- Inkscape Devel - Inkscape Mailman

Cafepress
by Ted Gould 11 Nov '03

11 Nov '03

The other night we were talking about setting up a cafepress store for the really cool inkscape logo - what do people think? Mental mentioned that he was concerned about where the money would go. I was thinking two things: 1) We just sell things at cost. I don't know if you can do this with Cafepress, but I would assume you could. 2) We donate all the profits (probably a couple bucks) back to Sourceforge. What do you guys think? I'm willing to set it up and get/write checks to make the second one happen. --Ted

3 2

icon.c again + patch
by Robert Crosbie 11 Nov '03

11 Nov '03

The benefit of the icon.c routines is that you can load an xpm, png or svg with the same call. You can replace all your .xpm images with svg and you don't need to change any code. It would certainly be cool if inkscape had all svg icons since it is an svg editor. The problem with icon.c is in sp_icon_new_full where we have: static iconlib = g_hash_table_new (g_str_hash, g_str_equal); but iconlib is never freed. I assume it is used to speed up the retrieval/loading of images/icons. Also every time an object is added to the hash table, the key is strduped: g_hash_table_insert (iconlib, g_strdup (c), icon->px); That key is never freed either. You create and destroy a few icons and inkscape/spdipodi goes kablooy! These things would be easier to manage if inkscape had true objects and classes. I almost have a fix for icon.c, but in the mean time, if you just want to have a stable 0.35, this patch should stop the transformation dialog from crashing: Index: transformation.c =================================================================== RCS file: /cvsroot/inkscape/inkscape/src/dialogs/transformation.c,v retrieving revision 1.3 diff -u -p -r1.3 transformation.c --- transformation.c 31 Oct 2003 06:28:39 -0000 1.3 +++ transformation.c 10 Nov 2003 04:02:58 -0000 @@ -226,7 +226,7 @@ sp_transformation_dialog_new (void) g_signal_connect (G_OBJECT (b), "clicked", G_CALLBACK (sp_transformation_dialog_close), dlg); /* Move page */ - img = sp_icon_new (SP_ICON_SIZE_NOTEBOOK, "move"); + img = gtk_image_new_from_file (INKSCAPE_GLADEDIR "/move.xpm"); gtk_widget_show (img); page = sp_transformation_page_move_new (G_OBJECT (dlg)); gtk_widget_show (page); @@ -234,7 +234,7 @@ sp_transformation_dialog_new (void) g_object_set_data (G_OBJECT (dlg), "move", page); /* Scale page */ - img = sp_icon_new (SP_ICON_SIZE_NOTEBOOK, "scale"); + img = gtk_image_new_from_file (INKSCAPE_GLADEDIR "/scale.xpm"); gtk_widget_show (img); page = sp_transformation_page_scale_new (G_OBJECT (dlg)); gtk_widget_show (page); @@ -242,7 +242,7 @@ sp_transformation_dialog_new (void) g_object_set_data (G_OBJECT (dlg), "scale", page); /* Rotate page */ - img = sp_icon_new (SP_ICON_SIZE_NOTEBOOK, "rotate"); + img = gtk_image_new_from_file (INKSCAPE_GLADEDIR "/rotate.xpm"); gtk_widget_show (img); page = sp_transformation_page_rotate_new (G_OBJECT (dlg)); gtk_widget_show (page); @@ -380,7 +380,7 @@ sp_transformation_page_move_new (GObject sp_unit_selector_set_unit (SP_UNIT_SELECTOR (us), sp_desktop_get_default_unit (SP_ACTIVE_DESKTOP)); } /* Horizontal */ - img = sp_icon_new (SP_ICON_SIZE_BUTTON, "arrows_hor"); + img = gtk_image_new_from_file (INKSCAPE_GLADEDIR "/arrows_hor.xpm"); gtk_table_attach (GTK_TABLE (tbl), img, 0, 1, 0, 1, (GtkAttachOptions)0, (GtkAttachOptions)0, 0, 0); adj = (GtkAdjustment *) gtk_adjustment_new (0.0, -1e6, 1e6, 0.01, 0.1, 0.1); g_object_set_data (obj, "move_position_x", adj); @@ -389,7 +389,7 @@ sp_transformation_page_move_new (GObject sb = gtk_spin_button_new (adj, 0.1, 2); gtk_table_attach (GTK_TABLE (tbl), sb, 1, 2, 0, 1, (GtkAttachOptions)( (GtkAttachOptions)( GTK_EXPAND | GTK_FILL ) ), (GtkAttachOptions)( (GtkAttachOptions)( GTK_EXPAND | GTK_FILL ) ), 0, 0); /* Vertical */ - img = sp_icon_new (SP_ICON_SIZE_BUTTON, "arrows_ver"); + img = gtk_image_new_from_file (INKSCAPE_GLADEDIR "/arrows_ver.xpm"); gtk_table_attach (GTK_TABLE (tbl), img, 0, 1, 1, 2, (GtkAttachOptions)0, (GtkAttachOptions)0, 0, 0); adj = (GtkAdjustment *) gtk_adjustment_new (0.0, -1e6, 1e6, 0.01, 0.1, 0.1); g_object_set_data (obj, "move_position_y", adj); @@ -555,7 +555,7 @@ sp_transformation_page_scale_new (GObjec sp_unit_selector_set_unit (SP_UNIT_SELECTOR (us), sp_unit_get_by_abbreviation ("%")); g_signal_connect (G_OBJECT (us), "set_unit", G_CALLBACK (sp_transformation_scale_set_unit), obj); /* Horizontal */ - img = sp_icon_new (SP_ICON_SIZE_BUTTON, "scale_hor"); + img = gtk_image_new_from_file (INKSCAPE_GLADEDIR "/scale_hor.xpm"); gtk_table_attach (GTK_TABLE (tbl), img, 0, 1, 0, 1, (GtkAttachOptions)0, (GtkAttachOptions)0, 0, 0); adj = (GtkAdjustment *) gtk_adjustment_new (0.0, -1e6, 1e6, 0.01, 0.1, 0.1); g_object_set_data (obj, "scale_dimension_x", adj); @@ -564,7 +564,7 @@ sp_transformation_page_scale_new (GObjec sb = gtk_spin_button_new (adj, 0.1, 2); gtk_table_attach (GTK_TABLE (tbl), sb, 1, 2, 0, 1, (GtkAttachOptions)( GTK_EXPAND | GTK_FILL ), (GtkAttachOptions)( GTK_EXPAND | GTK_FILL ), 0, 0); /* Vertical */ - img = sp_icon_new (SP_ICON_SIZE_BUTTON, "scale_ver"); + img = gtk_image_new_from_file (INKSCAPE_GLADEDIR "/scale_ver.xpm"); gtk_table_attach (GTK_TABLE (tbl), img, 0, 1, 1, 2, (GtkAttachOptions)0, (GtkAttachOptions)0, 0, 0); adj = (GtkAdjustment *) gtk_adjustment_new (0.0, -1e6, 1e6, 0.01, 0.1, 0.1); g_object_set_data (obj, "scale_dimension_y", adj); @@ -674,7 +674,7 @@ sp_transformation_page_rotate_new (GObje gtk_table_set_col_spacings (GTK_TABLE (tbl), 4); gtk_box_pack_start (GTK_BOX (vb), tbl, FALSE, FALSE, 0); - img = sp_icon_new (SP_ICON_SIZE_BUTTON, "rotate_left"); + img = gtk_image_new_from_file (INKSCAPE_GLADEDIR "/rotate_left.xpm"); gtk_table_attach (GTK_TABLE (tbl), img, 0, 1, 0, 1, (GtkAttachOptions)0, (GtkAttachOptions)0, 0, 0); adj = (GtkAdjustment *) gtk_adjustment_new (0.0, -1e6, 1e6, 0.01, 0.1, 0.1); g_object_set_data (obj, "rotate_angle", adj); Cheers, Rob. http://members.rogers.com/rcrosbie _________________________________________________________________ Is your computer infected with a virus? Find out with a FREE computer virus scan from McAfee. Take the FreeScan now! http://clinic.mcafee.com/clinic/ibuy/campaign.asp?cid=3963

