March 2004 - Inkscape Devel - Inkscape Mailman

is Python scripting supported yet?
by Kevin Altis 15 Mar '04

15 Mar '04

Is Python scripting of Inkscape supported yet? If not is it still planned? ka

2 1

RE: [Inkscape-devel] Bug 866103: non-well-formed preferences.xml causes data los
by bulia byak 15 Mar '04

15 Mar '04

>What's a good approach for dealing with this issue? Do we need to have >a validation step so that it's possible to descriminate good from bad >files? The most useful and probably easiest is to create a backup file each time preferences are written. This, combined with modal errorbox saying that the preferences failed to parse and will be replaced by defaults, will be enough to resolve this. Validation is not necessary. _________________________________________________________________ MSN Premium: Up to 11 personalized e-mail addresses and 2 months FREE* http://join.msn.com/?pgmarket=en-ca&page=byoa/prem&xAPID=1994&DI=1034&SU=ht…

1 0

Bug 866103: non-well-formed preferences.xml causes data loss
by Bryce Harrington 15 Mar '04

15 Mar '04

Summary: non-well-formed preferences.xml causes data loss Bug ID: 866103 ------------------------------------------------------------------------ 1. edit ~/.inkscape/preferences.xml so it's not well-formed (e.g. delete any > in a tag) 2. run inkscape 3. quit The file is overwritten with all the default preferences, all old data in it is lost. Expected behavior: use default values while the program is running but NOT overwrite data on quit if it was not read successfully, or at least leave a backup copy. --buliabyak Currently there's no way to discriminate between a failure to load preferences because the file did not exist yet and a failure because that file was not well-formed. -- mental ------------------------------------------------------------------------ What's a good approach for dealing with this issue? Do we need to have a validation step so that it's possible to descriminate good from bad files? This sounds like a fairly easy thing to fix once an approach is decided. Bryce

1 0

Advert: inkscape-devel@lists.sourceforge.net - Due Payment, account.
by Leased I. Clobber 14 Mar '04

14 Mar '04

1 0

re: line wrapping
by MenTaLguY 14 Mar '04

14 Mar '04

...and yes, I realize I've occasionally gone back and tweaked line wrapping myself. That doesn't mean it's a good idea (except maybe in really extreme cases, but those are better candidates for actually being rewritten). -mental

1 0

line wrapping
by MenTaLguY 14 Mar '04

14 Mar '04

While the coding standards make suggestions for how to wrap lines, I do not think line wrapping should be required. I've been getting complaints about authors who are coming back to find their code line-wrapped in ways they are not happy with. I don't mind people going back and fixing whitespace/indents/brace positioning, but I would prefer that wrapping be left alone unless you're at least also making functional changes to those lines. Otherwise it creates a lot of needless churn in the diffs, and it can't simply be ignored with -b. I also notice that our emacs modeline includes a fill-column:99 ... do we need to rethink this? I don't mind it so much if it's wrapping new code, but if it's forcibly reformatting existing code, that's not cool. -mental

