Hi All,
Any thoughts about when we're going to release Inkscape 0.48.3? Coming at this from an Ubuntu maintenance perspective, the next LTS release (12.04 - "Precise Pangolin") will hit feature freeze on Feb 16, so it would be great to update the binary-package in the Ubuntu-main repository before that date. Large numbers of Ubuntu desktop users could potentially end up sticking with it for a couple of years, so I think it makes sense to squash as many bugs as possible!
Does anyone else agree with releasing 0.48.3 in the very near future? Is there any reason to hold back?
If we're going to do this, we'd need to release a source tarball by ~9 Feb. We've already committed a fair number of bug fixes in 0.48.3[1], so it would probably already be useful to users in its current form. However, perhaps a two-week bug-squashing session would be nice before we go ahead.
Best wishes,
AV
Hey,
I would be happy to see us release a 0.48.3 before the next LTS. I will also say that if there is any way we could get fixes for the 4 items tagged and not fixed for it as well as get the win32 devlibs sorted so users on that platform could have tablet fun again, that would be awesome. It really seems like the solution for the win32 devlibs may be looking into versions of all the libs that The GIMP and MyPaint are currently using with working tablet support. We mainly just need a win32 dev who has a tablet and a little time on their hands to experiment.
Cheers, Josh
On Mon, Jan 2, 2012 at 3:59 AM, Alex Valavanis <valavanisalex@...400...> wrote:
Hi All,
Any thoughts about when we're going to release Inkscape 0.48.3? Coming at this from an Ubuntu maintenance perspective, the next LTS release (12.04 - "Precise Pangolin") will hit feature freeze on Feb 16, so it would be great to update the binary-package in the Ubuntu-main repository before that date. Large numbers of Ubuntu desktop users could potentially end up sticking with it for a couple of years, so I think it makes sense to squash as many bugs as possible!
Does anyone else agree with releasing 0.48.3 in the very near future? Is there any reason to hold back?
If we're going to do this, we'd need to release a source tarball by ~9 Feb. We've already committed a fair number of bug fixes in 0.48.3[1], so it would probably already be useful to users in its current form. However, perhaps a two-week bug-squashing session would be nice before we go ahead.
Best wishes,
AV
[1] https://launchpad.net/inkscape/+milestone/0.48.3
Ridiculously easy VDI. With Citrix VDI-in-a-Box, you don't need a complex infrastructure or vast IT resources to deliver seamless, secure access to virtual desktops. With this all-in-one solution, easily deploy virtual desktops for less than the cost of PCs and save 60% on VDI infrastructure costs. Try it free! http://p.sf.net/sfu/Citrix-VDIinabox _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
Hi Josh (et al),
Any further thoughts about the timetable for the 0.48.3 release? Are we still on track for a release in time for the Ubuntu LTS feature freeze on 2012-02-16? (NB we'll probably need about a week for the Ubuntu packaging process after the 0.48.3 source package is released).
There are only two remaining open bugs that are tagged for 0.48.3 milestone[1,2]. Are we planning to backport any of the remaining "backport-proposed" bugs?[3] Is it worth having a mini backporting-sprint?
Thanks,
Alex
[1] https://bugs.launchpad.net/inkscape/+bug/819209 [2] https://bugs.launchpad.net/inkscape/+bug/788560 [3] https://bugs.launchpad.net/inkscape/+bugs?field.tag=backport-proposed
On 2 January 2012 16:47, Josh Andler <scislac@...400...> wrote:
Hey,
I would be happy to see us release a 0.48.3 before the next LTS. I will also say that if there is any way we could get fixes for the 4 items tagged and not fixed for it as well as get the win32 devlibs sorted so users on that platform could have tablet fun again, that would be awesome. It really seems like the solution for the win32 devlibs may be looking into versions of all the libs that The GIMP and MyPaint are currently using with working tablet support. We mainly just need a win32 dev who has a tablet and a little time on their hands to experiment.
