- Inkscape Devel - Inkscape Mailman

better patch for mask loose
by Popolon 30 Sep '06

30 Sep '06

The patch attached I just attached in bug #1562520 corrects it and corrects bug #1561034 too in a better way. it also free a string that was not freed in the bug #1561034 patched function that lead me to wrong patch. <https://sourceforge.net/tracker/?func=detail&atid=604306&aid=1562520&group_…> There are nothing in guide line about comment and cast. I do a cast (gchar *) for a temporary 'gchar *' from a 'const gchar *' given by attribute() function. And about comment : in src/path-chemistery.cpp there are C style single line comments: /* foo bar */ but in src/splivarot.cpp there are c++ comments: // foo bar and some time not spaced text and c++ style comments //foo bar Is there no project to add something about that in coding style guide?

1 0

a little more on the blur crash
by Jon A. Cruz 30 Sep '06

30 Sep '06

I managed to track down more on that crash problem. down in desktop-style.cpp, things seem to have some issues right after this: SPGaussianBlur * b = new SPGaussianBlur(); I added this, and see a warning. if ( !SP_IS_GAUSSIANBLUR(b) ) { g_warning("problem with gaussian blur object"); } I'd sprinkled some of those debug messages starting at the point where I'd seen the lockups on OS X, and then moved things back until I'd gotten there. So I think the new call there is not getting all needed items setup. Given the "TODO" comment right above that, I think this is the right point to change things. I'm not familiar with the filter code enough, and the other object creation that doesn't use new has a bit more context for creating things. So I think I'm at the point where I've done what I can tonight.

1 0

[PATCH] bug in FileSaveDialog
by Kees Cook 30 Sep '06

30 Sep '06

I know inkscape is normally a patch-first-talk-later place, but I haven't done a "serious" commit to SVN in a while, so I thought I'd see what people thought of this patch... Here is the problem: in Ubuntu Edgy (glib 2.12.3) the filedialog throws an error if you give it a _file_ rather than a _path_ in set_current_folder(). src/file.cpp expected (at one time) to give the dialog a full path, not a directory, but at some point the logic slowly changed. This patch restores the logic of letting the dialog specify a full path name. (And as such, the "drawing.svg", "drawing-1.svg" default names re-appear too.) Example scenario: - start without ~/.inkscape dir and without any ~/drawing*.svg files - open inkscape - draw anything - click "save" expected results: put in home directory, with "drawing.svg" already in the filename entry box. - click "OK" - open new inkscape window - draw anything - click "save" expected results: put in home directory, with "drawing-1.svg" - click on "Filesystem" and scroll to "tmp". - click "OK" - switch back to first window - Select "File/Save As..." expected results: put in home directory, with "drawing.svg" already in the filename entry box. Thoughts? Index: src/ui/dialog/filedialog.h =================================================================== --- src/ui/dialog/filedialog.h (revision 13104) +++ src/ui/dialog/filedialog.h (working copy) @@ -172,6 +172,10 @@ */ virtual void change_title(const Glib::ustring& title) =0; + /** + * Change the default save path location. + */ + virtual void change_path(const Glib::ustring& path) =0; }; //FileSaveDialog Index: src/ui/dialog/filedialog.cpp =================================================================== --- src/ui/dialog/filedialog.cpp (revision 13104) +++ src/ui/dialog/filedialog.cpp (working copy) @@ -1138,6 +1138,7 @@ Glib::ustring getFilename(); void change_title(const Glib::ustring& title); + void change_path(const Glib::ustring& dir); private: @@ -1355,7 +1356,7 @@ // leaving a trailing backslash on the directory name leads to the infamous // double-directory bug on win32 if (len != 0 && udir[len - 1] == '\\') udir.erase(len - 1); - set_current_folder(udir.c_str()); + myFilename = udir; } //###### Add the file types menu @@ -1455,10 +1456,7 @@ bool FileSaveDialogImpl::show() { - Glib::ustring s = Glib::filename_to_utf8 (get_current_folder()); - if (s.length() == 0) - s = getcwd (NULL, 0); - set_current_folder(Glib::filename_from_utf8(s)); //hack to force initial dir listing + change_path(myFilename); set_modal (TRUE); //Window sp_transientize((GtkWidget *)gobj()); //Make transient gint b = run(); //Dialog @@ -1519,6 +1517,24 @@ this->set_title(title); } +/** + * Change the default save path location. + */ +void +FileSaveDialogImpl::change_path(const Glib::ustring& path) +{ + myFilename = path; + if (Glib::file_test(myFilename, Glib::FILE_TEST_IS_DIR)) { + //fprintf(stderr,"set_current_folder(%s)\n",myFilename.c_str()); + set_current_folder(myFilename); + } else { + //fprintf(stderr,"set_filename(%s)\n",myFilename.c_str()); + set_filename(myFilename); + Glib::ustring basename = Glib::path_get_basename(myFilename); + //fprintf(stderr,"set_current_name(%s)\n",basename.c_str()); + set_current_name(basename); + } +} Index: src/file.cpp =================================================================== --- src/file.cpp (revision 13104) +++ src/file.cpp (working copy) @@ -589,7 +589,8 @@ save_loc.append(formatBuf); } } else { - save_loc = Glib::path_get_dirname(doc->uri); + save_loc = Glib::build_filename(Glib::path_get_dirname(doc->uri), + Glib::path_get_basename(doc->uri)); } // convert save_loc from utf-8 to locale @@ -615,6 +616,8 @@ (char const *) _("Select file to save to"), default_extension ); + else + saveDialogInstance->change_path(save_loc); saveDialogInstance->change_title(dialog_title); bool success = saveDialogInstance->show(); @@ -640,7 +643,7 @@ if (success) prefs_set_recent_file(SP_DOCUMENT_URI(doc), SP_DOCUMENT_NAME(doc)); - save_path = fileName; + save_path = Glib::path_get_dirname(fileName); prefs_set_string_attribute("dialogs.save_as", "path", save_path.c_str()); return success; -- Kees Cook @outflux.net

