Ted,
Will you please push a Pre3? This will be the last pre-release before the final release. Thank you!
-=-=-
Everyone else,
This is the new thread to bring up any remaining blocker bugs, as it seems like almost all have been addressed. Let's also not forget, anything that can be lived with for a short period should not be considered a blocker. We do want to get 0.47 out, so we can get onto what can hopefully be a relatively short cycle for 0.48 (and then however many point releases we need until 0.49).
Cheers, Josh
On Thu, 2009-09-24 at 19:49 -0700, Josh Andler wrote:
Ted,
Will you please push a Pre3? This will be the last pre-release before the final release. Thank you!
-=-=-
Everyone else,
This is the new thread to bring up any remaining blocker bugs, as it seems like almost all have been addressed. Let's also not forget, anything that can be lived with for a short period should not be considered a blocker. We do want to get 0.47 out, so we can get onto what can hopefully be a relatively short cycle for 0.48 (and then however many point releases we need until 0.49).
Two important things left to fix:
1. Bug 382313 Image Filter broken for external images.
The problem is with the matrix inversion routine in lib2geom which compares the matrix determinant to a number that is too large to determine if the matrix is singular. The thread on the lib2geom-devel mailing list about this problem has died out. Unless they speak up with a fix, I can propose the attached patch which I've used for the past month to get around the problem.
2. Bug 179452 export to openclipart.org in dev version
Export to OCAL appears to be broken. The entry for this should be removed from the File menu. It would be quite an embarrassing thing to have a main menu item not work. (I would propose permanently removing this feature as using the OCAL web interface to upload pictures is more practical and is not subject to changes on OCAL's end.)
Tav
On Fri, Sep 25, 2009 at 12:21 PM, Tavmjong Bah <tavmjong@...8...> wrote:
On Thu, 2009-09-24 at 19:49 -0700, Josh Andler wrote:
Ted,
Will you please push a Pre3? This will be the last pre-release before the final release. Thank you!
-=-=-
Everyone else,
This is the new thread to bring up any remaining blocker bugs, as it seems like almost all have been addressed. Let's also not forget, anything that can be lived with for a short period should not be considered a blocker. We do want to get 0.47 out, so we can get onto what can hopefully be a relatively short cycle for 0.48 (and then however many point releases we need until 0.49).
Two important things left to fix:
- Bug 382313 Image Filter broken for external images.
The problem is with the matrix inversion routine in lib2geom which compares the matrix determinant to a number that is too large to determine if the matrix is singular. The thread on the lib2geom-devel mailing list about this problem has died out. Unless they speak up with a fix, I can propose the attached patch which I've used for the past month to get around the problem.
- Bug 179452 export to openclipart.org in dev version
Export to OCAL appears to be broken. The entry for this should be removed from the File menu. It would be quite an embarrassing thing to have a main menu item not work. (I would propose permanently removing this feature as using the OCAL web interface to upload pictures is more practical and is not subject to changes on OCAL's end.)
Tav
I vote for - Huge color blobs in pdf output https://bugs.launchpad.net/inkscape/+bug/248721 I provided a patch just 5 min ago.
what makes me scared is the huge number of open pdf and crash reports ...
Adib.
Come build with us! The BlackBerry® Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9-12, 2009. Register now! http://p.sf.net/sfu/devconf _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
On Fri, Sep 25, 2009 at 3:38 AM, the Adib <theadib@...1439...> wrote:
I vote for
- Huge color blobs in pdf output
https://bugs.launchpad.net/inkscape/+bug/248721 I provided a patch just 5 min ago.
what makes me scared is the huge number of open pdf and crash reports ...
Patch committed. Thanks Adib!
A lot of the PDF bugs can honestly wait for 0.48 because the PDF support in 0.47 is SO much better than 0.46 already.