Cheers, Josh
On Mon, Jan 2, 2012 at 3:59 AM, Alex Valavanis <valavanisalex@...400...> wrote:
Hi All,
Any thoughts about when we're going to release Inkscape 0.48.3? Coming at this from an Ubuntu maintenance perspective, the next LTS release (12.04 - "Precise Pangolin") will hit feature freeze on Feb 16, so it would be great to update the binary-package in the Ubuntu-main repository before that date. Large numbers of Ubuntu desktop users could potentially end up sticking with it for a couple of years, so I think it makes sense to squash as many bugs as possible!
Does anyone else agree with releasing 0.48.3 in the very near future? Is there any reason to hold back?
If we're going to do this, we'd need to release a source tarball by ~9 Feb. We've already committed a fair number of bug fixes in 0.48.3[1], so it would probably already be useful to users in its current form. However, perhaps a two-week bug-squashing session would be nice before we go ahead.
Best wishes,
AV
[1] https://launchpad.net/inkscape/+milestone/0.48.3
Ridiculously easy VDI. With Citrix VDI-in-a-Box, you don't need a complex infrastructure or vast IT resources to deliver seamless, secure access to virtual desktops. With this all-in-one solution, easily deploy virtual desktops for less than the cost of PCs and save 60% on VDI infrastructure costs. Try it free! http://p.sf.net/sfu/Citrix-VDIinabox _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
Hi Alex,
I talked to Ted and he said he should be good to cut a tarball for us prior to the freeze. I'd say that anyone who wants to get on backporting "safe" things should get going on it. Let's not worry about any official sprint, let's just have people with the time do what they can so we can give the go to Ted sooner rather than later. Any thoughts?
Cheers, Josh
On Mon, Jan 30, 2012 at 10:02 AM, Alex Valavanis <valavanisalex@...400...> wrote:
Hi Josh (et al),
Any further thoughts about the timetable for the 0.48.3 release? Are we still on track for a release in time for the Ubuntu LTS feature freeze on 2012-02-16? (NB we'll probably need about a week for the Ubuntu packaging process after the 0.48.3 source package is released).
There are only two remaining open bugs that are tagged for 0.48.3 milestone[1,2]. Are we planning to backport any of the remaining "backport-proposed" bugs?[3] Is it worth having a mini backporting-sprint?
Thanks,
Alex
[1] https://bugs.launchpad.net/inkscape/+bug/819209 [2] https://bugs.launchpad.net/inkscape/+bug/788560 [3] https://bugs.launchpad.net/inkscape/+bugs?field.tag=backport-proposed
On 2 January 2012 16:47, Josh Andler <scislac@...400...> wrote:
Hey,
I would be happy to see us release a 0.48.3 before the next LTS. I will also say that if there is any way we could get fixes for the 4 items tagged and not fixed for it as well as get the win32 devlibs sorted so users on that platform could have tablet fun again, that would be awesome. It really seems like the solution for the win32 devlibs may be looking into versions of all the libs that The GIMP and MyPaint are currently using with working tablet support. We mainly just need a win32 dev who has a tablet and a little time on their hands to experiment.
Cheers, Josh
On Mon, Jan 2, 2012 at 3:59 AM, Alex Valavanis <valavanisalex@...400...> wrote:
Hi All,
Any thoughts about when we're going to release Inkscape 0.48.3? Coming at this from an Ubuntu maintenance perspective, the next LTS release (12.04 - "Precise Pangolin") will hit feature freeze on Feb 16, so it would be great to update the binary-package in the Ubuntu-main repository before that date. Large numbers of Ubuntu desktop users could potentially end up sticking with it for a couple of years, so I think it makes sense to squash as many bugs as possible!
Does anyone else agree with releasing 0.48.3 in the very near future? Is there any reason to hold back?
If we're going to do this, we'd need to release a source tarball by ~9 Feb. We've already committed a fair number of bug fixes in 0.48.3[1], so it would probably already be useful to users in its current form. However, perhaps a two-week bug-squashing session would be nice before we go ahead.
