GNOME Bugzilla – Bug 513156
Get rid of libgnome(ui)
Last modified: 2009-06-24 20:42:32 UTC
After 2.4 relase i'll try to remove the GnomeApp stuff and use GtkWindow instead.
Created attachment 104102 [details] [review] patch This is the first patch. It is working without bonobo dockers. We have to discuss if we want use the bonobo dockers or maybe we can port the monodevelop docker to C. (Maybe it is the bonobo docker ported to C# i don't know)
Created attachment 104106 [details] [review] patch Maybe we could implement a toolbar system based in gdl. What do you think?
Created attachment 107138 [details] [review] plugins Apply it in plugins directory. About this files, shouldn't call the finalize signal? document-manager/plugin.c: finalize? editor/plugin.c: finalze? file-loader/plugin.c: finalize? file-wizard/plugin.c: finalize?
Changes summary... The patch looks good, please add the missing calls to the parent finalize in case they are missing. Thanks!
Created attachment 107145 [details] [review] plugins This should remove all GNOME_CALL* calls in plugins
Created attachment 107234 [details] [review] project-import This patch uses GtkAssistant instead of GnomeDruid.
(In reply to comment #5) > Created an attachment (id=107145) [edit] > plugins > > This should remove all GNOME_CALL* calls in plugins > Thanks for the patch it has been committed.
(In reply to comment #6) > Created an attachment (id=107234) [edit] > project-import > > This patch uses GtkAssistant instead of GnomeDruid. > Why did you not use glade for the widgets? Is it because glade-3 has no support for GtkAssistant?
I just checked glade-3 and it does have gtkassistant widget. Could fix the patch to use glade file instead?
(In reply to comment #2) > Created an attachment (id=104106) [edit] > patch > Thanks for the path, but it needs some changes: 1) Please keep the changes relevant and avoid style-only changes. There are lots of them in there. 2) Do not comment dead codes, remove them. SVN already takes care of that. 3) Could you please check if custom menu short-cut still work without GnomeApp? IIRC, GnomeApp does accel save/load and I don't recall how Anjuta handles is separately. > Maybe we could implement a toolbar system based in gdl. What do you think? > We thought about it at some point, but didn't really go any further. If libbonoboui is no more maintained and if it has other unnecessary drags for us, we could consider copying the toolbar dock code and have it in gdl.
(In reply to comment #9) > I just checked glade-3 and it does have gtkassistant widget. Could fix the > patch to use glade file instead? > I tried it with glade and i can't because libglade doesn't have support for GtkAssistant.
(In reply to comment #10) > (In reply to comment #2) > > Created an attachment (id=104106) [edit] > > patch > > > Thanks for the path, but it needs some changes: > > 1) Please keep the changes relevant and avoid style-only changes. There are > lots of them in there. Sorry about that. I am not going to do it anymore. > > 2) Do not comment dead codes, remove them. SVN already takes care of that. > > 3) Could you please check if custom menu short-cut still work without GnomeApp? > IIRC, GnomeApp does accel save/load and I don't recall how Anjuta handles is > separately. To do this we can do it as gedit or gtranslator do. Take a look to: http://svn.gnome.org/viewvc/gtranslator/trunk/src/application.c?r1=3464&r2=3478 > > > Maybe we could implement a toolbar system based in gdl. What do you think? > > > We thought about it at some point, but didn't really go any further. If > libbonoboui is no more maintained and if it has other unnecessary drags for us, > we could consider copying the toolbar dock code and have it in gdl. > Maybe we could port the monodevelop code to do this. And I think libbonoboui is going to become deprecated at some point.
Created attachment 107395 [details] [review] project import with changelog As you request. Now i'll try to remove the gnome-vfs stuff of that bug.
Created attachment 107400 [details] [review] project import I changed the title of the second page to "Summary" as jhs said.
(In reply to comment #13) > Created an attachment (id=107395) [edit] > project import with changelog > > As you request. Now i'll try to remove the gnome-vfs stuff of that bug. > This caused segfault for at the end of import. Could you please confirm it works? I tried with 2 different project imports and it crashed in both (repeated several times).