Crash reports are never good, but we always seem to have them. The more important of the crash bugs are those in which the autosave does not get done. Data loss is definitely more of a priority than just crashing (not to say that crashing is really acceptable either).
Adib, any chance you want to take a look the first bug Tavmjong mentioned? (382313)
If no one steps up to the second one Tavmjong mentioned (the OCAL export bug), I find the solution proposed would be good if no one is planning to actively maintain this feature... it's like another whiteboard issue imho.
Cheers, Josh
2009/9/25 Tavmjong Bah <tavmjong@...8...>
- Bug 179452 export to openclipart.org in dev version
Export to OCAL appears to be broken. The entry for this should be removed from the File menu. It would be quite an embarrassing thing to have a main menu item not work. (I would propose permanently removing this feature as using the OCAL web interface to upload pictures is more practical and is not subject to changes on OCAL's end.)
Tav+1, the OCAL code uses GnomeVFS which is long deprecated. In 0.49 we should
replace passing filenames in the extension system with passing open GIO streams, this way we can save to remote locations. As an intermediate step we can remove all GnomeVFS code and rely on gvfs-mount (e.g. the thing in ~/.gvfs). It just requires some tinkering with the GTK file dialogs and we get paths that can be opened like regular local files using POSIX and standard library functions).
Regards, Krzysztof
Tav,
from my standpoint your patch is (mathematically) not correct. The geometrical epsilon is not related to mathematical significance. The epsilon must be considered somewhere else but not in matrix_inverse operation. I attached a (similar) patch and more explanation in the launchpad.
HTH, theAdib.
On Fri, Sep 25, 2009 at 12:21 PM, Tavmjong Bah <tavmjong@...8...> wrote:
On Thu, 2009-09-24 at 19:49 -0700, Josh Andler wrote:
Ted,
Will you please push a Pre3? This will be the last pre-release before the final release. Thank you!
-=-=-
Everyone else,
This is the new thread to bring up any remaining blocker bugs, as it seems like almost all have been addressed. Let's also not forget, anything that can be lived with for a short period should not be considered a blocker. We do want to get 0.47 out, so we can get onto what can hopefully be a relatively short cycle for 0.48 (and then however many point releases we need until 0.49).
Two important things left to fix:
- Bug 382313 Image Filter broken for external images.
The problem is with the matrix inversion routine in lib2geom which compares the matrix determinant to a number that is too large to determine if the matrix is singular. The thread on the lib2geom-devel mailing list about this problem has died out. Unless they speak up with a fix, I can propose the attached patch which I've used for the past month to get around the problem.
- Bug 179452 export to openclipart.org in dev version
Export to OCAL appears to be broken. The entry for this should be removed from the File menu. It would be quite an embarrassing thing to have a main menu item not work. (I would propose permanently removing this feature as using the OCAL web interface to upload pictures is more practical and is not subject to changes on OCAL's end.)
Tav
Come build with us! The BlackBerry® Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9-12, 2009. Register now! http://p.sf.net/sfu/devconf _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
On Sun, 2009-09-27 at 18:05 +0200, the Adib wrote:
Tav,
from my standpoint your patch is (mathematically) not correct. The geometrical epsilon is not related to mathematical significance. The epsilon must be considered somewhere else but not in matrix_inverse operation. I attached a (similar) patch and more explanation in the launchpad.
HTH, theAdib.
You won't get any argument from me that what you did is correct. I was just aiming to get something in that was better than what was there without having to understand why EPSILON was being used in the first place. Thanks for getting in pushed through.
(Now if we can only get Export to OCAL disabled I'll be really happy.)
