2000-10-28 Michael Natterer <mitch@gimp.org>
* app/gimpdrawable.[ch]
* libgimp/gimpsizeentry.c: did some code review: s/0/NULL/,
s/sprintf/g_snprintf/, spacing.
2000-05-12 Sven Neumann <sven@gimp.org>
* gimpdrawable.c: enabled the (commented out) signal
"invalidate_preview".
* app/layers_dialog.c: connect to the "invalidate_preview"
signal to catch changes that need to be shown in the layer
previews. Synthetize an expose event when a layer changes.
Expose events are optimzed away by GTK+ if the widget is not
visible. Therefore, previews not visible in the layers_dialog
are not redrawn when they invalidate. Later the preview gets
validated by the image_preview in lc_dialog but is never
propagated to the layer_pixmap. We work around this by using an
additional flag "layer_pixmap_valid" so that the pixmap gets
updated once the preview scrolls into sight.
Fixes bugs #10549, #10300 and #8787.
* app/channel.[ch]
* app/layer.[ch]: code review and indentation
--Sven
* app/gimpdrawable.c: gimp_drawable_get_color_at() now silently
returns NULL again if the coordinates are out of range. A lot of
code using this function relies on this feature and correctly
checks the return value. No need to emit critical warnings here.
The GTK_CHECK_TYPE macro test for obj != NULL, no need to do this
check twice. Removed lots of unnecessary calls to g_return_if_fail().
* app/color_picker.c: with the old behaviour of
gimp_drawable_get_color_at() the code is a bit simpler.
* app/fuzzy_select.c: fuzzy_select relied on drawable_offsets()
returning off_x = off_y = 0 if drawable == NULL. Decided to change
this here, fixes bug #7077.
* app/gimpimage.[ch]: Even though we made bad experiences with the
changes in gimpdrawable.c, I have introduced similar argument checks
here.
* app/image_map.c: indentation
--Sven
2000-03-01 Michael Natterer <mitch@gimp.org>
* app/gimpdrawable.c: gimp_drawable_get_color_at(): put the check
below back but only require a non-indexed drawable if the drawable
doesn't have an image.
(The function may call gimp_image_get_color() with gimage == NULL
for RGB and GRAY colors because we don't need a colormap to
transform them to RGB space).
This was the obscure special case (TM).
2000-03-01 Michael Natterer <mitch@gimp.org>
* app/gimpdrawable.c: gimp_drawable_get_color_at(): don't
g_return_val_if_fail(!indexed, NULL) because the function seems to
work fine for all kinds of drawables. Fixes#6756. Or is this
check needed for some obscure special case?
2000-02-27 Garry R. Osgood <gosgood@idt.net>
* gimp/app/by_color_select.c
* gimp/app/color_picker.c
* gimp/app/gimpdrawable.c
* gimp/app/image_map.c
* gimp/app/paint_core.c
Inadvertent logic error in g_return_val_if_fail()
style sanity checks implemented in
gimp_drawable_get_color_at() gave rise to
segment violation reported in #6624;
error admitted out-of-bounds x&y that
do not map to tiles. Closes#6624.
GTK-critical warnings which result from
this new sanity check require that
gimp_drawable_get_color_at() clients
perform initial culling of out-of-bounds
x & y coordinates.
* tools/pdbgen/*.pl: added 2000 to the copyrights
* tools/pdbgen/lib.pl: it's foo_pdb.[ch] now
* tools/pdbgen/pdb/drawable.pdb
* tools/pdbgen/pdb/parasite.pdb: reenabled sanity checks for the
drawable IDs and the parasite names
* app/drawable.c
* app/gimpdrawable.c: added a bunch of g_return_if_fail's, instead
of the silent check and return in many functions. This should also
be done in gimpimage.c and the like too.
* app/gimpdrawable.h
* app/layer.c: cosmetic code fix
-Yosh
Tue Feb 15 23:27:42 GMT 2000 Andy Thomas <alt@gimp.org>
* gimp/app/gimpdrawable.c
* gimp/app/channel.c
* gimp/app/layer.c
* gimp/app/channel.h
* gimp/app/layer.h
* gimp/app/gimpimage.c
* gimp/app/gimpimage.h
* gimp/app/gimpdrawable.h
* gimp/tools/pdbgen/pdb/paths.pdb
* gimp/tools/pdbgen/pdb/layer.pdb
* gimp/tools/pdbgen/pdb/channel.pdb
* gimp/tools/pdbgen/pdb/gimage.pdb
* gimp/app/channel_cmds.c
* gimp/app/gimage_cmds.c
* gimp/app/internal_procs.c
* gimp/app/layer_cmds.c
* gimp/app/paths_cmds.c
New gimp_*_set_tattoo procedures. This allows save/load plugins
to save/restore tattoo states of layers, channels and paths. Note the
internal tattoo state can also be set, however rigorous checks are
performed to make sure that the internal tattoo states of layer,
channels and paths are consistent and that the new state value is
newval > MAX(MAX(layertattoo),MAX(channeltattoo),MAX(pathtattoo)).