1 0

Re: offset & simplify
by bulia byak 14 Mar '04

14 Mar '04

> > I suspect it's something with your code. > >indeed it is. the ConvertToShape() return 4, meaning >it fucked up the uncrossing. the path you give him is >a good example of polygons intersectors don't like: >"long" edges with lots of intersections in a small >area, and worst of all, small angles between segments. >there is also another tiny star in the svg that isn't >rendered properly for the same reasons. >i'll take a look at the problem when i have some time, >but you must realize that converting strokes to >polygons just to raster them is not a clever method. >dumbly rasterizing them segment by segment with a >supersampled antialiasing is faster (well, ought to be >faster) and much more robust to numerical problems. if >you want to get rid of rasterization bugs, i'm afraid >you'll have to get there one day. Yes, I never could fully understand why an intersector is needed in a renderer at all. This is design we inherited from Sodipodi. Do you think it's realistic to change that? How hard could that be? > > you add a check that would prevent it from doing any > > simplification on a > > subpath that contains only one node? That would stop > > it from growing > > indefinitely. I would do this myself but I only know > >here is also a big algorithmical problem: the >simplification algorithm (if you can call that an >algorithm) is not even able to recompose a single >cubic bezier curve from a polyline coming from a >single bezier curve; it will produce several cubic >curves, unless you give a very big treshhold, or >refine the fitting a lot (and it takes time). the >growth/rotate/oddities would probably not happen if >the algorithm worked properly on a single cubic curve. >before adding shortcuts to prevent degenerate cases, >it's better (IMHO) to make the simplification behave >correctly on such simple paths. I don't know, but if part of what Simplify does is adding more nodes sometimes, I'd say it's OK. The design appeal of Simplify is in its "randomness" and "generalization". The number of nodes is of secondary importance. The goal is not a mathematically perfect simplification, but one that "feels right". The current algorithm is very good in this respect, imho. _________________________________________________________________ Add photos to your messages with MSN Premium. Get 2 months FREE* http://join.msn.com/?pgmarket=en-ca&page=byoa/prem&xAPID=1994&DI=1034&SU=ht…

2 1

Re: FFDingbests
by bulia byak 14 Mar '04

14 Mar '04

>Hmm, this is getting stranger and stranger. Despite the warning >from pfaedit, the old one did work ok for me on both Linux and >Windows; On linux, that may be caused by different verions of fontconfig, freetype, or something else. But I didn't test that font on windows. > > Can you provide another example of a valid font which Inkscape > > sees on one platform but does not see on another? > >I'll try to find one when I get back to Windows next week. OK >I never thought about it, because I always edit text using Text >and Fonts. However, if Text and Fonts is going away, this >definitely would be useful (or perhaps "must be done" if we >don't want to lose existing functionality). OK, I will look into incorporating that patch. Another related question: Can you please test the status this bug: https://sourceforge.net/tracker/index.php?func=detail&aid=850966&group_id=9… Is this still a problem with the latest CVS? If yes, can you provide any additional pointers or information? What to look for to have that fixed? _________________________________________________________________ Tired of spam? Get advanced junk mail protection with MSN Premium http://join.msn.com/?pgmarket=en-ca&page=byoa/prem&xAPID=1994&DI=1034&SU=ht…

2 1

RE: FFDingbests
by bulia byak 13 Mar '04

13 Mar '04

>I don't know why they changed all the names. I downloaded the >font fresh, and found that all the names, including names in >the Postscript fonts (both the PC and Mac versions), have changed. > >Perhaps they are doing this to discourage people from using this >font and encourage them buy the paid version instead... :-/ There may be plenty of other reasons, too. ><< dingbest.ttf >> This font has fewer characters than Dingbests-Regular, and it seems broken. pfaedit says "TTF has bad glyph count field", and if i install it on linux, Inkscape sees no characters in it (all are empty rectangles). The TTF file you sent is 36697 bytes (in case it was truncated or corrupted). So maybe they redid and renamed the font exactly because the old one was broken (the new one works OK for me on both platforms). Can you provide another example of a valid font which Inkscape sees on one platform but does not see on another? On another topic, as a person using Chinese, can you comment on this post: http://sourceforge.net/mailarchive/message.php?msg_id=7449173 Do you think adding this capability to Inkscape would be useful? _________________________________________________________________ STOP MORE SPAM with the MSN Premium and get 2 months FREE* http://join.msn.com/?pgmarket=en-ca&page=byoa/prem&xAPID=1994&DI=1034&SU=ht…