1 0

Simplify (ctrl+L) remove mask & clip_path fix
by Popolon 29 Sep '06

29 Sep '06

I opened few weeks ago this bug report and just fixed it, the patch is in the bug report : https://sourceforge.net/tracker/index.php?func=detail&aid=1561034&group_id=… bug #1561034 Regards

1 1

pr launched for 0.44.1
by Jon Phillips 29 Sep '06

29 Sep '06

I finally launched out the press barrage for 0.44.1 >From the looks of the tracker, there aren't many packages up there. Also, could someone please post up a snippet or notice of the release up on the main Inkscape blog. I don't know the process for doing it after the site changeover. Thanks! Jon -- Jon Phillips San Francisco, CA USA PH 510.499.0894 jon@...235... http://www.rejon.org MSN, AIM, Yahoo Chat: kidproto Jabber Chat: rejon@...896... IRC: rejon@...897... Inkscape (http://inkscape.org) Open Clip Art Library (www.openclipart.org) Creative Commons (www.creativecommons.org) San Francisco Art Institute (www.sfai.edu)

2 2

metadata: general vs. specific in SVG
by Jon Phillips 29 Sep '06

29 Sep '06

So, one of the things I've wanted to work on is more specific metadata usage per-object which we could hav UI for in Inkscape and for use in Open Clip Art Library. I did some spelunking while doing research on pippin's gegl (next generation gimp library, http://pippin.gimp.org/gegl) and I found how to do this in SVG. The SVG spec defines global and specific per object metadata: SVG 1.1 standard http://www.w3.org/TR/SVG11/metadata.html SVG 1.2 working draft with more specifics http://www.w3.org/TR/2004/WD-SVG12-20041027/nonvisual.html#external-metadata Its really simple and is spelled out in the SVG 1.1 spec like so: <svg> <metadata>  </metadata> <rect> <metadata>  </metadata> </rect> </svg> Here is what the spec states about the specific metadata: "It is strongly recommended that at most one 'metadata' element appear as a child of any particular element, and that this element appear before any other child elements (except possibly 'desc' or 'title' elements) or character data content. If metadata-processing user agents need to choose among multiple 'metadata' elements for processing it should choose the first one." Thus, what I think would be a good approach is to generalize Kees' document metadata dialog into a panel which can be re-used and then either we add an object metadata dialog (using Jon's panel work), or just attach this functionality to the object properties dialog. Thus, one could select any object and attach the same good ole quality metadata we currently can set to the entire document. Kees what do you think? Others, what do you think? This would be awesome for asset management, right Andy, and would also allow some cool features in compositing multiple elements from Open Clip Art Library... Jon -- Jon Phillips San Francisco, CA USA PH 510.499.0894 jon@...235... http://www.rejon.org MSN, AIM, Yahoo Chat: kidproto Jabber Chat: rejon@...896... IRC: rejon@...897... Inkscape (http://inkscape.org) Open Clip Art Library (www.openclipart.org) Creative Commons (www.creativecommons.org) San Francisco Art Institute (www.sfai.edu)