2 1

Re: [Inkscape-devel] more keys
by bulia byak 11 Nov '03

11 Nov '03

> Tool switching > ============== > F1 and Ctrl F6 don't seem to do what they're supposed to; all of the > rest of the shortcuts worked as documented (both the letter and function > key variants). They work for me, and I have absolutely no idea why these might not work. They're just two shortcuts among others. The only reason may be that the underlying functions do not work, which I did not touch. Is there any diagnostics? Maybe a "make clean" or something would help? > It would be nice to have a shortcut to the Editing Window added (or even > better would be keyboard shortcuts to turn grid and snap on/off > directly). Now # toggles grid and | toggles guides, except that toggling guides does not work for some reason (it's disabled in the dialog too). > I noticed that ESC works for many of the dialogs but not all. The > export window, for example, didn't respond to Esc. I will try to see if I can do something about that. It's high priority because the simple keys like arrows and esc must work reliably everywhere. I've added a TODO section at the bottom of http://www.inkscape.org/cgi-bin/wiki.pl?KeyboardShortcuts This and many other ideas are there. Please discuss or add your own! Yemu: your suggestion is there too, edit as appropriate. > Zoom and scroll > =============== > + on the keypad works for zooming in, but the regular + key doesn't. I think I fixed this. Please check. > We probably need a keyboard mapping for "Fit the whole selection into > the window". Perhaps one of the number keys to keep with the scheme? > Also note that keypad-5 is mapped to "Fit the whole drawing into > window", so that could be added to the docs. However, the regular 5 key > isn't mapped. Since both regular and keyboard 0 and 1 are mapped, it > seems sensible for both 5's to be likewise. I think the old scheme was not very logical or convenient. Now I did this: 1 is 1:1, 2 is 1:2, 3 is selection, 4 is drawing, 5 is page. This makes it a more logical sequence (assuming drawing is smaller than page). > Mapping Ctrl-z and Ctrl-shift-z to the opposite actions is kind of > slick; makes it easy to quickly scroll up and down through the > transaction log. And having Ctrl-y mapped to the redo makes sense since > that's a commonly expected mapping. > > I wonder if for perfect balance, Ctrl-shft-y should be mapped to 'undo'? Why not. > Ctrl-w 'close window' only seems to work when the mouse is hovering over > the window when the window system is set to click-to-follow mode. Yes, it's a local binding for a canvas window and perhaps it does not work when mouse is away. It cannot be made global. > I'm not sure what Ctrl-shft-n does, but it didn't seem to work for me. No traces in the code. Removed. > Ctrl-shft-p did not pop up the print preview This one is enabled, but the function does not work (from menu, too). Someone needs to find out. > Ctrl-shft-s did not pop up the save file as box Should work now. > Ctrl-q only exits if the cursor is in the editing window; it doesn't > work when used over the toolbox nor if no document is loaded. Fixed by making it global. > Selector tool shortcuts > ======================= > The arrow keys all worked as advertised - that's going to be a really > nice enhancement. I noticed though that the keyboard arrows did not > always work - the kbd up and down are mapped to zoom up/down, and the rt > and left don't have actions. The kbd alt-arrows don't have action. > All kbd-arrows with the shift or shift-alt modifiers work as documented. By "keyboard", do you mean "keypad"? If so, that's strange. But after some thinking I remembered that some time ago I patched my X keyboard configuration so that keypad arrows _always_ produce arrows, with or without shift. (I can give details of the patch if someone is interested.) Without that patch, the KP arrows were arrows without shift and digits with shift, and vice versa if numlock is on. I never use KP digits but I like to use KP arrows with and without shift (for selecting) in Emacs, hence the patch. Now, if the original configuration is standard (it was by default on Mandrake), this means most people won't be able to use KP arrows with shift at all, unless they reconfigure X like I did. (Or they only can use KP arrows with shift but cannot without, if numlock is on.) I personally find this situation a bit weird, and we perhaps need to talk to an expert in X keyboard to see if there's another workaround. The idea to reconfigure XKB was the first to come to my mind and it worked for me, so I never researched the matter any deeper since that. We'll also need to find out how this is done outside of linux. Maybe we need to check for, e.g., GDK_KP_4 instead of (or in addition to) Shift with KP_Left? I will try this but I'll need someone else to test it on a standard linux and on other platforms. As for "kbd up and down are mapped to zoom up/down", no - it's just that the zoom value field in the bottom right corner grabs focus (very annoying!). I'll see if I can tame it somehow. > Ctrl-a 'select-all' also works but needs adding to the documentation. Done. I also made this local to the select tool, as it not always makes sense (and may be confusing) in other tools. > I'm wondering about ctrl-d though - in emacs and probably other apps > it's mapped to the same function as the DEL key, whereas here it > duplicates the selected item. In Xara, it's duplication as well, and I believe in Corel Draw too. Those are better matches for us than Emacs :) > Home/end/pgup/pgdn all work. The keyboard versions of those aren't > mapped similarly. I'm undecided if they should though; could be too > much functionality for the keyboard. For consistency, I enabled the KP keys too. > I had a thought... What if we made the keyboard mappings not do any > document changes, so that it's read only? What I'm thinking is that > being able to rely on the keyboard as a read-only navigation interface > would be handy since they could navigate around without having to worry > about accidentally hitting the wrong key and fussing up the document. I think it's too much to change, there are already many shortcuts that modify the document. Instead, isn't there a slideshow mode that only displays SVG without editing? > Ctrl-g and ctrl-shft-g work for group/ungroup. Ctrl-u does ungroup; > perhaps Ctrl-shft-u should group? Why not. > The mouse modifiers for stretch, rotate, skew, etc. all checked out. > Are some of these new or had I just not known about them before? > I am going to find Ctrl-move and alt-transform quite useful. :-) No, I did not touch these yet (but I plan to). They are all from Sodipodi. > Node edit tool shortcuts > ======================== > In editing mode, the regular arrows all worked as documented. Nice. > The keyboard keys didn't do anything though. See above for keypad ideas. > I'm wondering if in node-edit mode it might be more useful for the > keyboard arrowkeys to be left to view commands like moving the view > window, zooming, etc. Hmm, maybe we need a stronger defined concept of > what the keyboard does - either always precisely map to what the regular > arrow, paging, etc. keys do, or map to some other useful mode like > view-window control? You can zoom and pan in node editor, no problem. No need to deassign arrow keys. My concept is simple: everything must be consistent. If arrows move selection in selector, they must do the same in all other tools where "selection" makes sense. Ctrl-arrows pan canvas, and they do so everywhere. > You've got shift-k and shift-l listed twice. Fixed. > One thing that may be a bit awkward is with the unmodified keys, if you > happen to on the canvas instead of precisely over a node, the key can > cause you to leave node-edit mode and activate some other function. The > s key, for instance, ends up selecting the object, and the b key > activates the create-curve function. I thought about that and wanted to remove these unmodified keys to alt-keys or digits. But when I actually tried this, I found that they affect active node, and active node only exists when you are dragging it with your mouse. So psychologically, these are not "unmodified keys"; rather, they are "mousedrag+letter" keys and are unlikely to be confused with simple "letter" keys. The rest will have to wait until tomorrow... -- __________________________________________________________ Sign-up for your own personalized E-mail at Mail.com http://www.mail.com/?sr=signup Search Smarter - get the new eXact Search Bar for free! http://www.exactsearchbar.com/

