r14651@tres: ted | 2007-02-28 23:23:41 -0800
Wow, lost namespacing too. Err, SVK.
Wow, lost namespacing too. Err, SVK.
r14649@tres: ted | 2007-02-28 23:21:44 -0800
Somehow that didn't get deleted.
Somehow that didn't get deleted.
r14632@tres: ted | 2007-02-28 23:12:58 -0800
Splitting out the command line action class to try and leave SOME
dependencies out of main.cpp.
Splitting out the command line action class to try and leave SOME
dependencies out of main.cpp.
r14584@tres: ted | 2007-02-28 20:01:52 -0800
Effects now register two different verbs. One is the same one they've
always registered and one is the same ID with ".nopref" added to it.
This one executes the effect without the prefrences dialog. Default
values will be used.
Effects now register two different verbs. One is the same one they've
always registered and one is the same ID with ".nopref" added to it.
This one executes the effect without the prefrences dialog. Default
values will be used.
r14583@tres: ted | 2007-02-28 19:29:48 -0800
Selection is working too!
Selection is working too!
r14582@tres: ted | 2007-02-28 19:07:35 -0800
Oh Yeah! Verbs work!
Oh Yeah! Verbs work!
r14581@tres: ted | 2007-02-27 19:36:54 -0800
Calling everything on the command line. Good stopping point.
Calling everything on the command line. Good stopping point.
r14579@tres: ted | 2007-02-27 19:00:22 -0800
Adding in an extension init in arguably the wrong place, but it does
make the code much more readable. I guess that makes it the right place
then doesn't it? ;)
Adding in an extension init in arguably the wrong place, but it does
make the code much more readable. I guess that makes it the right place
then doesn't it? ;)
r14578@tres: ted | 2007-02-27 18:59:35 -0800
Protecting against the posibility of the INKSCAPE variable being NULL
Protecting against the posibility of the INKSCAPE variable being NULL
r14577@tres: ted | 2007-02-27 18:51:05 -0800
Filling in the list function to list all the verbs for the command line
option --verb-list
Filling in the list function to list all the verbs for the command line
option --verb-list
r14576@tres: ted | 2007-02-26 23:29:13 -0800
Basic construction in place to --verb-list
Basic construction in place to --verb-list
API change: render methods now take a cairo_t (not yet used)
* packaging/macosx/native-gtk/build-gtk.sh: Some updates to the script
to reflect the fact that gnome have switched from CVS to SVN. Also,
some changes to use the same script for building the Universal
Inkscape package.
to reflect the fact that gnome have switched from CVS to SVN. Also,
some changes to use the same script for building the Universal
Inkscape package.
render text in outline mode via cairo (still no endian-safe); factor out helper functions
Fix massive point queueing and tolerance issue
Optimize flood algorithm to only scan adjacent runs once
Re-add deleted preference setting for paint bucket tolerance
refactor sp_svg_transform_write to return a dynamically-allocated string
-- dynamic allocation won't hurt you, but buffer overruns will!
-- dynamic allocation won't hurt you, but buffer overruns will!
Fix layer transform issue
Fix issue with adding traced path to a layer with a transform
Change paint bucket tolerance toolbar to use percentages rather than color channel distance
Add paint bucket tolerance setting and toolbar widget to control tolerance
get rid of sp_repr_document_root and (commented) sp_repr_duplicate
increase padding
add paintbucket cursor
more comments
added new paint bucket; changed some colors; resaved with less precision and whitespace
Remove debug code from flood-context.cpp
more comments on cairo's and our own woes
Adjusted flood fill padding
Flood fill render area now taken from desktop view area and zoom level
dropping debian directory totally
add paint bucket
warn if unable to fill an unbounded area
rename flood to paintbucket, copyedit
move flood under calligraphic
add some obligatory shortcuts
Fix typo
get rid of sp_repr_document wrapper function
Check for GGO_UNHINTED not being defined
dead code removal
Update.
Typofix ("canceled").
Remove item handler from flood fill tool
Fix some memory issues with food fill
Fix g_free crash
Forgot to free some of the pixel storage
Add ability to flood fill areas of contiguous color
Prevent flood fills from leaking out of fill area
r14551@tres: ted | 2007-02-24 15:22:47 -0800
Merging in the latest libwpg and some code chages to make it work with
everything.
Merging in the latest libwpg and some code chages to make it work with
everything.