(In reply to comment #15) > (In reply to comment #13) > > Created an attachment (id=107395) [edit] > > project import with changelog > > > > As you request. Now i'll try to remove the gnome-vfs stuff of that bug. > > > This caused segfault for at the end of import. Could you please confirm it > works? I tried with 2 different project imports and it crashed in both > (repeated several times). > It seems that the problem is in the gtk_widget_destroy (pi->assistant); but i don't know why I can't destroy the assistant. Any idea?
Created attachment 107567 [details] [review] gtkwindow patch This patch adds support for toolbars using bonobo. Now I just need to add the accels support. This is really easy to do but I need to know where do you want to save the accels stuff. For example "~/.gnome2/anjuta.accels" Please let me now the url you want to use to save the accels data.
Created attachment 107683 [details] [review] gtkwindow port completed This patch should fix everything. But it needs revision because there are a few things that should be removed or changed. Take a look to the FIXME comments.
Created attachment 107814 [details] [review] url This should remove the gnome_url_show func.
Created attachment 107816 [details] [review] help And this one should remove the gnome help func.
Created attachment 107819 [details] [review] libanjuta Makefile.am It removes a duplicated variable and it adds the DATADIR variable.
(In reply to comment #16) > (In reply to comment #15) > > (In reply to comment #13) > > > Created an attachment (id=107395) [edit] > > > project import with changelog > > > > > > As you request. Now i'll try to remove the gnome-vfs stuff of that bug. > > > > > This caused segfault for at the end of import. Could you please confirm it > > works? I tried with 2 different project imports and it crashed in both > > (repeated several times). > > > It seems that the problem is in the gtk_widget_destroy (pi->assistant); but i > don't know why I can't destroy the assistant. Any idea? > I fixed the crash and committed your patch. Thanks.
(In reply to comment #19) > Created an attachment (id=107814) [edit] > url > > This should remove the gnome_url_show func. > Hmm, looks like you're overusing the array. Would it be possible to just do: char *open[3]; open[0] = "xdg-open"; open[1] = url; open[2] = NULL; ... Regards, -Naba
(In reply to comment #22) > (In reply to comment #16) > > (In reply to comment #15) > > > (In reply to comment #13) > > > > Created an attachment (id=107395) [edit] > > > > project import with changelog > > > > > > > > As you request. Now i'll try to remove the gnome-vfs stuff of that bug. > > > > > > > This caused segfault for at the end of import. Could you please confirm it > > > works? I tried with 2 different project imports and it crashed in both > > > (repeated several times). > > > > > It seems that the problem is in the gtk_widget_destroy (pi->assistant); but i > > don't know why I can't destroy the assistant. Any idea? > > > I fixed the crash and committed your patch. Thanks. > First thing. You forgot a GnomeDruid variable in the close callback. Second. How did you fix it?
(In reply to comment #23) > (In reply to comment #19) > > Created an attachment (id=107814) [edit] > > url > > > > This should remove the gnome_url_show func. > > > Hmm, looks like you're overusing the array. Would it be possible to just do: > > char *open[3]; > open[0] = "xdg-open"; > open[1] = url; > open[2] = NULL; > ... > > Regards, > -Naba > Yeah, you are right, it was thinked to use more programs than xdg-open, but as xdg-open is a freedesktop standard I only used it. So i'll rewrite the patch doing what you said.
Created attachment 108048 [details] [review] gnome druid variable Ok, you fixed it with the close signal. Sorry I didn't see it. This patch should remove the GnomeDruid variable you forgot to remove.
Created attachment 108049 [details] [review] url2 This should be what you want.
Created attachment 108050 [details] [review] url3 Sorry, I forgot to remove the #include <gnome.h>
(In reply to comment #26) > Created an attachment (id=108048) [edit] > gnome druid variable > > Ok, you fixed it with the close signal. Sorry I didn't see it. > This patch should remove the GnomeDruid variable you forgot to remove. > Committed. Thanks for point it out.