Best wishes,
AV
[1] https://launchpad.net/inkscape/+milestone/0.48.3
Ridiculously easy VDI. With Citrix VDI-in-a-Box, you don't need a complex infrastructure or vast IT resources to deliver seamless, secure access to virtual desktops. With this all-in-one solution, easily deploy virtual desktops for less than the cost of PCs and save 60% on VDI infrastructure costs. Try it free! http://p.sf.net/sfu/Citrix-VDIinabox _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
2012/1/30 Alex Valavanis <valavanisalex@...400...>:
This one should be fixed now. The other one is about extensions on OSX, so I can't help there.
Regards, Krzysztof
Thank you very much!!!
Cheers, Josh
2012/1/31 Krzysztof Kosiński <tweenk.pl@...400...>:
2012/1/30 Alex Valavanis <valavanisalex@...400...>:
This one should be fixed now. The other one is about extensions on OSX, so I can't help there.
Regards, Krzysztof
Keep Your Developer Skills Current with LearnDevNow! The most comprehensive online learning library for Microsoft developers is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3, Metro Style Apps, more. Free future releases when you subscribe now! http://p.sf.net/sfu/learndevnow-d2d _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
2012/1/30 Alex Valavanis <valavanisalex@...400...>:
There are only two remaining open bugs that are tagged for 0.48.3 milestone[1,2]. Are we planning to backport any of the remaining "backport-proposed" bugs?[3] Is it worth having a mini backporting-sprint? (...) [3] https://bugs.launchpad.net/inkscape/+bugs?field.tag=backport-proposed
I have fixed the following bugs in the stable branch:
https://bugs.launchpad.net/inkscape/+bug/167419 https://bugs.launchpad.net/inkscape/+bug/168942 https://bugs.launchpad.net/inkscape/+bug/184341 https://bugs.launchpad.net/inkscape/+bug/239430 https://bugs.launchpad.net/inkscape/+bug/340123 https://bugs.launchpad.net/inkscape/+bug/663667 https://bugs.launchpad.net/inkscape/+bug/810503 https://bugs.launchpad.net/inkscape/+bug/834721 https://bugs.launchpad.net/inkscape/+bug/889172
There are 7 remaining "backport-proposed" bugs. One of them is the unfixed bug about extensions on OSX, which I can't work on.* The other 6 are related WMF/EMF export and all have fixes in trunk. I will try to backport the WMF/EMF fixes shortly.
The fix for #810503 and #834721 introduces new translatable strings. I think it's worth fixing these bugs even if this causes a few untranslated messages to appear. Tell me if you think these fixes should be reverted.
I also have a GTK update ready that I haven't committed to the devlibs yet, which reportedly should solve tablet problems. This will go in before I start working on the WMF/EMF bugs.
*) What is our position on the legality of OSX virtualization for personal use?
Regards, Krzysztof
On 01/02/2012 21:46, Krzysztof Kosiński wrote:
2012/1/30 Alex Valavanis <valavanisalex@...400...>:
There are only two remaining open bugs that are tagged for 0.48.3 milestone[1,2]. Are we planning to backport any of the remaining "backport-proposed" bugs?[3] Is it worth having a mini backporting-sprint? (...) [3] https://bugs.launchpad.net/inkscape/+bugs?field.tag=backport-proposed
There are 7 remaining "backport-proposed" bugs. One of them is the unfixed bug about extensions on OSX, which I can't work on.* The other 6 are related WMF/EMF export and all have fixes in trunk. I will try to backport the WMF/EMF fixes shortly.
The report has a patch which I have verified to work with the official package (0.48.2) on OS X Lion as well tested on Mac OS X Leopard to not have a negative effect on other platforms. The patch inserts a single line of code into a shell script inside the application bundle (no recompiling and repackaging needed to test).
It has been confirmed by two other users as well.
Committing to trunk should be save, and back-porting as well (I'd rather have raised importance to 'High' because it also affects several input/output extensions).
