Time to unpause 0.45 release?
Hi all,
A few weeks back we decided to put the release on pause to give time for additional work on the PDF extension. Could someone outline the status of this work, and if it is still necessary for this to be blocking the release?
If not, would anyone have an issue if we resumed the release process at this point? I find myself with some unexpected extra freetime this month that I could devote to assisting with the release.
Bryce
On Wed, 6 Dec 2006 12:57:57 -0800, Bryce Harrington <bryce@...961...> wrote:
A few weeks back we decided to put the release on pause to give time for additional work on the PDF extension. Could someone outline the status of this work, and if it is still necessary for this to be blocking the release?
At this point it looks like the most major issues I care about have been addressed on the Inkscape side -- I think it's making all the right cairo calls for what it needs to export at this point. The problem is that cairo still isn't up to snuff.
Aaron Spike (and to a much lesser extent myself) have been working the cairo end of things, but Carl may still be the best person to give an estimate of how much more work is stil required. One of the most important things is getting push_group working right in Cairo-PDF -- right now it flat out rasterizes them, rather than creating PDF transparency groups with vectors.
If not, would anyone have an issue if we resumed the release process at this point? I find myself with some unexpected extra freetime this month that I could devote to assisting with the release.
In principle I think we could probably go ahead, since now that things are (mostly) fixed up on our side, the PDF export will get better as cairo gets better -- for Windows and autopackage at least, we could ship updates with updated cairo libs.
The problem is that with current cairo, the result (while better in some important ways) is really broken in others, so on balance it's about as bad as the PDF export we shipped with 0.44 (which was so bad that it [almost?] wasn't worth shipping IMO). And we're likely to be stuck with current cairo on most Linux distros for a while.
-mental
MenTaLguY wrote:
Aaron Spike (and to a much lesser extent myself) have been working the cairo end of things, but Carl may still be the best person to give an estimate of how much more work is stil required. One of the most important things is getting push_group working right in Cairo-PDF -- right now it flat out rasterizes them, rather than creating PDF transparency groups with vectors.
So far I've only been working on getting my self acclimated so that I might be able to work on the cairo end of things. I wish I had more experience and expertise. And I wish I could promise good things with a reasonable timeline. But I'm only stepping in because I want this release to get out soon and I see a need that no one has been willing to fill as of yet.
In principle I think we could probably go ahead, since now that things are (mostly) fixed up on our side, the PDF export will get better as cairo gets better -- for Windows and autopackage at least, we could ship updates with updated cairo libs.
I plan to offer occasional development snapshot autopackages of cairo as I've been doing for inkscape. This should allow linux to run any inkscape against a new cairo (using LD_LIBRARY_PATH) without screwing with their distros cairo. It should also help us test PDF output as cairo improves. (Please note from now on all of our users will have cairo installed on their systems due to our requirement on gtk2.8. this is worth a separate thread.)
Aaron Spike
On Wed, Dec 06, 2006 at 02:33:14PM -0800, MenTaLguY wrote:
On Wed, 6 Dec 2006 12:57:57 -0800, Bryce Harrington <bryce@...961...> wrote:
A few weeks back we decided to put the release on pause to give time for additional work on the PDF extension. Could someone outline the status of this work, and if it is still necessary for this to be blocking the release?
At this point it looks like the most major issues I care about have been addressed on the Inkscape side -- I think it's making all the right cairo calls for what it needs to export at this point. The problem is that cairo still isn't up to snuff.
Okay, well let's go ahead then.
If not, would anyone have an issue if we resumed the release process at this point? I find myself with some unexpected extra freetime this month that I could devote to assisting with the release.
In principle I think we could probably go ahead, since now that things are (mostly) fixed up on our side, the PDF export will get better as cairo gets better -- for Windows and autopackage at least, we could ship updates with updated cairo libs.
The problem is that with current cairo, the result (while better in some important ways) is really broken in others, so on balance it's about as bad as the PDF export we shipped with 0.44 (which was so bad that it [almost?] wasn't worth shipping IMO). And we're likely to be stuck with current cairo on most Linux distros for a while.
Hmm, well at least it's a step in the right direction. Perhaps Cairo will catch up, and new packages can be generated. At least, now when users complain about the PDF inadequacies we have a good place to point them.
Bryce
On 12/6/06, Bryce Harrington <bryce@...961...> wrote:
At this point it looks like the most major issues I care about have been addressed on the Inkscape side -- I think it's making all the right cairo calls for what it needs to export at this point. The problem is that cairo still isn't up to snuff.
Okay, well let's go ahead then.
I don't think we can release as is. In 0.44, we had PDF export which supported transparency - it had bugs but it basically worked. Current Cairo-PDF export simply rasterizes it. It's not acceptable.
