October 2005 - Inkscape Devel

win32 build working again
by Bob Jamison 08 Apr '24

08 Apr '24

The win32 build, which was broken for the weekend, is working again. I just made a "manual autobuild" again, so that we can see how it looks on Windows: http://troi.titan-aeu.com/~rjamison/inkscape/builds/ Bob

2 1

WEBFORM live bookmarks / rss
by pbhj 10 May '06

10 May '06

Message from Inkscape submission form Please note, that the submitter may not read this list. CC your answer to the submitter name: pbhj email: inkscape@...1029...%0Acc:pbhj@...19... submitted the following: RSS feed, great! No <link> in page header makes it v.hard for Firefox users to add a livemark though. <link rel=\"alternate\" type=\"application/rss+xml\" title=\"inkscape news\" href=\"/inkscape.rss\" /> Cheers pbhj

4 4

[ 1343164 ] Error compiling Inkscape from CVS
by Steven P. Ulrick 05 Nov '05

05 Nov '05

Hello Everyone I have filed a report on the Bug Tracker on my continuing issue compiling Inkscape from CVS on Fedora Core 3: https://sourceforge.net/tracker/index.php?func=detail&aid=1343164&group_id=… If you need any more info from me, please let me know. Steven P. Ulrick

2 9

Inkscape 0.43 mini-status
by Jon A. Cruz 04 Nov '05

04 Nov '05

I *believe* that we're in hard freeze and about set to do a Pre1. Translators should look into the file attached to RFE 1247694 "No ((parenthesis)) in menu labels.", and see if it should be applied or not. I think we've got a pending item still with Scislac updating the Calligraphy tutorial. There are a few translation patches that still need to be applied. I should be able to get to those tonight.

10 18

Useful info for potential Mac users
by William Swanson 31 Oct '05

31 Oct '05

Hi, I have a few friends who use Macs, and the X11 requirement seems to be a real tripping point for them when installing Inkscape. One friend figured it out and wrote up the following instructions: "If you are running later than OS X Tiger 10.4 you will need to download the X11 software from Apple at http://www.apple.com/support/downloads/x11formacosx.html. Or, if you already have OS X Tiger, you will need to insert your OS Disc #1 and run a Spotlight search in the CD for "X11User.pkg". Once you find and install that file, start another search for "X11SDK.pkg" and install that. Once you have both of those installed (with the OS Disc inside your drive), take out the disc, run the inkscape.dmg install file, and X11 will pop up. It might take a minute or two the first time you try, but it will load." I am not a Mac user, so I can't evaluate these directions myself. Perhaps someone using a Mac could verify them and put them on the website where folks downloading the .dmg will see it. Making this information more readily to Mac users should reduce the "I can't get Inkscape to work" troubles they seem to have. -s_tec

2 2

patch: crash on layer delete
by Ralf Stephan 31 Oct '05

31 Oct '05

Hello, regarding the recent 'crash on layer delete' I propose a patch that at least handles the simple case of that bug. As this is not my field, I'd ask you to review it. The idea is to setCurrentLayer() after the layer is deleted, and not before. setCurrentLayer() emits the layerChanged signal which is listened to by the layer selector widget. ralf --- verbs.cpp~ 2005-10-31 10:03:13.000000000 +0100 +++ verbs.cpp 2005-10-31 12:18:15.000000000 +0100 @@ -1123,11 +1123,11 @@ if (!survivor) { survivor = Inkscape::previous_layer(dt->currentRoot(), old_layer); } + old_layer->deleteObject(); + sp_object_unref(old_layer, NULL); if (survivor) { dt->setCurrentLayer(survivor); } - old_layer->deleteObject(); - sp_object_unref(old_layer, NULL); sp_document_done(SP_DT_DOCUMENT(dt));

