Hi all,
After chatting with Ishmal, Mental, Bulia, and others, it sounds like we ought to shoot for cutting a near-term release, in the mid- to late-January timeframe.
The reasoning for this is that we already have some noteworthy new features that would be worth getting out to users, and because the Gtkmm work, while progressing along well, won't be finished that quickly.
If no-one has any serious objections to this proposal, we would go back into a feature freeze within the first week or so of January, focus on bug fixing for a couple weeks, and try to push out the 0.41 release by the end of the month. (As a side PR benefit, there are a number of conferences and symposia in the Feb/Mar period, so this would give fresh bits for folks to play with.)
In parallel to the above release effort, the focus on the Gtkmm work will continue (and hopefully intensify), with the objective of merging immediately after the 0.41 release. The code won't be finished at that point, and won't be used for the primary UI (you'll have to use a special flag to activate the Gtkmm), but we'll finally have the code in the main codebase proper, which will hopefully enable the remainder of the work to get done swiftly.
Bryce
On Sun, 26 Dec 2004 11:29:13 -0800 (PST), Bryce Harrington <bryce@...260...> wrote:
Hi all,
After chatting with Ishmal, Mental, Bulia, and others, it sounds like we ought to shoot for cutting a near-term release, in the mid- to late-January timeframe.
It would be very nice if we could make scripting the main feature of 0.41. Thanks to Bob, this basically works already, though we need more sample scripts, more API, more documentation. Also Ben Crowell's Perl scripts and XSLT scripting are very nice features that should be more accessible, if possible, via the Scripts dialog or otherwise. And Ted has recently added the Effects menu, though I don't know how close it is to being useful.
With the close release in mind, developers please take time to fill in the corresponding placeholders at http://inkscape.org/cgi-bin/wiki.pl?ReleaseNotes. Describe what works already and leave placeholders for what you think you can finish before 0.41.
As for flowtext, it is now basically editable on canvas by the text tool, but is quite crashy and buggy, so I don't know if it will get any user-visible advancement in the new version. There's a lot of work to be done there.
From my perspective, the main reason for releasing 0.41 early is the
bugfixes. There are a few very important bugfixes in CVS already compared to 0.40, but we need to squash at least one of the killer Windows bugs to make the release worth its bandwidth. So:
http://troi.hous.es3.titan.com/~rjamison/inkscape/builds/Inkscape0412221011-...
This build uses a version 6.4 of boehm that may fix the Windows freeze bug:
http://sourceforge.net/tracker/index.php?func=detail&aid=1073459&gro...
So Windows folks who experienced this bug (Scislac, at least), please test this build and report your results here. This is important.
Another serious problem we're having is performance. Mental, you mentioned you had an idea of how to speed up the renderer with a relatively small change. Can you please try this now? It would be very cool to make 0.41 at least somewhat faster than 0.40, because this problem is becoming increasingly bothersome as people start to create more and more complex art with Inkscape.
And finally, as always, everyone please test CVS extensively to make sure we haven't at least broken something which used to work before.
On Sun, 26 Dec 2004, bulia byak wrote:
After chatting with Ishmal, Mental, Bulia, and others, it sounds like we ought to shoot for cutting a near-term release, in the mid- to late-January timeframe.
It would be very nice if we could make scripting the main feature of 0.41. Thanks to Bob, this basically works already, though we need more sample scripts, more API, more documentation. Also Ben Crowell's Perl scripts and XSLT scripting are very nice features that should be more accessible, if possible, via the Scripts dialog or otherwise. And Ted has recently added the Effects menu, though I don't know how close it is to being useful.
What Bob and I had discussed a few days back was that 0.41 would include scripting as an 'optional dependency', in order to get it out in the wild for testing by bleeding edgers, but not make it a required dependency for regular users, "just in case". Like you say, there's still a lot to be done to get scripting fully fleshed out, but this would allow us to get some basic packaging/testing sorted out.
Are there other major features besides scripting? If so, perhaps for this release there could be two packages? This would allow us to provide scripting for the bleeding edgers, and a plain (but enhanced stability) version for those who aren't.
Bryce
On Sun, 2004-12-26 at 15:31, bulia byak wrote:
As for flowtext, it is now basically editable on canvas by the text tool, but is quite crashy and buggy, so I don't know if it will get any user-visible advancement in the new version. There's a lot of work to be done there.
One important thing we need to do with flowtext (and really any SVG 1.2 features) is to start marking our documents with the appropriate SVG version.
We already have stuff in place for associating sodipodi, inkscape, and SVG versions with each SPObject. I think right now everything gets an SVG version of 1.0; we'll need to make sure that flowtext and other 1.2-specific features get an SVG version of 1.2. That shouldn't prove too difficult.
The overall SVG version (the one recorded in the root <svg> element) should be the maximum version of any of its children.
-mental
On Sun, 2004-12-26 at 16:31 -0400, bulia byak wrote:
It would be very nice if we could make scripting the main feature of 0.41. Thanks to Bob, this basically works already, though we need more sample scripts, more API, more documentation. Also Ben Crowell's Perl scripts and XSLT scripting are very nice features that should be more accessible, if possible, via the Scripts dialog or otherwise. And Ted has recently added the Effects menu, though I don't know how close it is to being useful.
The effect menu is usable, but I wouldn't say releasable. I wouldn't say that it will be ready for 0.41, but 0.42 is definitely a reality. I've purposely left the menu '#if 0'd out because I thought we might be doing another release before it is ready.
If people want to play with it, they are more than welcome, I just don't want it ending up in screenshots or anything like that. All you need to do is turn the #if 0 into a #if 1 in interface.cpp. If you search, it is the first one you come to. I've implemented an example called 'grid' which will generate a grid over the selected objects. It still needs more work, but can be fun to play with. (also, it can generate complex paths that slow Inkscape down for performance testing :) )
--Ted
On Sun, 2004-12-26 at 16:31 -0400, bulia byak wrote:
And finally, as always, everyone please test CVS extensively to make sure we haven't at least broken something which used to work before.
Has anyone had any problems with the timed unloading of extensions? I know it hasn't been in a long time, but I was just curious.
--Ted
participants (4)
-
Bryce Harrington -
bulia byak -
MenTaLguY -
Ted Gould