2000-05-16 Tor Lillqvist <tml@iki.fi>
* libgimp/gimpenv.c (gimp_path_parse)
* app/gimprc.c (transform_path): Don't try to handle tilde (~) at
all on Windows. The tilde is used when mangling long file names
into 8.3 form, and some people have a HOME environment variable
that contains this kind of 8.3 path with a tilde. This causes
interesting effects if we expand a tilde in some gimprc variable
with $HOME, which contains a tilde, which we expand with $HOME,
etc.
2000-05-14 Michael Natterer <mitch@gimp.org>
After a weekend of extensive mail exchange with the signal crew,
we finally found that #2742 is not a Gimp bug but a real OSF/1
bug. OSF/1 does _not_ reliably restart read() and write() calls
on certain slow devices (pipes).
* libgimp/gimpwire.[ch]: guard all read/write calls on the wire
with loops checking for EINTR. This has to be done "manually"
as glib's GIOChannels return G_IO_ERROR_UNKNOWN on the occurence
of EINTR (which is a bug, too).
s/int/gboolean and minor cleanups while I was on it (not changing
any logic).
This fix depends on the current (broken) state of the GIOChannel
implementation and is scheduled for removal as soon as glib
behaves nicely here.
Left SA_RESTART there for the moment in app/main.c. See it as
defensive programming or just my fear to change two #2742-related
places at the same time. We might choose to remove SA_RESTART
later. Many thanks again to Austin, Garry and Tim.
2000-05-12 Michael Natterer <mitch@gimp.org>
* libgimp/gimp.c (gimp_plugin_io_error_handler): fixed a
compiler warning about a missing return.
Slightly changed the messages that gimp<->plugin use different
protocol versions and a s/sprintf/g_snprintf/.
2000-05-10 Michael Natterer <mitch@gimp.org>
Another try to get the signal/dead child recovery stuff right.
Could the brave signal crew (TM) (Austin, Garry, Raphael, Tim)
please do bad tests to the new code? I removed all strange
constants (SA_NODEFER etc.) and used only glib and POSIX stuff.
* app/main.c
* libgimp/gimp.c:
- Call gimp_signal_private() with no flags to enforce a proper
sigaction() behaviour (block signals while handler is active).
- Removed the reentrancy guards from the handlers.
- Renamed the handlers.
- Ignore SIGPIPE in the app and in plugins.
- Re-introduced the SIGCHLD handler because it should work
now. Also added a SIGCHLD handler to libgimp/gimp.c.
* app/errors.c
* libgimp/gimp.c: in the signal handler, unblock all signals
with sigprocmask() before calling g_on_error_query() because
gdb otherwise inherits the blocked signals and does nothing.
Wrapped the statements with "if (TRUE) { }" blocks so it's
easy to make the stack trace a command line option.
* app/plug_in.c
* libgimp/gimp.c: listen for G_IO_ERR and G_IO_HUP on the read
channels. In the app, pop up an error message and clean up the
plugin. In plugins, clean up and exit.
* libgimp/gimpwire.c: removed the "plug-in chrashed?" message
and print the program's name with all error messages.
* plug-ins/helpbrowser/helpbrowser.c: typo.
2000-05-04 Tor Lillqvist <tml@iki.fi>
* gimprc.win32: Replace "color-cube" by "min-colors".
* */makefile.{cygwin,msc}: Updates. Install to /install/gimp.
* libgimp/gimp.def: Add gimp_min_colors.
* modules/cdisplay_highcontrast.c: Include gimpmath.h instead of
math.h. Use G_PI instead of M_PI.
2000-05-02 Michael Natterer <mitch@gimp.org>
* app/main.c
* libgimp/gimp.c: as noticed by Garry, Tim, Raphael and probably
many others, my latest signal change introduced a Linuxism
which doesn't compile on any other UNIX system, eek.
s/SA_NOMASK/SA_NODEFER/ makes it compile and is hopefully
the last chapter in this saga.
2000-05-01 Michael Natterer <mitch@gimp.org>
* libgimp/gimpsignal.[ch]
* libgimp/gimpui.c: don't panic, this is not another signal handling
change but only a namespace cleanup to make the html documentation
better. Also added documentation for gimp_ui_init().
2000-05-01 Michael Natterer <mitch@gimp.org>
* libgimp/libgimp-decl.txt
* libgimp/libgimp-sections.txt
* libgimp/tmpl/gimp.sgml
* libgimp/tmpl/gimpsignal.sgml
* libgimp/tmpl/gimpui.sgml
* libgimp/tmpl/libgimp-unused.sgml: updated.
2000-05-01 Michael Natterer <mitch@gimp.org>
* app/main.c
* libgimp/gimp.c: again... _do_ use gimp_signal_private() but
force it to behave like signal(). While on Linux it seems to
be ok to use both signal() and sigaction() (because signal() is
built on top of sigaction()), doing the same on other systems
is totally broken code. Thanks to Garry for pointing out this
portability issue.