2 1

Re: [Clipart] clipartbrowser 0.41
by Greg Steffensen 31 Oct '05

31 Oct '05

Ok, the clipartbrowser cvs module is now synched up again with the latest tarballs, and the cannonical download location for releases is now http://www.python.org/pypi/clipartbrowser (it used to be on Berlios). If you download a release (they work better than ever :), install using "python setup.py install" (assuming you have the dependencies met). The "easy_install" method mentioned earlier works too. Greg On 9/20/05, Greg Steffensen <greg.steffensen@...400...> wrote: > > I didn't explain the technical situtation with the install method very > well. Basically, if I write a file "setup.py" that defines certain > information, users should be able to execute "python setup.py install" to > install the program correctly, and I should be able to execute commands > "python setup.py bdist_win" to create actual win32 EXE files containing > pretty automated installer GUIs. In addition, users will be able to install > the way I described in the release email as well. In general, setup.pyuses the "setuptools" which are the successor to the "distutils", which is > the distribution/installation tools included with python. I'm still not > certain whether I should use the setuptools or the older distutils, and am > figuring stuff like that out, but they're both supposed to be the > python-friendly replacement for Makefiles, and I'm trying to reorganize the > code to do stuff the "correct" way. > > Greg > > On 9/20/05, Greg Steffensen <greg.steffensen@...400...> wrote: > > > > Hey, yeah, this is kinda what I was referring when I said that the > > packaging isn't as polished as it needs to be. For what its worth, most of > > the changes (bug fixes, performance improvements) actually are committed to > > CVS; I just didn't commit commit the change in the install method simply > > because I was having trouble figuring out how to get files like COPYING, > > NEWS, etc. included using that method, and I didn't want to put a totally > > messed up version into CVS. Basically, I actually really released more on a > > whim, not because I'd worked everything out, but because I'd gone too long > > without releasing code, wanted people to know that work was ongoing. So, at > > a minimum, I should have indicated that this was definitely a development > > release. > > > > So, solutions... in terms of where formal releases (formal tarballs) > > should go, I actually think the python package index is definitely the > > correct place... its now the "official" place where the python folk > > reccomend that python projects be listed, and it will offer a number of > > technical advantages, including publication of new releases, and > > installation automation. I'm very happy keeping CVS at Inkscape though (not > > that PyPI offers CVS anyway), and treating the code as a subproject of > > Inkscape. But as long as it offers the ability to be installed standalone as > > well, I'd like to make standalone tarballs available from PyPI. In general, > > I'm learning more about python, and am trying to make my practices for > > installation, etc. match accepted practice. > > > > I'll try to get CVS synched up with the current state of the project > > very soon; I'm still having some trouble with the python installation tools, > > and have sent an email to their mailing list for advice, but haven't heard > > back yet. Once I know how to get the documentation files included, and where > > to put the main install file (setup.py), I'll get that committed and CVS > > will contain the official version of the code gain. Again, I would have > > waited until this was done to release, but just felt bad about having gone > > too long without releasing already, and was overeager to get something out. > > > > Anyway, does this cover the stuff you were concerned about? If not, let > > me know. Also, you may recall that someone posted a link to a mockup for a > > very similar project that they'd found, and I replied that this was "very > > depressing" :). I used that mockup to develop version 0.4, and emailed > > the designer (as the community suggested); he's finally written me back, and > > said he's interested in contributing, which rocks. Later, > > > > Greg > > > > > > On 9/19/05, Jon Phillips <jon@...235...> wrote: > > > > > > <offlist /> > > > > > > On Mon, 2005-09-19 at 05:20 -0400, Greg Steffensen wrote: > > > > I'm releasing another version of the clip art browser. Lots of bug > > > > fixes, some performance improvements, code cleanups, and a new > > > > installation procedure (yes, yet another). > > > > > > > > Installation is now 3 steps, and hopefully easier than ever: > > > > > > > > 1) Satisfy the dependencies of Python 2.4 and PyGTK 2.6. > > > > > > > > 2) Download and run the following script, which installs the python > > > > setuptools: > > > > > > > > http://peak.telecommunity.com/dist/ez_setup.py > > > > > > > > 3) Run "easy_install clipartbrowser" > > > > > > > > As you can see, I've switched the installation system from Makefiles > > > > > > > to the python setuptools. The setuptools are still under development > > > > (they're apparently going to be included in python 2.5), and I'm > > > still > > > > very much learning how to use them, so the packaging isn't as > > > polished > > > > as it needs to be, but this makes installation a breeze. Eventually, > > > > it should allow arbitrary python dependencies to be intelligently > > > > installed as well using the same process. > > > > > > > > I'm still working on getting this integrated back into the Inkscape > > > > effects menu, that's next on the list. As always, feedback greatly > > > > appreciated. > > > > > > > Greg, > > > > > > I thought we were going to do the next release from Inkscape CVS? What > > > you have done is a non-standard release of the project now that we > > > consolidated the code into Inkscape. > > > > > > Also, you have now used a non-standard approach to packaging which > > > then > > > again breaks what users and developers are use to. > > > > > > The major thing though is that we were going to do a solid release > > > which > > > means you need to involve the people who would contribute to this in > > > the > > > release to help smooth out the bugs. > > > > > > Instead, now you have released from another location thus confusing > > > users and developers more. The problem with this is that it is not > > > pro-community. > > > > > > Anyhow, I'm curious what you think and also want to get these things > > > on > > > track with the community. > > > > > > It is so great that you are rocking these changes and I'm so proud of > > > what you are doing so I don't want to dismay you, but really I think > > > the > > > next release has to be really on track with the community and we > > > should > > > coordinate it, push it out through the sourceforge system we have in > > > place, and push the press release globally. > > > > > > Without these procedures, honestly, it is not very likely that others > > > will help on development nor use your work. > > > > > > So, lets talk some more and sort these things out. > > > > > > 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<http://www.openclipart.org> > > > ) > > > > > > > > >