(In reply to comment #28) > Created an attachment (id=108050) [edit] > url3 > > Sorry, I forgot to remove the #include <gnome.h> > Committed, thanks.
(In reply to comment #20) > Created an attachment (id=107816) [edit] > help > > And this one should remove the gnome help func. > Committed. (In reply to comment #21) > Created an attachment (id=107819) [edit] > libanjuta Makefile.am > > It removes a duplicated variable and it adds the DATADIR variable. > Committed.
(In reply to comment #18) > Created an attachment (id=107683) [edit] > gtkwindow port completed > > This patch should fix everything. But it needs revision because there are a few > things that should be removed or changed. Take a look to the FIXME comments. > +/* + * Accels + */ +static gchar * +get_accel_file (void) +{ + gchar *anjuta_dir = NULL; + + anjuta_dir = anjuta_utils_get_user_config_dir (); + + if (anjuta_dir != NULL) + { + return g_build_filename (anjuta_dir, + "anjuta-accels", + NULL); anjuta_dir leaks here. Couple of more things: 1. Please always fillup a changelog in your patches. 2. Please always take the diff from top directory. This one is fine, but all your other patches were not. 3. Please always make sure not to include 'whitespace only changes'. It makes it harder to review. Not to mention most of your style-only fixes doesn't really correspond to anjuta style. i.e. the function args indentation isn't right. 4. please read http://anjuta.org/hacking Please fix the patch as mentioned and submit again. Thanks. Regards, -Naba
Created attachment 108091 [details] [review] GtkWindow port As I said before, please have a look at FIXME comments.
Created attachment 108112 [details] [review] GtkWindow port 2 Added key bindings support.
Hi Ignacio, Thanks for the updated patch. But it needs more fixes: 1) You still have lot of 'style-only' changes. Please remove them all. 2) Your changes have tabsize 8, but anjuta uses tabsize 4 (using emacs should set it automatically for that file). 3) Your function args indentation different from anjuta's style (i.e. don't break line on each argument. break only for the purpose of trying to fit in 80 cols). It's clearly shown in sample code in HACKING file. 3) We can't start using gio, so please create the directory using something else. 4) The directory check/creation should be done in save_accels() function, just before the file is saved. This was one of your FIXMEs. 5) The other FIXME was about the accels. I think, your change is alright. The accels sync should be happening correctly within gtkwindow. But I can only try this for sure after your patch is applied. Thanks. Regards, -Naba
One more thing: 6. It should be anjuta_util_get_user_config_dir, not anjuta_utils_... (In reply to comment #35) > 4) The directory check/creation should be done in save_accels() function, just > before the file is saved. This was one of your FIXMEs. > I am sorry. That's not the case. I think the config directory should be checked and created in anjuta_util_get_user_config_dir(), so that whoever calls it first ensures the directory is created.
Created attachment 108287 [details] [review] GtkWindow port 3 I think that now I used the anjuta style. Sorry about that.
(In reply to comment #37) > Created an attachment (id=108287) [edit] > GtkWindow port 3 > > I think that now I used the anjuta style. Sorry about that. > Thanks a lot. It's been committed. One small note. I think the save/load_accels() should be in anjuta-ui.c and called whenever accel changes.
(In reply to comment #38) > (In reply to comment #37) > > Created an attachment (id=108287) [edit] > > GtkWindow port 3 > > > > I think that now I used the anjuta style. Sorry about that. > > > Thanks a lot. It's been committed. One small note. I think the > save/load_accels() should be in anjuta-ui.c and called whenever accel changes. > Ok. I'll have a look at soon as possible at anjuta-ui.c code to merge the save/load_accels() there.