* app/errors.c: s/vprintf/g_strdup_vprintf/
* libgimp/gimpsignal.[ch]: code formating paranoia.
2000-04-30 Michael Natterer <mitch@gimp.org>
* gimprc.in
* app/gimprc.[ch]: replaced "color-cube" by "min-colors".
* app/app_procs.c: read unitrc/gimprc before displaying the splash.
* app/colormaps.c: set min_colors and install_cmap before
initializing GdkRGB.
* app/gimprc.[ch]
* app/gimpunit.c: don't call the splash's progress_update function.
* app/plug_in.c: pass min_colors instead of color_cube to plugins.
* app/preferences_dialog.c: widget for min_colors.
* libgimp/gimp.[ch]: s/color_cube/min_colors/ but left
gimp_color_cube() there for source level compatibility.
* libgimp/gimpprotocol.[ch]: changed the GPConfig message
accordinly and increased the gimp protocol version number because
the change breaks binary compatibility. Also actually pass the
use_xshm variable over the wire (was only in the GPConfig struct
before).
Was it the right thing to do to increase the version number??
* libgimp/gimpui.c (gimp_ui_init): use the same code as the app
for initializing GdkRGB. Never explicitly activate Gdk's SHM
usage (only switch it off).
* app/main.c
* libgimp/gimp.c: reverted the handling of all signals except
SIGCHLD back to plain old signal() because those signals are
fatal anyway and sigaction() as used by gimp_signal_*() made
debugging (stacktrace) impossible.
* plug-ins/AlienMap/AlienMap.c
* plug-ins/AlienMap2/AlienMap2.c
* plug-ins/FractalExplorer/Dialogs.c
* plug-ins/bmp/bmp.c
* plug-ins/borderaverage/borderaverage.c
* plug-ins/dbbrowser/dbbrowser.c
* plug-ins/faxg3/faxg3.c
* plug-ins/fits/fits.c
* plug-ins/flame/flame.c
* plug-ins/fp/fp.c
* plug-ins/fp/fp_gtk.c
* plug-ins/gdyntext/Makefile.am
* plug-ins/gdyntext/gdyntext_ui.c
* plug-ins/gfig/gfig.c
* plug-ins/gflare/gflare.c
* plug-ins/gfli/gfli.c
* plug-ins/gimpressionist/gimpressionist.c
* plug-ins/helpbrowser/helpbrowser.c
* plug-ins/ifscompose/ifscompose.c
* plug-ins/imagemap/Makefile.am
* plug-ins/imagemap/imap_main.c
* plug-ins/maze/maze_face.c
* plug-ins/mosaic/mosaic.c
* plug-ins/pagecurl/pagecurl.c
* plug-ins/print/print.c
* plug-ins/rcm/rcm_dialog.c
* plug-ins/script-fu/script-fu-console.c
* plug-ins/script-fu/script-fu-scripts.c
* plug-ins/script-fu/script-fu-server.c
* plug-ins/sel2path/Makefile.am
* plug-ins/sel2path/sel2path.c
* plug-ins/sgi/sgi.c
* plug-ins/sinus/sinus.c
* plug-ins/struc/struc.c
* plug-ins/webbrowser/webbrowser.c
* plug-ins/winsnap/winsnap.c
* plug-ins/xjt/xjt.c: use gimp_ui_init().
* plug-ins/Lighting/lighting_ui.c
* plug-ins/MapObject/mapobject_ui.c: only switch Gdk SHM usage off,
never on. Don't use gimp_ui_init() here because of libgck.
2000-04-30 Garry R. Osgood <gosgood@idt.net>
* app/main.c
* libgimp/gimpsignal.c [Documentation only]
Patches arising from Tim Mooney to fix#2742
which should (hopefully) die at this point.
removes on_sig_chld() from app/main() which
offers no meaningful functionality in light
of SA_NOCHLDSTOP.
2000-04-26 Michael Natterer <mitch@gimp.org>
* app/free_select.c: fill empty "default" with a "break" (#9431).
* libgimp/gimpenv.c: g_path_is_absolute wants a parameter (#9400).
2000-04-23 Tor Lillqvist <tml@iki.fi>
* libgimp/gimpui.def: Add gimp_ui_init.
* libgimp/makefile.{cygwin,msc} (gimpui_OBJECTS): Add gimpui.
* app/makefile.{cygwin,msc}
* plug-ins/makefile.{cygwin,msc}: New installation directory.
* app/datafiles.c (is_script): Must not g_free() value returned by
g_getenv().
* plug-ins/gdyntext/font_selection.c: Don't include gdkx.h with
GTk+ 1.3.