What is still missing in the 0.48.x branch and needs confirmation by Michael IMHO are the changes (in trunk) to one of the packaging scripts (packaging/macosx/osx-app.sh), used to fill the bundle with files from MacPorts (due to changed locations of the gdk-pixbuf configuration files and loaders): https://bugs.launchpad.net/inkscape/+bug/819209/comments/13 http://bazaar.launchpad.net/~inkscape.dev/inkscape/trunk/revision/10026
hth, ~suv
Would it be OK to backport the Synfig Animation Studio (.sif) exporter into 0.48.3? And maybe the Digimemo (.dhw) importer as well.
~Nikita
2012/2/1 ~suv <suv-sf@...58...>
On 01/02/2012 21:46, Krzysztof Kosiński wrote:
2012/1/30 Alex Valavanis <valavanisalex@...400...>:
There are only two remaining open bugs that are tagged for 0.48.3 milestone[1,2]. Are we planning to backport any of the remaining "backport-proposed" bugs?[3] Is it worth having a mini backporting-sprint? (...) [3] https://bugs.launchpad.net/inkscape/+bugs?field.tag=backport-proposed
There are 7 remaining "backport-proposed" bugs. One of them is the unfixed bug about extensions on OSX, which I can't work on.* The other 6 are related WMF/EMF export and all have fixes in trunk. I will try to backport the WMF/EMF fixes shortly.
The report has a patch which I have verified to work with the official package (0.48.2) on OS X Lion as well tested on Mac OS X Leopard to not have a negative effect on other platforms. The patch inserts a single line of code into a shell script inside the application bundle (no recompiling and repackaging needed to test).
It has been confirmed by two other users as well.
Committing to trunk should be save, and back-porting as well (I'd rather have raised importance to 'High' because it also affects several input/output extensions).
What is still missing in the 0.48.x branch and needs confirmation by Michael IMHO are the changes (in trunk) to one of the packaging scripts (packaging/macosx/osx-app.sh), used to fill the bundle with files from MacPorts (due to changed locations of the gdk-pixbuf configuration files and loaders): https://bugs.launchpad.net/inkscape/+bug/819209/comments/13 http://bazaar.launchpad.net/~inkscape.dev/inkscape/trunk/revision/10026
hth, ~suv
I have updated the devlibs to GTK 2.24.8. Please report whether it fixes tablet problems.
If the exporters for .sif and .dhw are new features that are not present in the 0.48 series, I'm afraid we can't add them. 0.48.3 is intended to be a bugfix-only release.
Regards, Krzysztof
2012/2/3 Krzysztof Kosiński <tweenk.pl@...400...>:
I have updated the devlibs to GTK 2.24.8. Please report whether it fixes tablet problems.
Can anyone on windows confirm or deny possible tablet fixes?
As an aside, we are looking pretty good to go for 0.48.3 atm. Last call for any patches which may have been missed or concerns that anyone has.
Cheers, Josh
Maybe a bit of a long shot, but if anyone feels like sorting out the build with lcms 2 and backporting it, that would be great. Bug #885324 [1] suggests that lcms 1 will be dropped in Ubuntu Precise.
I don't know really anything about lcms or colour management in general, but it looks to me as if the transition to lcms 2 mostly a simple find/replace job.
Anyone interested?
[1] https://bugs.launchpad.net/inkscape/+bug/885324
On 8 February 2012 01:57, Josh Andler <scislac@...400...> wrote:
2012/2/3 Krzysztof Kosiński <tweenk.pl@...400...>:
I have updated the devlibs to GTK 2.24.8. Please report whether it fixes tablet problems.
Can anyone on windows confirm or deny possible tablet fixes?
As an aside, we are looking pretty good to go for 0.48.3 atm. Last call for any patches which may have been missed or concerns that anyone has.