Tav
Hi guys,
Be sure to forward any changes to our copy of 2geom to lib2geom itself! After 0.47 release, we'll probably update our copy with the latest 2geom and your changes will be lost. (see the readme file in the 2geom dir)
Ciao, Johan
-----Original Message----- From: Tavmjong Bah [mailto:tavmjong@...8...] Sent: zondag 27 september 2009 21:17 To: the Adib Cc: Inkscape Devel List; Ted Gould Subject: Re: [Inkscape-devel] Call for Pre3
On Sun, 2009-09-27 at 18:05 +0200, the Adib wrote:
Tav,
from my standpoint your patch is (mathematically) not correct. The geometrical epsilon is not related to mathematical significance. The epsilon must be considered somewhere else but not in matrix_inverse operation. I attached a (similar) patch and more explanation in the launchpad.
HTH, theAdib.
You won't get any argument from me that what you did is correct. I was just aiming to get something in that was better than what was there without having to understand why EPSILON was being used in the first place. Thanks for getting in pushed through.
(Now if we can only get Export to OCAL disabled I'll be really happy.)
Tav
Come build with us! The BlackBerry® Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9-12, 2009. Register now! http://p.sf.net/sfu/devconf _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
Johan,
Did you commit this on the 2geom end or submit it for review?
Cheers, Josh
On Mon, 2009-09-28 at 02:08 +0200, J.B.C.Engelen@...1578... wrote:
Hi guys,
Be sure to forward any changes to our copy of 2geom to lib2geom itself! After 0.47 release, we'll probably update our copy with the latest 2geom and your changes will be lost. (see the readme file in the 2geom dir)
Ciao, Johan
-----Original Message----- From: Tavmjong Bah [mailto:tavmjong@...8...] Sent: zondag 27 september 2009 21:17 To: the Adib Cc: Inkscape Devel List; Ted Gould Subject: Re: [Inkscape-devel] Call for Pre3
On Sun, 2009-09-27 at 18:05 +0200, the Adib wrote:
Tav,
from my standpoint your patch is (mathematically) not correct. The geometrical epsilon is not related to mathematical significance. The epsilon must be considered somewhere else but not in matrix_inverse operation. I attached a (similar) patch and more explanation in the launchpad.
HTH, theAdib.
You won't get any argument from me that what you did is correct. I was just aiming to get something in that was better than what was there without having to understand why EPSILON was being used in the first place. Thanks for getting in pushed through.
(Now if we can only get Export to OCAL disabled I'll be really happy.)
Tav
Come build with us! The BlackBerry® Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9-12, 2009. Register now! http://p.sf.net/sfu/devconf _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
Come build with us! The BlackBerry® Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9-12, 2009. Register now! http://p.sf.net/sfu/devconf _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
No I didn't. For any 2geom patch, I much rather have the original author(s) take care of it.
-----Original Message----- From: Joshua A. Andler [mailto:scislac@...400...] Sent: Monday, September 28, 2009 02:24 To: Engelen, J.B.C. (Johan) Cc: tavmjong@...8...; theadib@...1439...; inkscape-devel@lists.sourceforge.net Subject: Re: [Inkscape-devel] Call for Pre3
Johan,
Did you commit this on the 2geom end or submit it for review?
Cheers, Josh
On Mon, 2009-09-28 at 02:08 +0200, J.B.C.Engelen@...1578... wrote:
Hi guys,
Be sure to forward any changes to our copy of 2geom to
lib2geom itself!
After 0.47 release, we'll probably update our copy with the latest 2geom and your changes will be lost. (see the readme file in the 2geom dir)
Ciao, Johan
-----Original Message----- From: Tavmjong Bah [mailto:tavmjong@...8...] Sent: zondag 27 september 2009 21:17 To: the Adib Cc: Inkscape Devel List; Ted Gould Subject: Re: [Inkscape-devel] Call for Pre3
On Sun, 2009-09-27 at 18:05 +0200, the Adib wrote:
Tav,
from my standpoint your patch is (mathematically) not correct. The geometrical epsilon is not related to mathematical
significance.
The epsilon must be considered somewhere else but not in matrix_inverse operation. I attached a (similar) patch and more explanation in
the launchpad.
HTH, theAdib.