* libgimp/Makefile.am: gimpsignal.[ch] needs to go into libgimp,
since it is used in gimp.c
* libgimp/gimpsignal.c: use g_error instead of gimp_fatal_error
* plug-ins/common/png.c: typo fix for Nick's previous commit
-Yosh
2000-04-21 Garry R. Osgood <gosgood@idt.net>
* app/main.c
* libgimp/Makefile.am
* libgimp/gimp.c
* libgimp/gimpsignal.c (New )
* libgimp/gimpsignal.h (New )
Patch by Austin Donnelly <austin@gimp.org> to
address #2742. (I took the liberty of moving it to
libgimp for general consumption, & fixed a typo)
Thanks due to Tim Mooney
<mooney@dogbert.cc.ndsu.nodak.edu> for testing
this on various flavors of OSF/1 boxes where
#2742 seemed especially common.
2000-04-20 Michael Natterer <mitch@gimp.org>
* libgimp/gimpfileselection.[ch]: using the GimpPixmap here was a
bad idea because showing/hiding the pixmap widgets is no solution
(somebody may call gtk_widget_show_all() and we suddenly see both
pixmaps, eek).
Reverted it to use GtkPixmap and gtk_pixmap_set().
2000-04-19 Michael Natterer <mitch@gimp.org>
* app/bucket_fill.c
* app/by_color_select.c
* app/fuzzy_select.c
* app/preferences_dialog.c: made the "default_threshold" gimprc
variable work as advertized:
- initialize the thresholds with it.
- use it for "Reset".
- added a widget to the "Tool Options" preferences page.
- noticed that the "Reset" button of "By Color Select" doesn't
behave like all the other "Reset" buttons and changed it to
reset the ui, not the selection.
(There is now a "None" button and because it was so trivial, I
couldn't resist to add "All" and "Invert" buttons, too)
* libgimp/Makefile.am
* libgimp/gimpui.c: new file.
* libgimp/gimpui.h: new function gimp_ui_init() which will be
called by all plugins which have a ui (not only by those with a
preview because plugins should always follow gimp's colormap
installation policy).
Could someone please check if the FIXME stuff in the function
is the right thing to do (TM). Does GdkRGB allocate the correct
colors for the widgets in all cases or do we have to find another
way to ensure this across processes (gtk instances)?
2000-04-11 Tor Lillqvist <tml@iki.fi>
* libgimp/gimpenv.c (gimp_path_get_user_writable_dir): The stat()
function in the Win32 C runtime doesn't like if directory names
have an extra G_DIR_SEPARATOR suffixed. Remove such before calling
stat(). Those extra (back)slashes are added by for instance
gimp_path_parse(). Document that feature of gimp_path_parse().
* libgimp/gimputils.c (gimp_strescape): Correct documentation for
gimp_strescape(). The semantics of the exceptions parameter was
documented backwards. Clarify what the function does.
* plug-ins/gfig/gfig.c (plug_in_parse_gfig_path)
* plug-ins/FractalExplorer/FractalExplorer.c
(plug_in_parse_fractalexplorer_path)
* plug-ins/gflare/gflare.c (plug_in_parse_gflare_path,
gflare_save): Use painstakingly correct platform-specific
directory and search path separators, and escaped with
gimp_strescape when displayed as part of a gimprc entry.
* app/gimphelp.c: No debug messages on Win32, thanks.
2000-04-10 Tor Lillqvist <tml@iki.fi>
* README: Correct the info about Win32 mailing lists.
* gimprc.win32: Add perfect-mouse setting.
* libgimp/gimp.h
* app/main.c: Guard against redefinition of _stdcall.
* libgimp/gimpmath.h: Define macros ISNAN and ISINF. On Win32 use
_isnan() and _fpclass() from <float.h>, on other platforms use
isnan() and isinf(). We really should test for isnan and isinf in the
configure script.
* tools/pdbgen/pdb/gimage.pdb
* app/gimage_cmds.c: Include gimpmath.h, use above macros instead
of using the nonportable isnan and isinf.
* app/gdisplay_ops.c: Do use the old gtk_widget_hide/show hack on
Win32.
* libgimp/makefile.cygwin
* libgimp/makefile.msc
* app/makefile.cygwin
* app/makefile.msc
* plug-ins/makefile.cygwin
* plug-ins/makefile.msc: Updates.
* plug-ins/gfig/gfig.c (plug_in_parse_gfig_path): Simplify warning
message on Win32. We could use G_DIR_SEPARATOR in the sample
gimprc line, but then we would have to call g_strescape to quote
it, argh. Win32 users use prebuilt versions with a correct gimprc
file anyway.
(create_file_selection): Use g_get_tmp_dir().
(gfig_update_stat_labels): Guard against possibly NULL value from
g_get_home_dir().
* libgimp/gimpexport.c: implemented the changes Nick Lamb
suggested: If the background layer has no alpha channel,
suggest "Flatten" instead of "Merge Visible Layers".