* app/gimpdrawable.c: added an assert for name != NULL
* tools/pdbgen/pdb/channel.pdb
* tools/pdbgen/pdb/layer.pdb: sanity check the name parameters
-Yosh
Fri Oct 1 12:46:12 1999 Austin Donnelly <austin@gimp.org>
* gimprc.in: comment typo fix, plus add %D* to default
image-title-format string, so people get a '*' in the titlebar
if their image is dirty.
* app/fileops.c: initialise filename before using it.
* app/gdisplay.c: empty parameter list () is K&R - should be
stronger (void) in ANSI C.
* app/gimpdrawable.c: gimp_drawable_{dirty,clean} functions
removed - no one uses them anyway. Parasite undo type is
proper parasite undo type, not MISC_UNDO.
* app/gimpdrawableP.h: drawable dirty bit removed.
* app/gimpimage.c: don't change the resolution if there's no
difference from the old one. Call gdisplay_shrink_wrap() to
re-calculate scale factors and refresh the display on
resolution change. Layer undo doesn't have sub-types
anymore, uses main UndoType instead.
* app/layer.h: Remove LayerUndoType
* app/qmask.c: fix qmask undo so it actually works.
* app/undo.h: new types for undo_push_layer{,_mask} and
undo_push_qmask.
* app/undo.c: change way group boundaries are represented:
each Undo has a group_boundary boolean set to TRUE if this is
the start or the end of a group, and the type of the Undo is
the group's type. Within a group, each Undo keeps its own
type. This allows pop funcs and free funcs to do
type-specific things (eg needed by layer and channel stuff).
Don't maintain per-drawable dirty flags anymore. Floating
sel to layer and layer rename now uses meaningful undo types.
* app/undo_types.h: more specific undo types:
LAYER_{ADD,REMOVE}_UNDO, LAYER_MASK_{ADD,REMOVE}_UNDO,
LAYER_RENAME_UNDO, and PARASITE_{ATTACH,DETACH}_UNDO.
* app/undo_history.c: oops - undo stack was being placed into gtk
list in wrong order.
* app/edit_selection.c: push more descriptive LAYER_DISPLACE_UNDO
rather than MISC_UNDO.
* app/layers_dialog.c: better tagging of undo types
Mon Sep 20 12:51:30 EDT 1999 Austin Donnelly <austin@gimp.org>
Honest, guv, it's not a feature - it's a tightly integrated
package of undo system cleanups and fixes.
NEW FILES:
* app/undo_history.c: window showing recent undo (and redo) steps
available.
* app/undo_types.h: broken out of undo.h to fix circular includes.
MODIFIED FILES:
* app/Makefile.am: compile undo_history.c
* app/channel.h: use enum for channel undo type, not just magic
numbers.
* app/layer.h: same for layer undos.
* app/commands.c: edit_show_undo_history_cmd_callback() function to
pull up undo history window.
* app/commands.h: prototype for above.
* app/gdisplay.c: make undo / redo menu items sensitive according
to whether they would do anything. Would be easy to change
the text to say what would be undone/redone, but I don't know
the GTK.
* app/gimpimage.c: new signal emitted by gimage:
UNDO_EVENT. gimp_image_undo_event() function to emit it.
* app/gimpimage.h: prototype for above.
* app/gimpimageP.h: pushing_undo_group member is now an undo_type,
not an int. Keep undo history widget here too (if created).
* app/menus.c: add "Edit/Undo history..." to image menu.
* app/undo.c: new types: enums undo_type and undo_state rather than
ints and magic numbers. All undo_pop_* and undo_free_*
functions made static. New static function
undo_type_to_name(). Issue undo event signals on various
important events (eg undo pushed, undo popped etc).
undo_push() now takes a "dirties_image" arg to say whether
image should be dirtied. Layer moves now dirty the image. A
couple of g_return_if_fails () on undo_pop and undo_redo to
assert we're not in the middle of an undo group.
undo_get_{undo,redo}_name() to peek at names of top items on
undo and redo stacks resp. undo_map_over_{undo,redo}_stack()
to run a function for each item or group on stack. Layer and
channel undos use symbolic names rather than 0 or 1. Array
mapping undo types to names.