3 5

Re: [Inkscape-devel] Please clean up console output
by Johan Engelen 29 Sep '06

29 Sep '06

Bulia Byak wrote: [mailto:buliabyak@...400...] > Here are some specific problems I found: > > 1. Several "** Message: arg" lines every time GUI is started. > I think this is from the axonometric grid code by Johan > Engelen (though I may be wrong). > > 2. "(inkscape:12768): Gtk-CRITICAL **: gtk_style_detach: > assertion `style->attach_count > 0' failed" every time I quit > Inkscape. Has been there for a long time already, though I > don't know who's the culprit. > > 3. Three "(inkscape:12772): GLib-GObject-CRITICAL **: > g_object_set_qdata_full: assertion `quark > 0' failed" > warnings when I run Inkscape with a query, for example > "inkscape -W file.svg". This warning originates in > paramnotebook.cpp code, also by Johan - most likely the new > code is unaware that Inkscape can be run without GUI, tries > to do something GUIish and fails. > > 4. A stream of "(inkscape:12776): GLib-GObject-WARNING **: > invalid cast from `SPCAxonomGrid' to `SPCGrid'" when I'm > trying to use the new axonometric grid. > > 5. A "(inkscape:12778): Gtk-CRITICAL **: gtk_misc_set_alignment: > assertion `GTK_IS_MISC (misc)' failed" when I open the > Document Properties. Ad 1. Correct. Removed the stupid debug message. I forgot about it. It is actually not that handy to have a window in Inkscape which you have to enable to see these messages :( Ad 2. I cannot reproduce this one. Ad 3. Will look into this later. I already knew in the back of my head that Inkscape can be run without UI. On WinXP because the console output didn't work, I forgot about it; thanks for making me remember! Ad 5. Fixed. Had to do with the entry of the radiobutton-widget (which gridtype to use) in the widgetlist. Ad 4. Solution: do not use the axonometric grid with snapping! :P I haven't had time to code the snapping thing yet for the axonometric grid; and it is clear that some generalization of grid/snapping code and mechanism is needed for this to work nicely. At the moment, I haven't changed any of the snapping code, therefore invalid casts result. I am sorry that the axonometric grid commit gives so many warnings/errors. Especially the invalid cast to SPCGrid is nasty. I committed it so all can see what the grid looks like on which I am working, so I can get comments on how to improve things. Use on your own risk! :P

1 0

wouldn't penitentiary
by Flora Malone 29 Sep '06

29 Sep '06

warming,A said Mark Pagani, professor of geology and geophysics at Yale

1 0

Re: [Inkscape-devel] Please clean up console output
by Johan Engelen 28 Sep '06

28 Sep '06

> -----Original Message----- > From: bulia byak [mailto:buliabyak@...400...] > Sent: donderdag 28 september 2006 4:23 > > Some of the recent commits resulted in various debug stuff > and warnings output to console. Please don't leave it that > way, clean up after yourself promptly. Thanks for letting me know about these problems: I will try to fix them ASAP. I did not know about these warnings/errors because I don't see any: how can I view console output in windows XP? (running Inkscape from the commandline does not help) Regards, Johan

4 5

Re: [Inkscape-devel] Tango style icons
by Ronan Zeegers 28 Sep '06

28 Sep '06

The first idea with this icons theme was not to follow strictly the Tango guideline. I Just wanted a style that fit well with the default tango gtk/gnome icons (open, save, print, preferences...), very minimalist icons and that can be fast and easy to read. Some work and elements from the Tango icons have also been reused. So, Tangoished for this theme is maybe not the best term... I'm award of the new gimp icons that can be found here: http://tango.freedesktop.org/ArtLibreSet I think they don't fit exactly the minimalist idea behind my theme. But I can take a closer look to the Tango guidelines and see if I can change my icons and still stick to the original concept.... @graffoo: > Also can you post a download link? > Yep. Before that, I have to work on lots of stuffs (menu icons) and clean the icons.svg I also have to take a closer look the icons size. The toolbar seems to use a not very conventional 24x24 pix size... Ronan Zeegers

1 0