You won't get any argument from me that what you did is
correct. I
was just aiming to get something in that was better than what was there without having to understand why EPSILON was being
used in the
first place. Thanks for getting in pushed through.
(Now if we can only get Export to OCAL disabled I'll be really happy.)
Tav
Come build with us! The BlackBerry® Developer
Conference in SF,
CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9-12, 2009. Register now! http://p.sf.net/sfu/devconf _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
-------- Come build with us! The BlackBerry® Developer
Conference
in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile
applications
to market and stay ahead of the curve. Join us from
November 9-12,
- Register now! http://p.sf.net/sfu/devconf
Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
On Fri, Sep 25, 2009 at 6:49 AM, Josh Andler wrote:
This is the new thread to bring up any remaining blocker bugs, as it seems like almost all have been addressed.
https://bugs.launchpad.net/inkscape/+bug/214198
I confirm this on Linux now. Which makes color management broken on all/most supported platforms, it seems.
Alexandre
On Sep 25, 2009, at 3:49 AM, Alexandre Prokoudine wrote:
On Fri, Sep 25, 2009 at 6:49 AM, Josh Andler wrote:
This is the new thread to bring up any remaining blocker bugs, as it seems like almost all have been addressed.
https://bugs.launchpad.net/inkscape/+bug/214198
I confirm this on Linux now. Which makes color management broken on all/most supported platforms, it seems.
I have code that is *almost* done for Win32. The Linux and OS X specific code is in and works well.
So if anyone compiling on Win32 can lend a hand, this can be resolved very quickly.
On Sep 25, 2009, at 7:52 AM, Jon A. Cruz wrote:
I have code that is *almost* done for Win32. The Linux and OS X specific code is in and works well.
So if anyone compiling on Win32 can lend a hand, this can be resolved very quickly.
Can anyone on Win32 do a build of the latest and see if this was fixed as of SVN revision 22304?
What should be different is that the under the "Color Management" tab of the document properties the combo drop-down that lists profiles should have some by default on Windows once you click on the popup.
This is addressing bug #214198
https://bugs.launchpad.net/inkscape/+bug/214198
Thanks.
Hi Jon,
including lcms.h does not include windows.h, because our devlibs version has #define NON_WINDOWS 1 meaning a.o. that "#include windows.h" is skipped. I fixed the build in svn, by always including windows.h in color-profile.cpp.
Hope this helps keep svn head building.
Ciao, Johan
-----Original Message----- From: Jon A. Cruz [mailto:jon@...18...] Sent: zondag 27 september 2009 0:41 To: Inkscape Devel List Cc: Alexandre Prokoudine Subject: Re: [Inkscape-devel] Call for Pre3
On Sep 25, 2009, at 7:52 AM, Jon A. Cruz wrote:
I have code that is *almost* done for Win32. The Linux and OS X specific code is in and works well.
So if anyone compiling on Win32 can lend a hand, this can
be resolved
very quickly.
Can anyone on Win32 do a build of the latest and see if this was fixed as of SVN revision 22304?
What should be different is that the under the "Color Management" tab of the document properties the combo drop-down that lists profiles should have some by default on Windows once you click on the popup.
This is addressing bug #214198
https://bugs.launchpad.net/inkscape/+bug/214198
Thanks.
Come build with us! The BlackBerry® Developer Conference in SF, CA is the only developer event you need to attend this year. Jumpstart your developing skills, take BlackBerry mobile applications to market and stay ahead of the curve. Join us from November 9-12, 2009. Register now! http://p.sf.net/sfu/devconf _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
On Sep 26, 2009, at 6:12 PM, J.B.C.Engelen@...1578... wrote:
Hi Jon,
including lcms.h does not include windows.h, because our devlibs version has #define NON_WINDOWS 1 meaning a.o. that "#include windows.h" is skipped. I fixed the build in svn, by always including windows.h in color-profile.cpp.