2 1

Milestone 2 development work
by Bryce Harrington 11 Nov '03

11 Nov '03

As milestone 1 nears completion, we're ready to start looking at laying into M2. Also, a bunch of you are probably new to the codebase and looking for someplace to get your feet wet. The most sensible effort for new developers to dive into is bug fixing. Here's why: * Many bugs have been identified and documented in the Sodipodi bug tracker. A goodly number are also not that hard to fix. * A bug gives you a fixed goal to shoot for with a definite end. It is not a huge commitment of your time and doesn't carry any sort of ongoing responsibility - once you've got the fix in, you're done. * It will expose you to building the codebase, finding your way around in it, creating patches, etc. * Users *love* applications with few bugs. :-) There are two aspects to the bug work: 1. Bug Testing - Go through Sodipodi's bug list and test each report to see if it can be recreated in Inkscape and isn't already registered. If so, cut and paste it into the Inkscape bug tracker and mark that you've verified it still exists (and elaborate on how you found it). Include the Sodipodi bug ID number for traceability. 2. Bug Fixing - Grab a bug from the SF tracker and chase it down; try to find a fix for it, and post a patch to the patch tracker. Try to also improve the error checking behavior of the code, where possible. If those of us who are able to code C/C++ can close a couple bugs each, then we'll establish a sturdy base for building our next set of changes on. If this endeavor sounds interesting to you, feel free to drop me a line with any questions you have. Some other efforts that are going on, that you could participate in if you wish, include: * Research & plan GtkMM-ification (see mental or njh) * Design & develop new layer system (see mental) * Implement arrowhead support (see bryce) * Create unit tests for each source code file (see mental) * Implement new grid systems (see njh) Bryce

4 4

Announcing inkscape-announce list
by Bryce Harrington 10 Nov '03

10 Nov '03

A mailing list for Inkscape-related announcements is now available. https://lists.sourceforge.net/lists/listinfo/inkscape-announce This is intended to be a low-traffic list for people who don't wish to follow the regular development discussions but do want to know about new releases of Inkscape itself, plugins, events, and so forth. It's recommended that everyone on inkscape-devel@ subscribe to this list as well. Bryce