* app/fileops.c: In the MRU list and the doc_index always store
the absolute path to images, so you can open files from the
command-line and gimp will find them later if started from another
dir. Added tooltips that show the full path.
* app/menus.c: fixed a memleak (list entries were not freed on
removal from the MRU list).
* app/main.c
* app/user_install.[ch]: renamed functions and variables to match
the new filename.
--Sven
2000-04-06 Michael Natterer <mitch@gimp.org>
* app/commands.c
* app/layers_dialog.c
* app/resize.[ch]: behave like a widget:
removed resize_widget_free() and call
gtk_widget_destroy(resize->resize_shell) instead.
* app/interface.c: s/0/FALSE/
* app/color_area.c
* libgimp/gimpchainbutton.c: show the correct background also for
engine driven themes. The chainbutton's lines still don't draw
with the pixmap theme, which is probably a bug of the theme's
draw_polygon method.
2000-03-31 Michael Natterer <mitch@gimp.org>
* app/gimpdnd.[ch]: minor cleanups.
* libgimp/gimpwidgets.c: don't set a border width for the
radio_group's vbox if no frame around it was requested.
* plug-ins/common/sample_colorize.c: add a separator between
image and sample.
* plug-ins/common/iwarp.c: use the libgimp ui functions, removed
callbacks, a global s/gfloat/gdouble/, replaced the booleans
toggling the transform mode with a single variable containing an
enum value, indentation.
2000-03-08 Tor Lillqvist <tml@iki.fi>
* app/cursorutil.c (gtkutil_compress_motion)
* app/edit_selection.c (process_event_queue_keys): Guard against
gdk_event_get returning NULL (which can happen at least on Win32).
* libgimp/gimp.def: Add a couple of new entry points.
* plug-ins/makefile.{cygwin,msc}: Update according to the source
file changes. Fix some typos in the .msc file.
Fixes by Hans Breuer:
* app/resize.c: Add some more includes.
* libgimp/gimpenv.c
* plug-ins/gflare/gflare.c: Win32 header lossage fixup.
2000-03-07 Michael Natterer <mitch@gimp.org>
* app/by_color_select.c: immediate cursor_update feedback on
modifier events.
* libgimp/gimpwidgets.c: one more s/private_tip/help_data/
2000-03-03 Michael Natterer <mitch@gimp.org>
* app/crop.c: s/"Only"/"only"/
* app/iscissors.c: one more cursor_update fix. This time I don't
claim that it's _really_ correct.
* app/tool_options.c: don't add a separator after
opacity/paint_mode if a paint pressure options box follows.
* cursors/bad.xbm
* cursors/bad_mask.xbm: made it FAT (no need to use thin lines
which show as much as possible of the image below because the
cursor indicates that no operation is possible).
* libgimp/gimpprotocol.[ch]: s/int/gboolean/ where appopriate,
indentation paranoia.
_gp_*_read(): free the already allocated parts of the message if
reading a subsequent part fails. These cleanups will probably occur
shortly before the process crashes, but at least they make the
search for real leaks easier.
* plug-ins/common/uniteditor.c: some more tooltips.
* plug-ins/common/xbm.c: store the image comment in the
"gimp-comment" parasite and the hot spot in the new "hot-spot"
parasite. Added ui for entering the hot spot.
* docs/parasites.txt: documented the new "hot-spot" parasite.
2000-03-03 Michael Natterer <mitch@gimp.org>
* libgimp/gimpprotocol.c (_gp_proc_return_destroy): fixed a
memleak: free the message's name before freeing the message itself.
2000-02-28 Michael Natterer <mitch@gimp.org>
* libgimp/gimpfileselection.c
* libgimp/gimpunitmenu.c: behave like other dialogs (close) when
pressing "OK" or "Cancel".
* libgimp/stdplugins-intl.h
* plug-ins/script-fu/script-fu-intl.h: bind the "gimp-libgimp"
textdomain also if HAVE_LC_MESSAGES is undefined.
* plug-ins/print/print-ps.c
* plug-ins/print/print.c: applied a patch from Robert L Kravitz
<rlk@alum.mit.edu> which adds a "Center" button, activates all
text entries and corrects the rounding behaviour. And a minor fix
for the PS driver.
* plug-ins/dbbrowser/Makefile.am
* plug-ins/dbbrowser/dbbrowser.h: removed.
* plug-ins/dbbrowser/dbbrowser.c
* plug-ins/dbbrowser/dbbrowser_utils.c
* plug-ins/dbbrowser/dbbrowser_utils.h
* plug-ins/common/plugindetails.c: ui and code cleanup.
libgimp ui stuff and a proper include structure for the dbbrowser.
* plug-ins/script-fu/script-fu-console.c: include
"plug-ins/dbbrowser/dbbrowser_utils.h".