* app/undo.h: split out undo types to undo_types.h. Prototypes
for functions described above. undo_event_t enum.
undo_history_new() prototype lives here too.
Random other fixes:
* app/gimpdrawable.c
* app/image_render.c: default labels in switches to keep egcs happy.
* app/nav_window.c: some fixes to (sort of) cope with image res !=
screen res. Still needs work to handle non-square pixels
properly.
* app/paths_dialog.c: bad idea to call gimp_image_dirty()
directly. Even though it's currently commented out.
* libgimp/Makefile.am: add gimpmath.h
* app/gtkwrapbox.[ch]
* app/gtkhwrapbox.[ch]: wrapbox widget from gle
* app/Makefile.am: added those files
* app/interface.c: use an hwrapbox for the toolbar. Still not perfect
yet, working on it.
* app/gimpdrawable.c
* app/about_dialog.c
* app/airbrush.c
* app/blend.c: some minor code cleanup
-Yosh
Mon Aug 23 10:15:32 EDT 1999 Austin Donnelly <austin@gimp.org>
Dirty flag now correct in all cases. Can be displayed in image
window title too. See NOTE near gimp_image_dirty() for details.
* app/fileops.c: gimp_image_clean_all() after reverting an image.
* app/gdisplay.c: register handlers for gimage dirty and clean
signals to update image title. New image-title-format
expansion: %Dx expands to x if the image is dirty.
* app/gdisplay_ops.c: gimage->dirty flags != 0 is the correct
condition to test to see if an image is dirty.
* app/gimpdrawable.c: gimp_image_dirty() should never be called
except from an undo_push_* function. Call
undo_push_cantundo() if you want to dirty the image but can't
be bothered writing an undo handler (be ashamed of yourself!).
* app/gimpimage.c: new gimage signal: clean. Emitted when an undo
operation takes place. Gimage changes when either dirty or
clean is emitted, so if you need to update previews etc, look
for both! Move group_count into gimage structure, since
leaving it as a static in undo.c is bad if two undo groups are
started on different images at the same time. More changes
of gimp_image_dirty() to undo_push_cantundo()
(parasite-related, plus layer moves). See the NOTE on dirty
counter near gimp_image_dirty() for the full story.
gimp_image_dirty() and gimp_image_clean() simplified - counter
can go negative.
* app/gimpimageP.h: group_count moved from undo.c
* app/layers_dialog.c: push undo for layer name change, rather
than dirtying the image.
* app/undo.c: layer rename undo functions
added. undo_push_cantundo() convenience functions added.
group_count made per-gimage since everything else is. When
blowing away redo stack, make image infinitely dirty if redo
info contained file save point.
* app/undo.h: added undo_push_layer_rename() and
undo_push_cantundo().
* TODO: added idea for undo history window.
1999-08-22 Michael Natterer <mitschel@cs.tu-berlin.de>
* app/color_area.[ch]
* app/color_panel.[ch]
* app/gimpdnd.[ch]: enabled dnd for colors (compatible with the
standard gtk/gnome color dnd). The color selection is now invoked
on a real click, not on button_down to avoid confusion with dnd.
* app/channels_dialog.c
* app/disp_callbacks.c
* app/interface.c
* app/layers_dialog.c: minor dnd updates/fixes.
* app/channel.[ch]
* app/channel_pvt.h
* app/docindex.[ch]
* app/docindexif.[ch]
* app/drawable.[ch]
* app/floating_sel.c
* app/gimage.[ch]
* app/gimage_mask.c
* app/gimpdrawable.[ch]
* app/gimpdrawableP.h
* app/gimpimage.[ch]
* app/gimpimageP.h
* app/layer.[ch]
* app/layer_pvt.h
* app/undo.c
* app/xcf.c: wanted to do some s/int/gboolean/ in the layer files
where appropriate and found myself spending the whole night doing
a big code review for layers/channels/drawables/images:
s/int/gboolean/, s/<type>/g<type>/, lots of indentation, removed
some old global variables and deprecated functions, #include
cleanups, proper prototypes, copyright headers, ...
* app/*.[ch]: Actually use the enum types GimpImageType,
GimpImageBaseType, LayerModeEffects, PaintApplicationMode,
BrushApplicationMode, GimpFillType and ConvertPaletteType, instead
of just int or gint. Hopefully I catched most of the places
where these should be used.