Cheers, Josh
Keep Your Developer Skills Current with LearnDevNow! The most comprehensive online learning library for Microsoft developers is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3, Metro Style Apps, more. Free future releases when you subscribe now! http://p.sf.net/sfu/learndevnow-d2d _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
Second round of backports done. Most of these are Windows-specific EMF issues. Also added some build fixes related to devlibs updates.
https://bugs.launchpad.net/inkscape/+bug/407394 https://bugs.launchpad.net/inkscape/+bug/681262 https://bugs.launchpad.net/inkscape/+bug/858369 https://bugs.launchpad.net/inkscape/+bug/888793 https://bugs.launchpad.net/inkscape/+bug/898538 https://bugs.launchpad.net/inkscape/+bug/919728
The list of bugs with proposed backports is now exhausted.
There are four milestoned bugs still in progress. https://launchpad.net/inkscape/+milestone/0.48.3
446842 can be ignored as it only manifests in console output. 594441 is partially fixed, ut there's still a minor regression. 427514 is a problem with mirroring data in docked dialogs belonging to different windows. 911079 is Ubuntu-specific. I'm not sure what to do with these.
Regards, Krzysztof
On Feb 8, 2012, at 2:33 AM, Alex Valavanis wrote:
Maybe a bit of a long shot, but if anyone feels like sorting out the build with lcms 2 and backporting it, that would be great. Bug #885324 [1] suggests that lcms 1 will be dropped in Ubuntu Precise.
I don't know really anything about lcms or colour management in general, but it looks to me as if the transition to lcms 2 mostly a simple find/replace job.
Anyone interested?
Unfortunately it's not just as simple as it might seem.
I have the configure.ac changes all set, but at least on OSX there are some runtime issues.
I'm poking about with Ubuntu 11.11 now... so there's a chance of getting some platforms in.
Hi,
De : Josh Andler <scislac@...400...> 2012/2/3 Krzysztof Kosiński <tweenk.pl@...400...>:
I have updated the devlibs to GTK 2.24.8. Please report whether it fixes tablet problems.
Can anyone on windows confirm or deny possible tablet fixes?
I'm currently testing the new devlibs with a bamboo fun tablet on XP, and it seems that some of the most annoying bugs, such as #845354 "Critical Wacom tablet glitches (0.48.2)" [https://bugs.launchpad.net/inkscape/+bug/845354], are fixed. More test in progress. Combined to the half-dozen UI bugs fixed, gtk 2.24.8 looks very good compared to the previous one.
Regards, -- Nicolas
On Sat, Feb 11, 2012 at 5:35 AM, Nicolas Dufour <nicoduf@...48...> wrote:
Hi,
De : Josh Andler <scislac@...400...> 2012/2/3 Krzysztof Kosiński <tweenk.pl@...400...>:
I have updated the devlibs to GTK 2.24.8. Please report whether it fixes tablet problems.
Can anyone on windows confirm or deny possible tablet fixes?
I'm currently testing the new devlibs with a bamboo fun tablet on XP, and it seems that some of the most annoying bugs, such as #845354 "Critical Wacom tablet glitches (0.48.2)" [https://bugs.launchpad.net/inkscape/+bug/845354], are fixed. More test in progress. Combined to the half-dozen UI bugs fixed, gtk 2.24.8 looks very good compared to the previous one.
Awesome! Thanks for the update... this is sounding promising. :)
Cheers, Josh
On 01/02/2012 22:06, ~suv wrote:
On 01/02/2012 21:46, Krzysztof Kosiński wrote:
2012/1/30 Alex Valavanis <valavanisalex@...400...>:
There are only two remaining open bugs that are tagged for 0.48.3 milestone[1,2]. Are we planning to backport any of the remaining "backport-proposed" bugs?[3] Is it worth having a mini backporting-sprint? (...) [3] https://bugs.launchpad.net/inkscape/+bugs?field.tag=backport-proposed
There are 7 remaining "backport-proposed" bugs. One of them is the unfixed bug about extensions on OSX, which I can't work on.* The other 6 are related WMF/EMF export and all have fixes in trunk. I will try to backport the WMF/EMF fixes shortly.