2 1

Adobe Illustrator CS screenshots and comments.
by Alan Horkan 13 Mar '04

13 Mar '04

So much for an unbiased write up, the Grid Mesh Tool rocks! (Well it produces great looking results, I haven't quite gottne the hang of using it yet). I've taken screenshots and grabbed a few samples http://www.maths.tcd.ie/~horkana/inkscape/illustrator/ Screenshot of Adobe Illustrator CS showing the About Inkscape SVG (139 kb) http://www.maths.tcd.ie/~horkana/inkscape/illustrator/Adobe-Illustrator-CS/… I was interested to notice that Adobe Illustrator CS doesn't render the About Inkscape SVG correctly (it also seems the version of librsvg used by the EOG viewer doesn't render it correctly either). The white part was not shown, I didn't make any changes, although I have selected it to show it is still there. It seems Adobe Illustrator isn't correctly interpreting the compound paths. Screenshot of Adobe Illustrator CS showing the Graph Tool(s) and tooltips for various graphs (81 kb). http://www.maths.tcd.ie/~horkana/inkscape/illustrator/Adobe-Illustrator-CS/… The screenshot shows many of the hidden tool options that you access by clicking and holding on tool button for a few seconds (and I should point out that usability people think click and hold is generally a bad idea). The screenshot was altered to include all the tooltips at once and usually only the active tool is coloured. Screenshot of Adobe Illustrator CS in fullscreen mode with menubar showing a simple Bar Chart created using the Graph Tool. (53 kb) http://www.maths.tcd.ie/~horkana/inkscape/illustrator/Adobe-Illustrator-CS/… As soon as I tried the graph tool I immediately saw the potential for Gnome Office integration and the reuse of existing components if developers are interested. Embeddding Gnumeric graphs into Inkscape! Gnumeric can already be embedded using bonobo which unfortunately is not a cross platform technology but Gnumeric is evolving towards being crossplatform so Jody Goldberg (Gnumeric maintainer,) may be able to suggest a better way to do easily produce a portable graph tool dialog based on Gnumeric. The Gnumeric graphing engine is in a seperate library adn the graphs have an SVG backend, hopefully this would allow the graphs to be embedded and the resulting graphs could be embedded as SVG so that Gnumeric would only be needed for the creation not the viewing of graphs. >From my brief playing around with the Adobe Illustrator Graph tool I didn't think it was particularly well done. It wasn't very obvious how to adjust the data of the graph (figured it out eventually though) and it was even less obvious how I might convert from one graph type to another without needing to create a whole new graph. My point is that if Inkscape were to have this feature it could easily do a better job of it. Screenshot of Adobe Illustrator in fullscreen mode with menubar show the sample file Flower.ai that is included with Adobe Illustrator (195 kb) http://www.maths.tcd.ie/~horkana/inkscape/illustrator/Adobe-Illustrator-CS/… I turned on a whole load of differnt pallettes and tried to show as many of the tools as I possibly could, and I made some selections so you could see the Grid Mesh. I was told that it would be important to get some postscript out so I converted the sample Flower.ai file into EPS and SVG formats (with some slightly differnt options in each case) and the files are here: http://www.maths.tcd.ie/~horkana/inkscape/illustrator/Adobe-Illustrator-CS/ I only have a 30 day Trial of Adobe Illustrator and although I'll try my best to produce lots of screenshots and write ups I urge you all (even the busy developers) to help out and try Adobe Illustrator CS as well or perhaps try Corel Draw (Corel Graphics Suite 12) or Macromedia Freehand (or Xara X from Bryce, or Deneba Canvas* from ACD Systems, or Jasc Web Draw) or any other vector graphics program that inspires you. Perhaps even some CAD or 3D graphics programs would inspire you, from what I've read Corel Graphics Suite 12 includes many advanced alignment guides and tools that make it quite useful for simpler technical drawing. Enthusiastically Yours Alan Horkan http://advogato.org/person/AlanHorkan/ PS Thanks for the software. PPS I bet I've forgotten something, I probably should have broken this into several smaller emails. * I got a 'free' full copy of Canvas 7 (current version is 9) inlcuded with Computer Arts magazine (annoying online registration required so I haven't tried it yet), published in the UK by Future publishing. Anyone else in the UK might be interested in trying it out.

4 4