There is a problem in trunk now: The toolbar position are not remembered correctly and it is very difficult to aling toolbars correctly. Test case: - Start Anjuta - Align all toolbars in one line - Close/Restart Anjuta - Toolbars are no longer in one line
(In reply to comment #40) > There is a problem in trunk now: The toolbar position are not remembered > correctly and it is very difficult to aling toolbars correctly. > > Test case: > - Start Anjuta > - Align all toolbars in one line > - Close/Restart Anjuta > - Toolbars are no longer in one line > You are right, I'll have a look as soon as I have some time.
Created attachment 110028 [details] [review] accels This patch moves the accel funcs as Naba requested.
Created attachment 111072 [details] [review] project import alignment I attached this patch here to don't create a bug for a 2 lines patch. This patch changes the alignment of the label to 0.0
Thanks for the patches. Both have been committed. What 'libgnome(ui)' portions are remaining now?
I think we need to port the project-wizard plugin (seb is working on this, or he will) and make something with the gnome-session stuff (This is I already don't know how to do it). I have to port the main func too removing the GnomeProgram stuff and then should be everything without libgnome(ui).
A quick grep also revealed some more. Many of them are quite straightforward: ./libanjuta/anjuta-launcher.c:#include <libgnome/gnome-macros.h> ./libanjuta/anjuta-preferences-dialog.c:#include <libgnome/gnome-macros.h> ./libanjuta/anjuta-session.c:#include <libgnome/gnome-config.h> ./libanjuta/anjuta-ui.c:#include <libgnome/gnome-macros.h> ./libanjuta/cell-renderer-captioned-image.c:#include <libgnome/gnome-macros.h> ./libanjuta/e-splash.c:#include <libgnome/gnome-macros.h> ./plugins/debug-manager/data_view.c:#include <libgnome/gnome-i18n.h> ./plugins/debug-manager/debug_tree.c:#include <libgnome/gnome-i18n.h> ./plugins/debug-manager/info.c:#include <libgnome/gnome-i18n.h> ./plugins/debug-manager/sparse_view.c:#include <libgnome/gnome-i18n.h> ./plugins/gdb/parser.c:#include <libgnome/gnome-i18n.h> ./plugins/project-wizard/druid.c:#include <libgnome/gnome-i18n.h> ./plugins/project-wizard/property.c:#include <libgnome/gnome-i18n.h> ./plugins/sourceview/anjuta-utils.c:#include <libgnome/gnome-url.h> ./plugins/symbol-browser/an_symbol.c:#include <libgnome/gnome-macros.h> ./plugins/symbol-browser/an_symbol_iter.c:#include <libgnome/gnome-macros.h> ./plugins/symbol-browser/an_symbol_view.c:#include <libgnomeui/gnome-stock-icons.h> ./plugins/symbol-browser/an_symbol_view.c:#include <libgnome/gnome-macros.h> ./plugins/symbol-db/symbol-db-view-search.c:#include <libgnome/gnome-i18n.h> ./plugins/tools/editor.c:#include <libgnomeui/libgnomeui.h> ./plugins/valgrind/plugin.c:#include <libgnome/gnome-util.h> ./plugins/valgrind/plugin.c:#include <libgnomeui/gnome-about.h> ./plugins/valgrind/process.c:#include <libgnome/gnome-i18n.h> ./plugins/valgrind/vgcachegrindprefs.c:#include <libgnome/gnome-i18n.h> ./plugins/valgrind/vgdefaultview.c:#include <libgnome/gnome-i18n.h> ./plugins/valgrind/vggeneralprefs.c:#include <libgnome/gnome-i18n.h> ./plugins/valgrind/vghelgrindprefs.c:#include <libgnome/gnome-i18n.h> ./plugins/valgrind/vgmemcheckprefs.c:#include <libgnome/gnome-i18n.h> ./plugins/valgrind/vgrule-editor.c:#include <libgnome/gnome-i18n.h> ./plugins/valgrind/vgrule-list.c:#include <libgnome/gnome-i18n.h> ./plugins/valgrind/vgsearchbar.c:#include <libgnome/gnome-i18n.h> ./src/main.c:#include <libgnome/gnome-program.h>
And some more in .h files: ./libanjuta/anjuta-status.h:#include <libgnomeui/gnome-appbar.h> ./libanjuta/anjuta-status.h:#include <libgnome/gnome-macros.h> ./plugins/valgrind/vggeneralprefs.h:#include <libgnomeui/gnome-file-entry.h>
Ok, I'll try to remove the trivial ones this week but until two weeks I can't dedicate to many time because I am in examns.
