2.5.1.4 --> 2.5.1.5

FYI, while making sure that the embedded sample still worked (it didn't even compile) I discovered some bugs in the new wx module initialization sequence that caused a fatal crash in the embedded sample. So that is the main change in 2.5.1.5 from 2.5.1.4. I also changed a couple of the samples in the demo to not behave badly on OSX.

···

--
Robin Dunn
Software Craftsman
http://wxPython.org Java give you jitters? Relax with wxPython!

What's done is done, but I still think that us guinea pigs should be testing the exact same binaries that the rest of the community gets later. It is a pain to go through release candidates every few days, but if we are trying out the exact same files that the rest of the users get then we should avoid things like having to release a 2.5.1.6 immediately due to a stupid mistake (fingers crossed 2.5.1.5 is okay). That's why we started this list in the first place. OTOH, I know you have this thing for doing releases on Friday, so I guess I understand... :slight_smile:

ka

···

On Apr 2, 2004, at 9:23 AM, Robin Dunn wrote:

FYI, while making sure that the embedded sample still worked (it didn't even compile) I discovered some bugs in the new wx module initialization sequence that caused a fatal crash in the embedded sample. So that is the main change in 2.5.1.5 from 2.5.1.4. I also changed a couple of the samples in the demo to not behave badly on OSX.

Robin Dunn wrote:

FYI, while making sure that the embedded sample still worked (it didn't even compile) I discovered some bugs in the new wx module initialization sequence that caused a fatal crash in the embedded sample. So that is the main change in 2.5.1.5 from 2.5.1.4. I also changed a couple of the samples in the demo to not behave badly on OSX.

Wow! that was fast. I havn't had a chance to do mauch testing, but 2.5.1.4 broke all my mouse dvent handling in FloatCanvas, at lest on OS-X. Having not tested it carefully, it could be my error. I'll report when I get a chance to look closer. By the way, I have made major additions and improvements to FloatCanvas. I didn't feel that is was well enough tested and packaged to send to yet. Oh well, maybe for the next release.

While we're on the topic, what is the motivation for releasing 2.5.* now? My understandiong is that wxWidgets 2.5.* is an unstable release, so I wonder whether to recommend folks start using wxPython 2.5.1.*. I want to as I like a lot of the changes, adn it does work better on OS-X in particular. However, I understand there's been a lot of great work done for OS-X in 2.5.* that you haven't included yet. Should we wait for that? Or will 2.6.* be more or less source-compatible with what you have just released anyway?

-Chris

···

--
Christopher Barker, Ph.D.
Oceanographer
                                         
NOAA/OR&R/HAZMAT (206) 526-6959 voice
7600 Sand Point Way NE (206) 526-6329 fax
Seattle, WA 98115 (206) 526-6317 main reception

Chris.Barker@noaa.gov

Kevin Altis wrote:

FYI, while making sure that the embedded sample still worked (it didn't even compile) I discovered some bugs in the new wx module initialization sequence that caused a fatal crash in the embedded sample. So that is the main change in 2.5.1.5 from 2.5.1.4. I also changed a couple of the samples in the demo to not behave badly on OSX.

What's done is done, but I still think that us guinea pigs should be testing the exact same binaries that the rest of the community gets later. It is a pain to go through release candidates every few days, but if we are trying out the exact same files that the rest of the users get then we should avoid things like having to release a 2.5.1.6 immediately due to a stupid mistake (fingers crossed 2.5.1.5 is okay). That's why we started this list in the first place. OTOH, I know you have this thing for doing releases on Friday, so I guess I understand... :slight_smile:

Normally I would, but since the only change (besides the demo tweaks and some BUILD/INSTALL docs updates) was a three line change that would only take effect for embedded situations (which I don't think anybody here is doing) I didn't think it was worth the time/effort. For non-embedded situations the change was such that it would either have no effect if done correctly, or would be immediately fatal if not. Since it didn't crash I figured it was safe.

···

On Apr 2, 2004, at 9:23 AM, Robin Dunn wrote:

--
Robin Dunn
Software Craftsman
http://wxPython.org Java give you jitters? Relax with wxPython!

Chris Barker wrote:

Robin Dunn wrote:

FYI, while making sure that the embedded sample still worked (it didn't even compile) I discovered some bugs in the new wx module initialization sequence that caused a fatal crash in the embedded sample. So that is the main change in 2.5.1.5 from 2.5.1.4. I also changed a couple of the samples in the demo to not behave badly on OSX.

Wow! that was fast.

Eight preview releases and five release candidates over 6 months is fast? :wink:

I havn't had a chance to do mauch testing, but 2.5.1.4 broke all my mouse dvent handling in FloatCanvas,

Did they work in 2.5.1.3?

at lest on OS-X. Having not tested it carefully, it could be my error. I'll report when I get a chance to look closer. By the way, I have made major additions and improvements to FloatCanvas. I didn't feel that is was well enough tested and packaged to send to yet. Oh well, maybe for the next release.

While we're on the topic, what is the motivation for releasing 2.5.* now?

To allow people to start using the new features.

My understandiong is that wxWidgets 2.5.* is an unstable release,

Where "unstable" only means that the API is allowed to change and the C++ classes don't have to be binary compatible between releases, not that it is any more buggy than a "stable" release. In fact, the unstable releases often have less bugs because sometimes they are easier to fix when not constained by stableness.

so I wonder whether to recommend folks start using wxPython 2.5.1.*. I want to as I like a lot of the changes, adn it does work better on OS-X in particular. However, I understand there's been a lot of great work done for OS-X in 2.5.* that you haven't included yet. Should we wait for that? Or will 2.6.* be more or less source-compatible with what you have just released anyway?

There will be more 2.5.x releases before 2.6. Probably even more 2.5.1.x releases. That said, the further 2.5.x and 2.6 releases will be more similar API-wise with 2.5.1.5 than with 2.4.2.4.

Once the current flurry of changes in CVS settles down then I'll start the release cycle again. (Not just the classic/carbon split and the new control compositing changes for wxMac, but also changes dealing with window attribute (font, colours, etc.) defaults and inheritance, and a few other things too.)

···

--
Robin Dunn
Software Craftsman
http://wxPython.org Java give you jitters? Relax with wxPython!