1 0

Fwd: [Inkscape-devel] OSX?: Patch
by T Ingham 10 Nov '03

10 Nov '03

I figured out how to get this done, here's the patch. Probably contains more data than necessary, but I'm learning. On Nov 9, 2003, at 11:54 PM, T Ingham wrote: > // Wl, removed > ./Makefile.am > > //malloc.h replaced with unistd.h > > ./src/helper/action.c > ./src/xml/repr-io.c > ./libarikkei/arikkei-token.c > > > <Makefile.am><action.c><arikkei-token.c><repr-io.c>

1 0

Bug on debian package
by Carlos Ingelmo 10 Nov '03

10 Nov '03

Since I don't know if this would go to the bug tracker or not, I write it here. It seems that 'debian/changelog' must be change in order to build debian package, by replacing sodipodi by inkscape or by creating a new one, I don't know anything about debian packaging. What I've done to build it is replace sodipodi by inkscape, although like it is a new project you might one to create a new changelog (by the way, it doesn't seem very updated). Thanks for reading... Carlos Ingelmo

2 1

isometric grid
by Robert Crosbie 10 Nov '03

10 Nov '03

Just wanted to say that the new grid system pictured on the website is an awsome awsome awsome idea. I can't wait to try it out. It would have been particularly useful when I was drawing triangular celtic knots. Cheers, Rob. http://members.rogers.com/rcrosbie _________________________________________________________________ STOP MORE SPAM with the new MSN 8 and get 2 months FREE* http://join.msn.com/?page=dept/bcomm&pgmarket=en-ca&RU=http%3a%2f%2fjoin.ms…

1 0

Re: [Inkscape-devel] OSX?
by T Ingham 10 Nov '03

10 Nov '03

Roger this, for the record. -t On Nov 9, 2003, at 11:26 PM, Daniel Díaz wrote: > Hello, Thomas. > > > On Sun, 2003-11-09 at 22:21, T Ingham wrote: >> The previous issue was related to Apple's version of glibtoolize, >> replacing with fink's copy fixes the ranlib issue. >> >> Now however, I have a more interesting note. I got the damn thing >> compiled! >> >> Here are some images showing a very unhappy Inkscape: >> >> http://www.coalmarch.com/images/inkscape/inkscape_osx_01.png > > I had that in my Yellowdog. I had to `make install' for the icons to be > loaded. Otherwise, the X server will display garbage instead of the > graphics. > > Hope that's it. > > Greetings! > > Daniel Díaz > yosoy@...31... > > > >> http://www.coalmarch.com/images/inkscape/inkscape_osx_02.png >> http://www.coalmarch.com/images/inkscape/inkscape_osx_03.png >> >> You can see that some of the icons appear to be loading in a corrupted >> fashion. This could be related to my cvs checkout ( done anon ) or >> perhaps the files needed for these icons simply aren't present. I >> can't >> say for sure. >> >> The other problem is that it appears as though our code base suffers >> from the same font problem that sodipodi did. Pogma at #fink on >> freenode.net says there is a patch available, but as I'm not really >> qualified for such business, I figured we could get someone to apply >> the necessary change? If I'm wrong, may I be flogged for thinking so. >> >> Laters, >> >> -thomas >> >> ps. I'm not sure if these qualify as bugs, since they appear to be >> platform and session specific. > > >

1 1

shortcuts
by Yemu 09 Nov '03

09 Nov '03

Witaj Bryce, W Twoim liście datowanym 9 listopada 2003 (20:38:47) można przeczytać: BH> On Sat, 8 Nov 2003, bulia byak wrote: >> Still more keys today. See the wiki page. Notably, esc now deselects and cancels rubberband selection. Also, lots more ways to switch tools and call >> dialogs. The program is becoming actually >> usable :) >> >> Actually, the bulk of the Sodipodi shortcuts programming task is now ready. However, quite a few bits remain, and still more are in my plans. It's >> addictive :) >> >> Sorry I've been committing all this right before the release, but I think it's unlikely to break anything. However, anyone please test the new keys >> before making a release. >> thanks bulia! shortcuts are great! they really speed up things! i have one suggestion concerning shortuts: it would be great if there were shourtcuts for aligning, like for example: c - center horizontally e - center vertically l - left r - right t - top b - bottom (these are key bindings from Corel which i like the most, and you must admit they are quite intuitive) best regards yemu

1 0