Thanks, but you don't have to them now. Take your time with exams.
Some hints: libgnome/gnome-i18n.h can simple be replaced by glib/gi18n.h Some stuff from gnome-macros.h needs either porting to macros included in glib or some more explicit casting gnome-appbar needs real porting to GtkStatusBar but this also means we need to take care of the progressbar ourselfs because GtkStatusBar doesn't support a progressbar. GnomeFileEntry should be ported to GtkFileChooser.
Created attachment 112672 [details] [review] i18n glib/gi18n.h instead of libgnome/gnome-i18n.h
Created attachment 112678 [details] [review] template Removed GNOME_CALL from anjuta plugin template.
Created attachment 112729 [details] [review] anjuta-status This patch removes GnomeAppBar. It needs testing.
(In reply to comment #51) > Created an attachment (id=112672) [edit] > i18n > > glib/gi18n.h instead of libgnome/gnome-i18n.h > (In reply to comment #52) > Created an attachment (id=112678) [edit] > template > > Removed GNOME_CALL from anjuta plugin template. > Thanks for the patches. They are committed now.
(In reply to comment #53) > Created an attachment (id=112729) [edit] > anjuta-status > > This patch removes GnomeAppBar. It needs testing. > Committed the patch with some modifications. Thanks for it.
Created attachment 112770 [details] [review] libanjuta macros Removed all gnome-macros from libanjuta.
Thanks!
Created attachment 114237 [details] [review] mem leak I've put this patch to not open a bug for a one line patch.
Created attachment 114346 [details] [review] mark strings Another little patch
Created attachment 114519 [details] [review] another string Another string marked for translation in subversion.
For the project wizard, I miss a replacement for GnomeIconEntry. It appears in the property page, typically to choose an icon for your new program. Then, it's not possible to build a GtkAssistent from .glade file. I can use GtkBuilder but I don't know how to generate such file. I think, it's better to wait for anjuta 2.8 to fix this, as it's a quite big change which can break several stuff.
(In reply to comment #61) > For the project wizard, I miss a replacement for GnomeIconEntry. It appears in > the property page, typically to choose an icon for your new program. About this, maybe there is something we could use in libegg. > > Then, it's not possible to build a GtkAssistent from .glade file. I can use > GtkBuilder but I don't know how to generate such file. > > I think, it's better to wait for anjuta 2.8 to fix this, as it's a quite big > change which can break several stuff. >
Created attachment 119445 [details] [review] libgnome Another patch.
Hmm, does the statusbar stuff work atm? I looks like we have been mixing Gtk+ and Libgnome stuff there.
It seems it was mixed in that func only. I think it is not called by any func if not would make anjuta crash or something. The other funcs are ok.
Created attachment 121333 [details] [review] remove gnome-config
This patch does break session loading/saving for me. Can you please have a second look at it. Steps to reproduce: - Load project session - Close anjuta - Restart anjuta => The session is not restored correctly. More files are loaded than before, sometimes the jump-to-line does not work.
There is compatibility problem. GKeyFile doesn't allow set list separator "%%%". Should we save backward compatibility?
yes for loading we should maintain compatibility! thanks
I have fixed the tools plugin.
I'm working on the project wizard currently, trying to use GtkAssistant instead of GnomeDruid. It's not easy though, because GtkAssistant is not supported by glade (so I have to use GtkBuilder) and is quite different from GnomeDruid.
(In reply to comment #71) > I'm working on the project wizard currently, trying to use GtkAssistant instead > of GnomeDruid. It's not easy though, because GtkAssistant is not supported by > glade (so I have to use GtkBuilder) and is quite different from GnomeDruid. http://library.gnome.org/devel/gtk/stable/gtk-migrating-GtkAssistant.html If you run into issues it might be useful to enhance the migration documentation for the next hacker trying this for module $foo.