Hope this helps keep svn head building.
Thanks.
What should be different is that the under the "Color Management" tab of the document properties the combo drop-down that lists profiles should have some by default on Windows once you click on the popup.
This is addressing bug #214198
After the latest modifications (r22312 on XP), I can no longer draw anything with any tool. Fill and stroke are "Unset" on the status bar, and nothing appears in the xml editor.
Preben
On Sep 26, 2009, at 11:20 PM, Preben Soeberg wrote:
After the latest modifications (r22312 on XP), I can no longer draw anything with any tool. Fill and stroke are "Unset" on the status bar, and nothing appears in the xml editor.
22312 does not even appear to be changes for Inkscape proper. Do you happen to have any idea which may have be the last revision that did work for you?
If you are building your own, you can check out from several revisions past and see if that works. If so, try a revision between those, etc. Knowing which revision actually caused your problems will me most helpful.
-----Original Message----- From: Jon A. Cruz [mailto:jon@...18...] Sent: 27 September, 2009 13:32 To: Preben Soeberg Cc: inkscape-devel@lists.sourceforge.net Subject: Re: [Inkscape-devel] Call for Pre3
On Sep 26, 2009, at 11:20 PM, Preben Soeberg wrote:
After the latest modifications (r22312 on XP), I can no longer draw anything with any tool. Fill and stroke are "Unset" on the status bar, and nothing appears in the xml editor.
22312 does not even appear to be changes for Inkscape proper. Do you happen to have any idea which may have be the last revision that did work for you?
If you are building your own, you can check out from several revisions past and see if that works. If so, try a revision between those, etc. Knowing which revision actually caused your problems will me most helpful.
Latest revision that worked for me was r22294, without checking those between r22294 and r22312.
r22312 works if I reuse the old preferences.xml from r22294, but it does not with a freshly generated preferences.xml.
I will have time look more at it in a couple of hours.
Preben
On 25/9/09 16:52, Jon A. Cruz wrote:
On Sep 25, 2009, at 3:49 AM, Alexandre Prokoudine wrote:
On Fri, Sep 25, 2009 at 6:49 AM, Josh Andler wrote:
This is the new thread to bring up any remaining blocker bugs, as it seems like almost all have been addressed.
https://bugs.launchpad.net/inkscape/+bug/214198
I confirm this on Linux now. Which makes color management broken on all/most supported platforms, it seems.
I have code that is *almost* done for Win32. The Linux and OS X specific code is in and works well.
Inkscape 0.47pre2 and r22312 crash when using a linked color profile with spaces like 'Black & White' and clicking in one of the RGB sliders in the 'Fill & Stroke > Fill > CMS' tab.
I added a comment including crash report to bug #405143 “implement validator for icc color profile names” https://bugs.launchpad.net/inkscape/+bug/405143, before I realized that it happens with both 'Black---White' (r22312) and 'Black & White' (0.47pre2-3) internal name versions. It seems this crash happens independent of r22260 and r22309. Should I open a separate bug?
~suv
On Sep 27, 2009, at 4:50 PM, ~suv wrote:
I added a comment including crash report to bug #405143 “implement validator for icc color profile names” https://bugs.launchpad.net/inkscape/+bug/405143, before I realized that it happens with both 'Black---White' (r22312) and 'Black & White' (0.47pre2-3) internal name versions. It seems this crash happens independent of r22260 and r22309. Should I open a separate bug?
Yes. That definitely sounds like a new bug.
The existing work on 405143 was observed to fail on multiple systems (mine, Juca's, etc.), then after the fix was observed to work.
On 28/9/09 03:59, Jon A. Cruz wrote:
On Sep 27, 2009, at 4:50 PM, ~suv wrote:
I added a comment including crash report to bug #405143 “implement validator for icc color profile names” https://bugs.launchpad.net/inkscape/+bug/405143, before I realized that it happens with both 'Black---White' (r22312) and 'Black & White' (0.47pre2-3) internal name versions. It seems this crash happens independent of r22260 and r22309. Should I open a separate bug?