The report has a patch which I have verified to work with the official package (0.48.2) on OS X Lion as well tested on Mac OS X Leopard to not have a negative effect on other platforms. The patch inserts a single line of code into a shell script inside the application bundle (no recompiling and repackaging needed to test).
It has been confirmed by two other users as well.
Committing to trunk should be save, and back-porting as well (I'd rather have raised importance to 'High' because it also affects several input/output extensions).
What is still missing in the 0.48.x branch and needs confirmation by Michael IMHO are the changes (in trunk) to one of the packaging scripts (packaging/macosx/osx-app.sh), used to fill the bundle with files from MacPorts (due to changed locations of the gdk-pixbuf configuration files and loaders): https://bugs.launchpad.net/inkscape/+bug/819209/comments/13 http://bazaar.launchpad.net/~inkscape.dev/inkscape/trunk/revision/10026
Other issues related to OS X packaging only:
1) GTK+ 2.24.9 (current version in MacPorts) Due to a recent regression in GTK+ 2.24.9: Bug 543520 – Cups custom options are not sent properly: https://bugzilla.gnome.org/show_bug.cgi?id=543520#c39 Inkscape 0.48.2, 0.48.x and trunk crash with GTK+ 2.24.9 when trying to print (print preview and printing to file seem ok).
Before building and bundling the application package, I would recommend to revert to gtk2 @2.24.8 in MacPorts or rebuild and test the port using a local patch from the gtk-2.24 branch: http://git.gnome.org/browse/gtk+/commit/?h=gtk-2-24&id=66c99016d3d063aee0e00793da8f087c80172012
2) Cairo 1.10.2 in MacPorts Cairo (and cairo-devel) in MacPorts have xcb and xlib-xcb enabled unconditionally (the variant is fake for legacy reasons) while the cairo developers do not consider it as stable yet (planned for cairo 0.12 AFAIU). This is the reason why Inkscape.app crashes for users on Mac OS X 10.5.8 Leopard with Apple's default X11 installed (no xcb support) even though the bundle was built and packaged on Leopard [2].
'xlib-xcb' also causes rendering errors in the GUI (text overspilling into other widgets if auto-hinting is enabled [1], incorrect rendering of shades and gradients e.g. for scrollbar sliders and button states on Lion (not reported)). I would recommend to rebuild the cairo port locally and disable all xcb options before creating the application bundle.
(Personally, I'm using cairo 1.10.2 with '--enable-xlib-xcb' removed from the portfile (local repo) on OS X Lion, and haven't noticed any drawbacks due to the change).
3) gdk-pixbuf.loaders (from gdk-pixbuf-2.22.1) Inkscape.app (0.48.2) has a small regression compared to 0.48.1: it appears that due to a missing entry for SVG in the gdk-pixbuf.loaders file, Inkscape fails to load external SVG files linked to with the <image> tag (works ok with Inkscape.app (0.48.1)). For details please see comments #31 and #32 in Bug #168244 “image tag does not find SVG source” [3].
4) outdated custom theme file Bug #875577 “[osx] Clearlooks configuration option "menuitemstyle" is not supported and will be ignored” [4] has a patch for the custom theme file used by the bundled application, updating it to the newer version of the Clearlooks engine (certain old settings are no longer supported). Patch was done with intention to keep visual appearance unchanged.
I also disabled the button and menu icons (not used in native apps, AFAICT) - both can be re-enabled by changing the value to '1' for 'gtk-button-images' and 'gtk-menu-images', at the bottom of pre_gtkrc.
Please test and consider for inclusion.
hth, ~suv
[1] https://bugs.launchpad.net/inkscape/+bug/625668 [2] https://bugs.launchpad.net/inkscape/+bug/878368 [3] https://bugs.launchpad.net/inkscape/+bug/168244 [4] https://bugs.launchpad.net/inkscape/+bug/875577
On 03/02/2012 14:28, ~suv wrote:
On 01/02/2012 22:06, ~suv wrote:
On 01/02/2012 21:46, Krzysztof Kosiński wrote:
2012/1/30 Alex Valavanis <valavanisalex@...400...>:
There are only two remaining open bugs that are tagged for 0.48.3 milestone[1,2]. Are we planning to backport any of the remaining "backport-proposed" bugs?[3] Is it worth having a mini backporting-sprint? (...) [3] https://bugs.launchpad.net/inkscape/+bugs?field.tag=backport-proposed
There are 7 remaining "backport-proposed" bugs. One of them is the unfixed bug about extensions on OSX, which I can't work on.* The other 6 are related WMF/EMF export and all have fixes in trunk. I will try to backport the WMF/EMF fixes shortly.