5 8

a couple of "effects"
by unknown＠example.com 30 Oct '05

30 Oct '05

Yesterday at Bryce's prompting I wrote a little effect that makes the selected objects a <marker /> in the <defs /> of the current document. Inkscape doesn't really seem to deal well with this but there is some code there to work with if anyone has ideas on how it could be used. I think Bryce made some modifications to his copy as well. Because there is a lack of interactivity you must position the connection point of the marker at (0,0). And there are likely a few other nuances I haven't thought of yet. http://www.ekips.org/comp/inkscape/inx/makemarker.tar.gz I also wrote a python script that embeds images in SVG because I didn't feel comfortable editing the perl script. It functions as an effect and can be called from the effects menu. If objects are selected it will only replace the images in the selection, otherwise it should replace the images in the entire document (although in retrospect I didn't test that, oops). Hopefully this effect will be obsolete soon and we can have "embed all" in a menu and "embed" in the images' context menu. http://www.ekips.org/comp/inkscape/inx/embedimage.tar.gz Aaron Spike

4 3

Additional info regarding Inkscape CVS compile error
by Steven P. Ulrick 29 Oct '05

29 Oct '05

Hello, Everyone For the past few days, I had been getting error messages like this when running "make" on Inkscape from CVS: file=./`echo it | sed 's,.*/,,'`.gmo \ && rm -f $file && /usr/bin/msgfmt -c -o $file it.po it.po:1989: message catalog has plural form translations... it.po:10: ...but header entry lacks a "plural=EXPRESSION" attribute it.po:1989: message catalog has plural form translations... it.po:10: ...but header entry lacks a "nplurals=INTEGER" attribute make[2]: *** [it.gmo] Error 1 make[2]: Leaving directory `/home/steve/CVS/Assorted-CVS-Modules/inkscape/po' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/home/steve/CVS/Assorted-CVS-Modules/inkscape' make: *** [all] Error 2 For a while, I tried going back to old revisions of the *.po file that "make" seemed to choke on. Sometimes I would get the same error, but on a different *.po file. In each different error message, the numbers next to the name of the *.po file that was compiling when "make" crashed was different. Here is a new error message. Now "make" is choking on sr.po. I haven't seen it bomb on that one before: Making all in po make[2]: Entering directory `/home/steve/CVS/Assorted-CVS-Modules/inkscape/po' file=./`echo sr | sed 's,.*/,,'`.gmo \ && rm -f $file && /usr/bin/msgfmt -c -o $file sr.po sr.po:3224: format specifications in 'msgid' and 'msgstr' for argument 1 are not the same sr.po:4774: format specifications in 'msgid' and 'msgstr[0]' for argument 1 are not the same /usr/bin/msgfmt: found 2 fatal errors make[2]: *** [sr.gmo] Error 1 make[2]: Leaving directory `/home/steve/CVS/Assorted-CVS-Modules/inkscape/po' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/home/steve/CVS/Assorted-CVS-Modules/inkscape' make: *** [all] Error 2 Further, I noticed the following in the output for "configure": Running aclocal-1.7 ... Running autoheader ... Running libtoolize --copy --force ... Running automake-1.7 --add-missing ... Running autoconf ... Running glib-gettextize --copy --force ... Copying file mkinstalldirs Copying file po/Makefile.in.in Running intltoolize --copy --force --automake ... patching file po/Makefile.in.in Hunk #4 FAILED at 95. Hunk #5 succeeded at 194 (offset 15 lines). 1 out of 6 hunks FAILED -- saving rejects to file po/Makefile.in.in.rej This is the error that I posted that I referred to in the response to Erik's "What is happening?" message. I have reported this to the list, but no one apparently has seen it yet, or has knowledge about the possible cause. So that you can get all of the information that I have sent to the list about this problem, here is a link to my original thread in the archives: http://sourceforge.net/mailarchive/forum.php?thread_id=8808340&forum_id=360… As always, if you need any more information, let me know. Oh yeah, I said it elsewhere, but I am attempting to compile this on Fedora Core 3. Steven P. Ulrick