Add an enum ConvolutionType, suffix the too general constants
NORMAL, ABSOLUTE and NEGATIVE with _CONVOL. Use NORMAL_MODE
instead of NORMAL in some places (this was what was intended). Fix
some minor gccisms.
* app/apptypes.h: New file. This file contains the above
enumeration types, and some opaque struct typedefs. It was
necessary to collect these in one header that doesn't include
other headers, because when we started using the above mentioned
types in the headers, all hell broke loose because of the
spaghetti-like cross-inclusion mess between headers.
(An example: Header A includes header B, which includes header C
which includes A. B uses a type defined in A. This is not defined,
because A hasn't defined it yet at the point where it includes B,
and A included from B of course is skipped as we already are
reading A.)
Wed Jun 23 23:52:54 BST 1999 Andy Thomas <alt@gimp.org>
* gimp/app/gimppreviewcache.c
* gimp/app/gimppreviewcache.h
* gimp/app/drawable_cmds.c
* gimp/app/gdisplay.c
* gimp/app/gimpdrawableP.h
* gimp/app/gimage_cmds.c
* gimp/app/Makefile.am
* gimp/app/layers_dialog.c
* gimp/app/channel.c
* gimp/app/lc_dialog.c
* gimp/app/lc_dialog.h
* gimp/app/lc_dialogP.h
* gimp/app/layer.c
* gimp/app/gimpdrawable.c
* gimp/app/internal_procs.c
* gimp/libgimp/gimp.h
* gimp/libgimp/gimpimage.c
* gimp/libgimp/gimpdrawable.c
* gimp/libgimp/gimpmenu.c
* gimp/tools/pdbgen/pdb/drawable.pdb
* gimp/tools/pdbgen/pdb/gimage.pdb
Added thumbnail image preview functions.
Previews are visible on the L&C&P dialogs as well as in the
drawables/channels/ menus generated for plugins
(see the bumpmap & Mapobject plugins).
PDB interface exists to simply extract a thumbnail preview
of a given size. This is much quicker & more efficient
than getting the image data tile-by-tile if you only need a small
image since a "preview cache" has been implemented. This cache also
reduces the number of times the tiles cached is scanned since smaller
previews are always generated from large ones if they exists and
are valid.
Some possible usages (I don't intend to implement these ideas. Just
suggestions). More plugins using the thumbnail preview (ie any that
use multiple images). Indication of "active image" somewhere.....
Actually almost anywhere a drawable/image name appears.
* app/edit_selection.c: removed some nonfunctional code.
* app/paint_core.c: remove the alt toggles perfectmouse behaviour.
* app/paintbrush.c: when ctl (or alt) is held set the fg (or bg) color.
* app/gimpparasite.[ch]: made char *name parameters const.
* app/parasitelist.c: removed unused static variable.
* app/gimpdrawable.c, app/gimpimage.c, app/undo.[ch]: added
support for undoing parasite changes.
* libgimp/gimp.h, libgimp/gimpimage.c: added
gimp_undo_push_group_start and gimp_undo_push_group_end
* libgimp/parasite.[ch]: added undoable flag.
* plug-ins/gdyntext/font_selection.c: fixed c++ style comment.
* plug-ins/gdyntext/gdyntext.c: use the new undoable parasites.
* plug-ins/rcm/rcm_misc.c: arctg can't be inline because it is
used in other .c files
* plug-ins/waterselect/waterselect.c,
* plug-ins/rotators/rotators.c, app/tips_dialog.c, app/plug_in.c:
fixed some warnings
* app/Makefile.am: build color_cmds, lut_funcs, and pixel_processor
* app/app_procs.c: feedback in the splash screen when loading
parasites.
* app/boundary.c: Optimized find_empty_segs.
* app/brightness_contrast.[ch]
* app/levels.[ch]
* app/posterize.[ch]:
moved pdb and lut calculation code. These files now contain only
GUI functions.
* app/channel.c: Optimized channel_bounds (fewer compares, better
use of registers). Use color_region instead of channel_*_segment
in channel_combine_rect. Optimized channel_combine_ellipse by
skipping pixels inside of the ellipse. Use
pixel_region_process_parallel in channel_combine_mask. Use a
GimpLut in channel_invert, and channel_sharpen.
* app/invert.c
* app/equalize.c: moved the lut functions to lut_funcs.c
* app/gimpdrawable.c, app/gimpdrawableP.h
* app/gimpimage.c, app/gimpimageP.h: removed unused gimpmatrix
variables/includes.
* app/gimplut.[ch]: added new function gimp_lut_process_inline
that operates on a single PixelRegion.