The report has a patch which I have verified to work with the official package (0.48.2) on OS X Lion as well tested on Mac OS X Leopard to not have a negative effect on other platforms. The patch inserts a single line of code into a shell script inside the application bundle (no recompiling and repackaging needed to test).
It has been confirmed by two other users as well.
Committing to trunk should be save, and back-porting as well (I'd rather have raised importance to 'High' because it also affects several input/output extensions).
Done (r10951 in trunk, r9866 in 0.48.x)
What is still missing in the 0.48.x branch and needs confirmation by Michael IMHO are the changes (in trunk) to one of the packaging scripts (packaging/macosx/osx-app.sh), used to fill the bundle with files from MacPorts (due to changed locations of the gdk-pixbuf configuration files and loaders): https://bugs.launchpad.net/inkscape/+bug/819209/comments/13 http://bazaar.launchpad.net/~inkscape.dev/inkscape/trunk/revision/10026
Done (r9868 in 0.48.x)
Other issues related to OS X packaging only:
- GTK+ 2.24.9 (current version in MacPorts)
Due to a recent regression in GTK+ 2.24.9: Bug 543520 – Cups custom options are not sent properly: https://bugzilla.gnome.org/show_bug.cgi?id=543520#c39 Inkscape 0.48.2, 0.48.x and trunk crash with GTK+ 2.24.9 when trying to print (print preview and printing to file seem ok).
Before building and bundling the application package, I would recommend to revert to gtk2 @2.24.8 in MacPorts or rebuild and test the port using a local patch from the gtk-2.24 branch: http://git.gnome.org/browse/gtk+/commit/?h=gtk-2-24&id=66c99016d3d063aee0e00793da8f087c80172012
GTK+ 2.24.10 has been released in the meantime and is available in MacPorts. It includes the fix to avoid the crash when printing, but now instead of crashing, Inkscape (stable and trunk) fails to print at all for me (printing works in gtk-demo, and when printing a PDF file saved in Inkscape with Evince, all using the same GTK+ stack). Based on my (limited) tests [1] I would still recommend to use 2.24.8 for the application package - if printing directly from Inkscape is considered as important.
[1] GTK+/X11 2.24.10 on Lion, GTK+/Quartz 2.24.10 on Leopard Brother HL-2035 B&W Laser Printer (A4)
- Cairo 1.10.2 in MacPorts
Cairo (and cairo-devel) in MacPorts have xcb and xlib-xcb enabled unconditionally (the variant is fake for legacy reasons) while the cairo developers do not consider it as stable yet (planned for cairo 0.12 AFAIU). This is the reason why Inkscape.app crashes for users on Mac OS X 10.5.8 Leopard with Apple's default X11 installed (no xcb support) even though the bundle was built and packaged on Leopard [2].
'xlib-xcb' also causes rendering errors in the GUI (text overspilling into other widgets if auto-hinting is enabled [1], incorrect rendering of shades and gradients e.g. for scrollbar sliders and button states on Lion (not reported)). I would recommend to rebuild the cairo port locally and disable all xcb options before creating the application bundle.
(Personally, I'm using cairo 1.10.2 with '--enable-xlib-xcb' removed from the portfile (local repo) on OS X Lion, and haven't noticed any drawbacks due to the change).