2 3

0.43 status as at Oct12
by Peter Moulder 29 Oct '05

29 Oct '05

SPDesktop now looks in good shape, which is good as that initially looked like the major show stopper. (Fixes thanks to rwst with some GC work from mental.) Inkboard -------- A couple of people have reported either difficulty in connecting: bryce made a bug report, and sciboy today mentioned similar problems -- see 17:52 and previous at http://inkscape.gristle.org/2005-10-11.txt. sciboy says that it works if connecting to gristle.org, but that none of jabber.org, jabber.com, or jabber.org.au work. Later that day, Michael & sciboy did get a connection. Issues: - Lots of output on the terminal, should be turned off for general use before release. (If these messages are still useful to developers, then wrap the status messages in a macro or function that does nothing in the main version but that interested developers can "switch on".) I wouldn't want to release with these status messages: e.g. because they mask real errors that we'd want to be reported. - Michael & sciboy's session got out of sync somehow: at some point sciboy cleared everything but a star was left on michael's session. Not a show-stopper: this was the only thing we were aware of being out of sync. - Undo is global, i.e. it often undoes what the other person did rather than what the undoer did. This is a usability problem, but changing this is non-trivial, so I don't consider this a show-stopper. Btw, Michael was using MacOS X, so it seems that libloudmouth works acceptably there (as well as on Linux/Gnu). I gather Windows still doesn't work acceptably well, due to threading differences. sciboy did experience a segfault later in his session with Michael (18:47 in the gristle log), though he didn't have a backtrace at the time. My feeling is that if the crash is specific to inkboard then we not consider it a show-stopper, but just make a note of it in the release notes. Connectors ---------- Progress has been made. The bug tracker no longer has any connector bugs marked with priority >= 8. We still have some strange behaviour. Michael comments: "The only known connector crash now is the one bryce reported, which I believe is fixed but he hasn't yet confirmed. I used to be able to reproduce it, but after fixing some other problems I now longer can. Once this sp_use_update fix is in that Bulia and I were just discussing (give me 20 minutes), then I think it's releasable. The other bugs are related to overlapping objects and are things I need to rework in libavoid after the release. The one remaining thing I'd like to look at if it's easy is the slowdown moving endpoints with the xml editor open." menus.xml localization ---------------------- In principle fixed (there's one huge menus-skeleton.h string to be translated), but in practice none of the translations has been updated. It would be nice if we went back to smaller strings to facilitate keeping the translations up-to-date (and so that we could re-use the existing translations). Personally I'm still inclined to treat this as a show-stopper, particularly as it looks easy enough to fix. Other ----- Sorting bug tracker items (http://sourceforge.net/tracker/?group_id=93438&atid=604306) by decreasing priority shows the following marked at priority 9: - a text-anchor regression (cyreve, can you comment?); - the corrupted about.svg problem I posted on just now; - a bunch of windows-specific bugs that I haven't looked at: I'll leave these for our Windows devs to decide on; - "unexpected printout size", where we're wrongly converting to postscript at 90px/inch or 72px/inch rather than 120px/inch. Personally I wouldn't let this one hold up the release. pjrm.

19 38