(In reply to comment #72) > If you run into issues it might be useful to enhance the migration > documentation for the next hacker trying this for module $foo. It's not really a problem of documentation. This wizard generates pages dynamically, the next page depends on the user input on the current one. So it was relying on "next" and "back" signal which are not available anymore.
Ah. Yeah, Gilles had the same problem when he took a look (read: not done) at porting Evolution: http://www.mail-archive.com/evolution-hackers@gnome.org/msg01635.html If you work on this, feel free to share implementation details in http://bugzilla.gnome.org/show_bug.cgi?id=567260 to avoid having another person taxing his brain about this problem. :)
I have committed the change on the project wizard plugin. I have replaced GnomeDruid by GtkAssistant and GnomeIconList by GtkIconView. I still need GnomeIconEntry though as there is currently no easy replacement for it. (In reply to comment #74) > If you work on this, feel free to share implementation details in > http://bugzilla.gnome.org/show_bug.cgi?id=567260 to avoid having another person > taxing his brain about this problem. :) I have added a small comment there.
Created attachment 127503 [details] [review] Remove GnomeDialog and GnomeMenuShell as well as GtkCList
Created attachment 127504 [details] [review] replace gnome_shell_execute
Thanks very much for your 2 patches Jens, I have committed them in svn trunk. I have fixed the IANJUTA_DEBUGGER_OK in your first patch, but sending signal is still not possible because the Info menu is disabled when you run the debugger. This needs some work but not in this bug.
Hi Yuriy (In reply to comment #68) > There is compatibility problem. GKeyFile doesn't allow set list separator > "%%%". Should we save backward compatibility? I'm agree with Johannes that backward compatibility should be preserved. But it would be useful to plan how to solve this situation. By example, we could add a version group, if it is not here we use the old Anjuta function to load the file else we use the GKeyFile function. And 1 or 2 release we can remove the old Anjuta function. If we go in this direction, we can perhaps directly switch from key file format to xml or something really different.
I don't know if these warnings come from the last patch or if I'm doing something wrong, but after launching Anjuta I see a bunch of messages (anjuta:26550): Gtk-WARNING **: SendSignal: missing action ActionDmaSignalSend (anjuta:26550): Gtk-WARNING **: UpdateSignalList: missing action ActionDmaSignalUpdate (anjuta:26550): Gtk-WARNING **: UpdateShlList: missing action ActionDmaSharedlibsUpdate (anjuta:26550): Gtk-WARNING **: SendSignal: missing action ActionDmaSignalSend (anjuta:26550): Gtk-WARNING **: UpdateSignalList: missing action ActionDmaSignalUpdate (anjuta:26550): Gtk-WARNING **: UpdateShlList: missing action ActionDmaSharedlibsUpdate (anjuta:26550): Gtk-WARNING **: SendSignal: missing action ActionDmaSignalSend (anjuta:26550): Gtk-WARNING **: UpdateSignalList: missing action ActionDmaSignalUpdate (anjuta:26550): Gtk-WARNING **: UpdateShlList: missing action ActionDmaSharedlibsUpdate (anjuta:26550): Gtk-WARNING **: SendSignal: missing action ActionDmaSignalSend (anjuta:26550): Gtk-WARNING **: UpdateSignalList: missing action ActionDmaSignalUpdate (anjuta:26550): Gtk-WARNING **: UpdateShlList: missing action ActionDmaSharedlibsUpdate (anjuta:26550): Gtk-WARNING **: SendSignal: missing action ActionDmaSignalSend (anjuta:26550): Gtk-WARNING **: UpdateSignalList: missing action ActionDmaSignalUpdate (anjuta:26550): Gtk-WARNING **: UpdateShlList: missing action ActionDmaSharedlibsUpdate (anjuta:26550): Gtk-WARNING **: SendSignal: missing action ActionDmaSignalSend (anjuta:26550): Gtk-WARNING **: UpdateSignalList: missing action ActionDmaSignalUpdate (anjuta:26550): Gtk-WARNING **: UpdateShlList: missing action ActionDmaSharedlibsUpdate (anjuta:26550): Gtk-WARNING **: SendSignal: missing action ActionDmaSignalSend (anjuta:26550): Gtk-WARNING **: UpdateSignalList: missing action ActionDmaSignalUpdate (anjuta:26550): Gtk-WARNING **: UpdateShlList: missing action ActionDmaSharedlibsUpdate (anjuta:26550): Gtk-WARNING **: SendSignal: missing action ActionDmaSignalSend (anjuta:26550): Gtk-WARNING **: UpdateSignalList: missing action ActionDmaSignalUpdate (anjuta:26550): Gtk-WARNING **: UpdateShlList: missing action ActionDmaSharedlibsUpdate
(In reply to comment #80) > I don't know if these warnings come from the last patch or if I'm doing > something wrong, but after launching Anjuta I see a bunch of messages Yes, I'm quite sure it's due to the latest patch. I have missed them due to the other VCS warnings. I will take care of them.