Mental, can you just restore the old PDF export that you removed? Then we can release with it (by the way it had several bugfixes compared to 0.44) and switch to Cairo for one of the next versions when it's progressed enough.
On Thu, 2006-12-07 at 22:21 -0400, bulia byak wrote:
Mental, can you just restore the old PDF export that you removed? Then we can release with it (by the way it had several bugfixes compared to 0.44) and switch to Cairo for one of the next versions when it's progressed enough.
Ok, per our conversation on Jabber, I'd like to do it on the release branch after the freeze.
-mental
With the decision regarding the PDF work, we're ready to proceed with the release.
At this point we can move into the 'Frost' portion of the release, to focus on bugfixes:
Frost: *** + Wrap up development work + Select an about screen + Release Notes is 90% filled in + Bug Hunt: 300 points + Post inkscape-0.45pre0.tar.gz
Feature Freeze: + SVN is branched + No further development work on release branch + Focus on critical bug fixing + Finalize all tutorials, docs, extensions, etc. + Translators create translations + Inkscape must pass 'make distcheck' + Packagers encouraged to test package creation + Post inkscape-0.45pre1.tar.gz
Hard Freeze: + Focus on 'MUST FIX' bug fixing + Finalize translations + Finalize release notes + Test installation of packages for all major platforms + Post inkscape-0.45pre2.tar.gz
Release + Final verification of late work, packaging, release notes, etc. + Post inkscape-0.45.tar.gz + Post packages for major platforms + Post official announcements
Bryce
Hi all,
We're ready to move into the bug fixing portion of the release. Traditionally, we specify a point target for us to reach, and award us points based on the severity of the bug: 3, 6, or 9 points for low, medium, and high.
But this time, to encourage work on old bugs, lets give double points to fixes for any bugs that have been open for a year or more.
Bug Hunt Goal: 300 points Current: 0 points Low: 3 - 6 pts Med: 6 - 12 pts High: 9 - 18 pts
We especially need testers (and bug fixers) for Windows and OSX. A lot of bugs have accumulated that only show on those platforms, so they are difficult to test.
If you are a Windows or OSX user, you too can be a big help, simply by browsing the bug tracker and see if you can reproduce the bug, and then add a comment on the bug one way or the other. No technical experience required. :-)
Thanks, Bryce
Bug Pts Desc Assigned To ------------------------------------------------------------------------ 1576258 3 Color artifact in SVN gaussian blur nobody ------------------------------------------------------------------------ Total 3
On Sat, Dec 09, 2006 at 10:59:28AM -0800, Bryce Harrington wrote:
Hi all,
We're ready to move into the bug fixing portion of the release. Traditionally, we specify a point target for us to reach, and award us points based on the severity of the bug: 3, 6, or 9 points for low, medium, and high.
But this time, to encourage work on old bugs, lets give double points to fixes for any bugs that have been open for a year or more.
Bug Hunt Goal: 300 points Current: 0 points Low: 3 - 6 pts Med: 6 - 12 pts High: 9 - 18 ptsWe especially need testers (and bug fixers) for Windows and OSX. A lot of bugs have accumulated that only show on those platforms, so they are difficult to test.
If you are a Windows or OSX user, you too can be a big help, simply by browsing the bug tracker and see if you can reproduce the bug, and then add a comment on the bug one way or the other. No technical experience required. :-)
Thanks, Bryce
1576258 3 Color artifact in SVN gaussian blur nobody 1612814 3 Guides ans Gridlines vanish on corresp. alpha... nobody 1513218 6 Win32: ask if old menus.xml should be kept when... theadib ------------------------------------------------------------------------ Total 12
Bryce
1576258 3 Color artifact in SVN gaussian blur nobody 1612814 3 Guides ans Gridlines vanish on corresp. alpha... nobody 1513218 6 Win32: ask if old menus.xml should be kept when... theadib 1604391 3 MouseWheel NodeSelection > Undo = Floating Node nobody 1572450 3 Dialogs pop up off screen ------------------------------------------------------------------------ Total 18
On 12/6/06, MenTaLguY <mental@...3...> wrote:
The problem is that with current cairo, the result (while better in some important ways) is really broken in others, so on balance it's about as bad as the PDF export we shipped with 0.44 (which was so bad that it [almost?] wasn't worth shipping IMO).
I disagree with this assessment. Many people welcomed the transparency-enabled PDF export in 0.44 and were happy with it, despite its bugs. It was a huge improvement over 0.43. Rasterizing everything instead, as the Cairo currently does, is in my opinion an unacceptable regression.
participants (4)
-
Aaron Spike -
Bryce Harrington -
bulia byak -
MenTaLguY