- gdk-pixbuf.loaders (from gdk-pixbuf-2.22.1)
Inkscape.app (0.48.2) has a small regression compared to 0.48.1: it appears that due to a missing entry for SVG in the gdk-pixbuf.loaders file, Inkscape fails to load external SVG files linked to with the <image> tag (works ok with Inkscape.app (0.48.1)). For details please see comments #31 and #32 in Bug #168244 “image tag does not find SVG source” [3].
- outdated custom theme file
Bug #875577 “[osx] Clearlooks configuration option "menuitemstyle" is not supported and will be ignored” [4] has a patch for the custom theme file used by the bundled application, updating it to the newer version of the Clearlooks engine (certain old settings are no longer supported). Patch was done with intention to keep visual appearance unchanged.
I also disabled the button and menu icons (not used in native apps, AFAICT) - both can be re-enabled by changing the value to '1' for 'gtk-button-images' and 'gtk-menu-images', at the bottom of pre_gtkrc.
Please test and consider for inclusion.
[1] https://bugs.launchpad.net/inkscape/+bug/625668 [2] https://bugs.launchpad.net/inkscape/+bug/878368 [3] https://bugs.launchpad.net/inkscape/+bug/168244 [4] https://bugs.launchpad.net/inkscape/+bug/875577
On 11/02/2012 08:18, ~suv wrote:
On 03/02/2012 14:28, ~suv wrote:
Other issues related to OS X packaging only:
- GTK+ 2.24.9 (current version in MacPorts)
Due to a recent regression in GTK+ 2.24.9: Bug 543520 – Cups custom options are not sent properly: https://bugzilla.gnome.org/show_bug.cgi?id=543520#c39 Inkscape 0.48.2, 0.48.x and trunk crash with GTK+ 2.24.9 when trying to print (print preview and printing to file seem ok).
Before building and bundling the application package, I would recommend to revert to gtk2 @2.24.8 in MacPorts or rebuild and test the port using a local patch from the gtk-2.24 branch: http://git.gnome.org/browse/gtk+/commit/?h=gtk-2-24&id=66c99016d3d063aee0e00793da8f087c80172012
GTK+ 2.24.10 has been released in the meantime and is available in MacPorts. It includes the fix to avoid the crash when printing, but now instead of crashing, Inkscape (stable and trunk) fails to print at all for me (printing works in gtk-demo, and when printing a PDF file saved in Inkscape with Evince, all using the same GTK+ stack). Based on my (limited) tests [1] I would still recommend to use 2.24.8 for the application package - if printing directly from Inkscape is considered as important.
[1] GTK+/X11 2.24.10 on Lion, GTK+/Quartz 2.24.10 on Leopard Brother HL-2035 B&W Laser Printer (A4)
More details about the printing failure:
The error reported e.g. by CUPS 1.5.0 (OS X 10.7.2 Lion) when trying to print directly from Inkscape (stable and trunk) with GTK+ 2.24.10:
stopped "cgpdftoraster: cupsRasterInterpretPPD had an error: Page header uses unsupported values."
whereas Inkscape stable and trunk with GTK+ 2.24.8 printed just fine on the same system (no system updates since), without errors.
The print job was sent with these arguments:
argv[3]="Print test-A4-1.svg"
<snip>
argv[5]="InputSlot=Tray1 BRSleepTime=Default BRHalfTonePattern=Brother PageSize=Custom.Custom.595.28x841.89 number-up=1 TonerSaveMode=Off BRMediaType=Plain Resolution=300dpi job-uuid=urn:uuid:d674f959-7fa4-3a2c-6eb5-2274f8d444b1 job-originating-host-name=localhost time-at-creation=1329032636 time-at-processing=1329032636"
The repeated 'Custom' in "PageSize=Custom.Custom.595.28x841.89" seems odd to me, but I'm not familiar with CUPS nor PPDs to really know whether this page size argument is correct or not.
Has anyone else been able to test printing from Inkscape with GTK+ 2.24.10 yet?
~suv
participants (7)
-
Alex Valavanis -
Jon Cruz -
Josh Andler -
Krzysztof Kosiński -
Nicolas Dufour -
Nikita Kitaev -
~suv