Yes. That definitely sounds like a new bug.
Bug #437927 https://bugs.launchpad.net/inkscape/+bug/437927
The existing work on 405143 was observed to fail on multiple systems (mine, Juca's, etc.), then after the fix was observed to work.
~suv
On 25/9/09 04:49, Josh Andler wrote:
This is the new thread to bring up any remaining blocker bugs, as it seems like almost all have been addressed. Let's also not forget, anything that can be lived with for a short period should not be considered a blocker. We do want to get 0.47 out, so we can get onto what can hopefully be a relatively short cycle for 0.48 (and then however many point releases we need until 0.49).
one blocker:
Bend LPE crashes badly, incompatible to LPE generated with 0.46, but limited to the option 'Original Path is vertical'. Bug #436304 “bend path effect doesn't work in 0.47 but does in 0.46”: https://bugs.launchpad.net/inkscape/+bug/436304
questions:
1) Python extensions on Snow Leopard not working I cannot verify it myself but had yet another (e-mail) report last week about failing extensions on SL. Unfortunately there is not separate bug filed. Does Inkscape 0.47pre2-3 fix this? Bug #421305 “Inkscape does not run on Mac OSX 10.6”: https://bugs.launchpad.net/inkscape/+bug/421305
2) there are many 3dbox related bugs and even crashes: Has anyone yet had a chance to look at the patch attached in Bug #205667 “Inkscape crashes when using 3d box after attempting to render latex, gears, tree, barcode, maybe others.”: https://bugs.launchpad.net/inkscape/+bug/205667 ?
~suv
On Sat, 2009-09-26 at 06:09 +0200, ~suv wrote:
one blocker:
Bend LPE crashes badly, incompatible to LPE generated with 0.46, but limited to the option 'Original Path is vertical'. Bug #436304 “bend path effect doesn't work in 0.47 but does in 0.46”: https://bugs.launchpad.net/inkscape/+bug/436304
Fixed... woohoo!
questions:
- Python extensions on Snow Leopard not working
I cannot verify it myself but had yet another (e-mail) report last week about failing extensions on SL. Unfortunately there is not separate bug filed. Does Inkscape 0.47pre2-3 fix this? Bug #421305 “Inkscape does not run on Mac OSX 10.6”: https://bugs.launchpad.net/inkscape/+bug/421305
Sounds like a packaging issue, which can luckily be fixed without a point release if necessary. (Although if it could be taken care of before release that would rock).
- there are many 3dbox related bugs and even crashes: Has anyone yet
had a chance to look at the patch attached in Bug #205667 “Inkscape crashes when using 3d box after attempting to render latex, gears, tree, barcode, maybe others.”: https://bugs.launchpad.net/inkscape/+bug/205667 ?
The patch was reviewed, and could be an okay "quick fix"... However, bulia has a patch from Max which fixes this properly (and probably many other 3dbox issues). Unfortunately is not safe to commit at this late of a stage in release. Since we are shooting for a shorter dev cycle for 0.48, it's really just better to live with this for a short period.
Cheers, Josh
On Sep 24, 2009, at 7:49 PM, Josh Andler wrote:
This is the new thread to bring up any remaining blocker bugs, as it seems like almost all have been addressed. Let's also not forget, anything that can be lived with for a short period should not be considered a blocker.
I think I may have seen some issue with 0.47 Mac using the wrong XDG directory. I'll hunt that down a bit, but shouldn't block a main release. Definitely needs to be addressed for OS X though.
participants (10)
-
unknown＠example.com -
Alexandre Prokoudine -
Jon A. Cruz -
Josh Andler -
Joshua A. Andler -
Krzysztof Kosiński -
Preben Soeberg -
Tavmjong Bah -
the Adib -
~suv