* app/gimpparasite.[ch]: new functions to save/load parasiterc
* app/parasitelist.[ch]: new functions to save/load ParasiteLists
in/from files.
* libgimp/parasite.[ch]: new functions to load/save parasites.
* app/internal_procs.c: get some procs from new location in
color_cmds.h.
* app/pixel_region.[ch]: moved pixel_regions_process_parallel
related functions to a new file.
* app/color_cmds.[ch]: new files for PDB
definitions/implementations of color correction functions.
* app/lut_funcs.[ch]: new files to hold lut creation functions.
* app/pixel_processor.[ch]: new files that contain the
pixel_regions_process_parallel routines. Added some new
capabilities that are currently unused.
done in pieces, don't expect to be able to compile on Win32 from
these sources yet. Ans of course, the official version of GTk+
doesn't include the Win32 stuff yet.
* app/gimplut.[ch]: new source files that implement pixel Look Up
Table functions.
* app/Makefile.am: build gimplut.[ch]
* app/brightness_contrast.c
* app/curves.c
* app/invert.c
* app/levels.c
* app/posterize.c: Use the new lut functions. Use
pixel_region_process_parallel in the PDB versions of these routines.
* libgimp/parasite.h
* libgimp/parasite.c: new functions parasite_name and
parasite_compare.
* app/gimpdrawable.c:
* app/gimpdrawable.h: new function
gimp_drawable_get_color_at(...) returns the RGBA[color index]
value at a specified position in the drawable. Don't set the dirty
bit on the image if a new parasite is the same as the old.
* app/gimpimage.c
* app/gimpimage.h new function
gimp_image_get_color_at(...) returns the RGBA[color index]
value at a specified position in the drawable. Don't set the dirty
bit on the image if a new parasite is the same as the old.
* app/by_color_select.c
* app/color_picker.c: use the new gimp_*_get_color_at
functions instead of messing with the tiles.
* app/layer.c: fixed a minor warning.
* app/commands.c:
don't scale the image if the new size == the old size
* app/channel.c: optimized channel_bounds by only checking the
pixels in a tile if it is not already entirely within the
currently computed bounds.
* app/gimpdrawable.c (gimp_drawable_set_name): Fix adding of
numbers after drawable copies. We used to get pairs of duplicates
after adding the same name several times.
(Try by adding a new layer several times in the layers dialog. You will
get New Layer, New Layer#2, New Layer#3, New Layer#3 (again!), New Layer#4,
New Layer#4 (again!), etc.)
Wed Jan 27 21:52:43 GMT 1999 Austin Donnelly <austin@gimp.org>
* app/scale.c: oops - should use the image resolution, not the
screen resolution. D'oh!
* app/gimpdrawable.c: only merge shadow image if there actually is
one.
* plug-ins/emboss/emboss.c: small optimisation: use
gimp_drawable_mask_bounds() to calculate only the pixels that
are actually in the selection.
* libgimp/gserialize.c
* libgimp/gserialize.h: Changed the enum values to allow for
simpler future expansion.
* libgimp/parasite.c
* libgimp/parasite.h: s/persistant/persistent/.
new accessor functions for parasites. #defines for new flags.
* app/paintbrush.c: added timeing code for brush strokes.
It is #ifed out, and is only valid for shift clicks.
* app/parasite_cmds.h: fixed a warning
* app/parasitelist.h
* app/parasitelist.c: added _for_each and _length functions
* app/gimpdrawable.c: set the dirty flag when adding or removing a
persistent parasite
* app/gimpimage.c: set the dirty flag when adding or removing a
persistent parasite. Fixed bug and removed debug statements in
merge_down.
* app/xcf.c: save and load resolution, parasites, and tattoos.
* app/main.c: updated the deserialize test.
* plug-ins/tiff/tiff.c
* plug-ins/gif/gif.c: use PARASITE_PERSISTENT define instead of 1
* plug-ins/bmp/bmp.c
* plug-ins/bmp/bmp.h: declare some struct variable as extern.
* app/paint_funcs.c: Lots of optimizations aimed at speeding up
painting. Should see a 2-4X speed up on most painting
(depending on paint modes, brush size etc.)
* app/drawable.c: check for NULL drawable in drawable_ID.
this stops us from being crashed by ill-behaved plug-ins
objects. (And to automatically manage their signals, not implemented
yet)
Moved drawable_apply_image to drawable.c
Created a global context object (image_context) to handle the
collection of images that the app manages.
Started separating crud out of drawables.
Isolated the id system of images entirely within pdb. Even the
window titles and menus use pointers instead of ids. Should at
least remind people that this is a developers' version. :)
Michael Natterer