GimpInterpolationType and
GimpTransferMode to libgimpbase (from core and libgimp). The
remaining enums in app/base/base-enums.h need special treatment.
* app/base/base-enums.c
* libgimp/gimpenums.h
* libgimpbase/gimpbaseenums.[ch]
* tools/pdbgen/enums.pl: regenerated.
* libgimpbase/gimpbase.def
* libgimp/gimp.def: updated.
2005-01-22 Sven Neumann <sven@gimp.org>
* libgimpbase/gimputils.c (gimp_enum_value_get_desc): no point in
crashing if gimp_enum_get_desc() returns NULL.
* modules/cdisplay_colorblind.c
* modules/cdisplay_proof.c: register enum descriptions, use
GimpEnumComboBox.
2004-12-14 Sven Neumann <sven@gimp.org>
* configure.in: added GIMP_API_VERSION to the generated gimpversion.h.
* libgimpbase/gimpenv.c (gimp_toplevel_directory): use
GIMP_API_VERSION instead of GIMP_MACRO_VERSION.GIMP_MINOR_VERSION
when building a path to test the plug-in executable path against.
2004-11-23 Michael Natterer <mitch@gimp.org>
* app/core/gimpparasitelist.c: cosmetics, untabified.
* libgimpbase/gimpparasiteio.[ch]: added g_return_if_fail()'s
to all functions.
(gimp_pixpipe_params_parse): changed "gchar*" param to "const
gchar*" (sortof API change, but these files are most probably only
used by GIMP itself). Still uses strtok() on the internal copy,
but at least not on the passed string.
* plug-ins/common/csource.c
* plug-ins/common/gif.c
* plug-ins/common/gih.c
* plug-ins/common/jpeg.c
* plug-ins/common/png.c
* plug-ins/common/tiff.c: use parasite getters instead of
accessing the scruct members directly. Always use g_strndup()
instead of just g_strdup() to get strings stored in parasites
because there is no guarantee that they are nul-terminated.
2004-11-22 Michael Natterer <mitch@gimp.org>
* libgimpbase/gimputils.c (gimp_any_to_utf8): use g_strndup()
instead of g_strdup() if a length was passed.
* app/dialogs/info-window.c: g_strndup() the comment parasite's
data and pass -1 as length to gimp_any_to_utf8() so we don't
encounter the questionable (buggy?) behavior of g_utf8_validate()
to fail upon finding '\0' within the "length" passed.
Fixes bug #159051.
2004-09-21 Hans Breuer <hans@breuer.org>
* app/dialogs/makefile.msc : [new file]
app/dialogs/Makefile.am : added to EXTRA_DIST
* **/makefile.msc app/gimpcore.def : updated
* app/gimp.rc : let wilber be first
* app/widgets/gimppropwidgets.c : msvc6 can't cast uint64 either
* libgimpbase/gimpwin32-io.h : make up recent loss of ftruncate in GLib
* libgimpthumbnail/gimpthumbnail.c : <process.h> for getpid() on win32
* plug-ins/helpbrowser/dialog.c : include gimpwin32-io.h
* plug-ins/script-fu/siodwrapper.c plug-ins/script-fu/scrip-fu.c : there
is no script-fu-server on win32
2004-11-13 Sven Neumann <sven@gimp.org>
* libgimpbase/gimputils.c (gimp_strip_uline): changed based on a
patch by Joao S. O. Bueno to remove mnemonics as used in languages
like Chinese. Fixes bug #157561.
2004-10-25 Michael Natterer <mitch@gimp.org>
* libgimpbase/gimpbaseenums.[ch]: register some missing enums.
* tools/pdbgen/enumcode.pl: removed code to generate
plug-ins/script-fu/script-fu-constants.c, generate code to
explicitely initialize and query all of libgimp*'s enums
and write it to libgimp/gimpenums.c.tail
* libgimp/gimpenums.h: regenerated.
* libgimp/Makefile.am: append gimpenums.c.tail to gimpenums.c
* libgimp/gimp.c (gimp_main): call g_type_init() and
_gimp_enums_init().
* libgimp/gimp.def: added gimp_enums_get_type_names().
* plug-ins/script-fu/Makefile.am
* plug-ins/script-fu/script-fu-constants.[ch]: removed these files.
* plug-ins/script-fu/siod-wrapper.c: dynamically register all
constants using gimp_enums_get_type_names() and introspection.
Also register the built-in unit types.
* plug-ins/script-fu/script-fu.c: changed accordingly.
2004-10-25 Michael Natterer <mitch@gimp.org>
Don't store human readable and translatable enum/flag strings in
GEnumValue's and GTypeValue's fields but attach them to their
GType using separate structs and utility functions:
* tools/gimp-mkenums: added params and perl voodoo to support
generating a second array of values, which is used by the
Makefiles below to create and register arrays of value
descriptions.
* libgimpbase/gimpbasetypes.[ch]: added API to attach/retreive
arrays of translatable strings to/from enum and flags types. Added
structs GimpEnumDesc and GimpFlagsDesc for that purpose.
* libgimpbase/gimputils.[ch]: changed existing enum utility
functions, added new ones and added a symmetric API for flags.
* app/base/Makefile.am
* app/core/Makefile.am
* app/display/Makefile.am
* app/paint/Makefile.am
* app/text/Makefile.am
* app/tools/Makefile.am
* app/widgets/Makefile.am
* libgimp/Makefile.am
* libgimpbase/Makefile.am: changed *-enums.c generation rules
accordingly.
* app/base/base-enums.c
* app/core/core-enums.c
* app/display/display-enums.c
* app/paint/paint-enums.c
* app/text/text-enums.c
* app/tools/tools-enums.c
* app/widgets/widgets-enums.c
* libgimpbase/gimpbaseenums.c: regenerated.
* app/widgets/gimpenumstore.c
* app/widgets/gimpenumwidgets.c
* app/widgets/gimptemplateeditor.c
* libgimpwidgets/gimppreviewarea.c: follow the enum utility
function API changes.
2004-10-04 Sven Neumann <sven@gimp.org>
* app/config/gimpguiconfig.[ch]
* app/config/gimprc-blurbs.h: added gimprc option "show-help-button".
* app/dialogs/preferences-dialog.c: added a GUI for it.
* app/dialogs/file-save-dialog.c
* app/dialogs/image-new-dialog.c
* app/dialogs/quit-dialog.c
* app/display/gimpdisplayshell-close.c
* app/widgets/gimphelp-ids.h: don't set help-ids on confirmation
dialogs.
* libgimpbase/gimpprotocol.[ch]
* libgimp/gimp.[ch]: added boolean "show_help_button" to the
config message.
* app/plug-in/plug-in-run.c: pass the new preference to the plug-in.
* libgimpwidgets/gimpdialog.[ch]: added new function that allows to
set whether new dialogs should get a help button added.
* app/gui/gui.c
* libgimp/gimpui.c: call gimp_dialogs_show_help_button() according
to the gimprc settings.
2004-10-04 Michael Natterer <mitch@gimp.org>
* libgimpbase/gimpwire.c: added assertions to make sure "count" is
always >= 0. Turns the crash described in bug #154266 into a
warning plus corrupted wire state :) Real fix (in script-fu) will
follow. Untabified.
Michael Natterer