(In reply to comment #80) > I don't know if these warnings come from the last patch or if I'm doing > something wrong, but after launching Anjuta I see a bunch of messages Yes, it was due to this patch. I have fixed it, now it should run without any warning.
Other trivial issues to port (list not complete) for volunteers: ./anjuta/plugins/class-inheritance/class-inherit.c: gnome_app_fill_menu (GTK_MENU_SHELL (plugin->menu), canvas_menu_uiinfo, ./anjuta/libanjuta/anjuta-preferences.c: gnome_font_picker_set_font_name (GNOME_FONT_PICKER (prop->object), ./anjuta/plugins/gdb/utilities.h: /* Equivalent to gnome_execute_shell(), but does not overwrite ./anjuta/plugins/tools/execute.c: gnome_execute_terminal_shell (dir, cmd); (hint: use xdg-terminal)
Yuriy: Any news on the gnome-config patch? I think it shouldn't be difficult to make this backward-compatible in some way.
gnome-config is removed in revision 4489. I think my patch is not needed now.
To update the status here: Grepping for "libgnome" and "<gnome" in anjuta master comes up with several hits in ./plugins/project-wizard/templates/*.wiz plus ./libanjuta/gbf-project.c: #include <libgnome/gnome-macros.h> ./libanjuta/libanjuta-1.0.pc.in: Requires: libgnomeui-2.0 ./plugins/gbf-am/gbf-am-project.c: #include <libgnome/gnome-macros.h> ./plugins/gbf-am/test.c: #include <libgnome/gnome-init.h> ./plugins/gbf-mkfile/gbf-mkfile-project.c: #include <libgnome/gnome-macros.h> ./plugins/gtodo/egg-datetime.c: /* from libgnomeui */ ./plugins/project-manager/gbf-project-view.c: #include <libgnome/gnome-macros.h> ./plugins/project-wizard/property.c:#include <gnome.h> ./plugins/tools/execute.c: #include <libgnome/libgnome.h> ./configure.in: PKG_CHECK_MODULES(GNOME, libgnome-2.0 >= $GNOME_REQUIRED libgnomeui-2.0 >= $GNOME_REQUIRED)
I will take care of removing GnomeIconEntry from the project wizard. I think the other occurrences are in unused code and should be easier to fix.
I think I have finally complete this bug. - I have replaced the remaining gnome macros in the old gnome build part. - I have replaced the GnomeIconEntry with a GtkButton displaying an image and calling a file browser dialog - I have written a function execute_terminal_shell in anjuta_utils I have update anjuta-extra to make it works with the new libanjuta which doesn't require libgnome anymore. But 2 plugins in anjuta-extra, profile and class inheritance need gnome canvas and I don't know any replacement for this. Perhaps we should open another bug for it. I think we should remove the gnome project template in anjuta project wizard and perhaps use ui file everywhere instead of glade file.