Add flood fill tool
Factor out bpath_to_Path
move a global to a class property; reduce the use of is_scrolling to prevent the lack-of-redraw issues
add 'hide all except selected' checkbox
some more refactoring
make selection_contains_both_clone_and_original reusable
fix crash when tracing with too small clones
fix crash when tracing with too small clones
guaranteed to pan the canvas all the way from mouse-press point to mouse-release point in any case, even if sometimes it fails to show the intermediate positions
Fixed crash when draw height was zero.
Disabling transientize callback - it's currently causing some data loss
when switching between windows.
when switching between windows.
When selecting 'none', the marker wasn't getting removed from the line.
This fixes that.
This fixes that.
add commented-out cairo stroke rendering; refactoring
factor out some cairo-related helpers; add cairo clipping - commented out for now due to a cairo bug
Adjusting how items in the marker menu are selected, so the recently
selected marker at the top of the menu shows as the selection.
selected marker at the top of the menu shows as the selection.
Fix initialisation problem.
Fixes issue where a modified stock marker's preview image appears in the
marker dropdown with the same preview image as the unmodified stock
marker. This adds the source document's uri to the cache index.
marker dropdown with the same preview image as the unmodified stock
marker. This adds the source document's uri to the cache index.
Adding handler for switching documents, to regenerate the marker menus
fix 1662589: increase blur margins so that the cutoff is not noticeable (now at 2.4*radius, it is at 3e-3; previously at 2*radius it was 1.8e-2 and quite visible)
Crudely improve check-markup for { markup, fix translations with bugs in that area (dz and zh_TW).
remove old *.mingw files
remove old *.mingw files
Improve null check
fully drop fixed_g_ascii_strtod hack
Completes fix for 980157. Markers in current document's defs is now
shown in top of the marker dropdown menu. When you select a marker from
the dropdown, all three menus are updated to show what you selected at
the top.
Note that to implement this, a marker cache is introduced, because
otherwise the regeneration of the menus would be noticeably too slow.
There is still a modest pause during the regeneration, but it's fairly
reasonable on my machine.
shown in top of the marker dropdown menu. When you select a marker from
the dropdown, all three menus are updated to show what you selected at
the top.
Note that to implement this, a marker cache is introduced, because
otherwise the regeneration of the menus would be noticeably too slow.
There is still a modest pause during the regeneration, but it's fairly
reasonable on my machine.
Moving cache management code from .h to .cpp.
Making the marker dropdowns static so they can be updated.
Making the marker dropdowns static so they can be updated.
Moving functions from svg_preview_cache.h to a new .cpp
Partial fix for bug 980157.
This commit disables the code that prevents stock markers from appearing
at the top of the marker menu, as discussed in the bug report.
Additionally, this patch factors out the pixmap rendering code, which is
nearly identical to the rendering code in icon.cpp (and perhaps
elsewhere).
This commit disables the code that prevents stock markers from appearing
at the top of the marker menu, as discussed in the bug report.
Additionally, this patch factors out the pixmap rendering code, which is
nearly identical to the rendering code in icon.cpp (and perhaps
elsewhere).
fix the order of searching for linked image file to match that of inkscape itself: first href, only if it fails absref
Cleaned up DOS line ends that had snuck in.
Fixed unintialized variables and minor misc warnings
now at least the first opening of the first gtkmm dialog opens it in-place without jumping. still can't make it work for all cases.
speed up default scrolling a bit
make sure scrolling acceleration does not hiccup on complex slow-rendering drawings by reducing time interval between events by the time of painting the slowest buffer on the canvas
smooth panning: suppress painting for so long as scrolling events come one after the other; add comments on tiles array; make tile size a define; simplify work with tiles in paint(); remove redundant redraw requests in sp_canvas_scroll_to as it gets requested by expose event anyway
New dir for UI cache management code
remove old workaround for a bug which seems to be fixed in latest gtk
added forward decl to fixed_g_ascii_strtod to fix compiler issue on WinXP
relocated unix print dialog outside of extension call path
standardize converting from number to string and back
reverse the order of markers, because since rev 13483 stroke-style.cpp reads markers in direct order, not reversed as before
Finishing fleshing out some design documentation for markers
separate crash handlers, allow "in place" crashes to happen
17 years agoa bunch of cairo fixmes in anticipation of a more direct rendering with less intermed...
a bunch of cairo fixmes in anticipation of a more direct rendering with less intermediate buffers and unpacked 32bit RGBA throughout