Beta-5 is out
This is an archived entry from the Movable Type 3.2 beta test.
Entries from the current beta release (v3.3) can be found here.
Beta-5 can now be found on the left sidebar of this page.
While you're waiting for that to download Read this (!!!).
And finally, below is a list of all of the issues we addressed in Beta-5. Enjoy!
- SpamLookup now supports blog-level configuration settings.
- Implemented better handling of DBI module requirement and reporting when it's unavailable.
- Fixed all contextual help links ([?]) currently in the application to point to the new user docs. Some pages to which the links point are still not published, but they will be very soon. (Case 13654)
- Now producing a useful error message if the directory containing a SQLite database isn't writeable by the web server. (Case 12716)
- Removed ALL MTLink tags from the default templates because they were causing problems with localized versions and in cases where peopple changed either the name or the output file of a template linked to with the tag. Good idea, bad implementation. Next time... (Many cases)
- Fixed a bug which cause mt-check.cgi to produce a blank screen or an error "Can't call method "translate" on an undefined value at mt-check.cgi" (Case 14097)
- Fixed a bug which caused an internal server error under Perl 5.004: "Can't locate object method 'rel2abs'" (Case 14344)
- Fixed a bug which could produce 404s under dynamic publishing if the Site URL of the blog is changed (Case 13675)
- Fixed a bug which cause MT to ignore the sort_order attribute of MTEntries under dynamic publishing (Case 14396)
- Fixed a "Use of uninitialized value" error which prevented run-periodic-tasks from successfully running (Case 14401)
- Fixed a bug on PostgreSQL that caused an infinite login process upon upgrade (Case 13814)
- Fixed a problem in which a missing config file would cause a 500 server error instead of producing a useful error. (Case 15228)
- Fixed a bug with dynamic publishing which caused the contents of the MTDateHeader/Footer container tags to be printed in each iteration of an MTEntries loop (Case 14396)
- Inserted a missing slash in the service.post link in default Atom feed (Case 13927)
- Now displating "No excerpt" for Trackbacks with no excerpt text so that a user can get to its individual editing page (Case 14435)
- Fixed a bug causing double-encoding of ampersands in category labels in power edit mode (Case 13912)
- Fixed a bug which prevented saving of multiple category assignments in some cases (Case 14578)
- Fixed a bug which caused an error for some plugins under mod_perl: "Missing configuration file" (Case 12856)
- IndexBasename is now stripped from the comment redirect script like it is from entry and archive links (Case 14768)
- Non-working formatting buttons now re-suppressed in Safari. Pester Apple about their Javascript implementation. (Case 14926)
- Fixed a bug which forced admins to restart their servers under mod_perl in order to take advantage of changed plugin configuration or enabled state. (Case 14781)
- Fixed a bug which prevented MT::Serialize from working under Perl 5.004.05 (Case 14881)
- Fixed a bug which required re-login before upgrade for users using Safari (Case 13572)
- Now sorting log items with the exact same timestamp in reverse-chron order by ID under all database except for BerkeleyDB, because it can't handle that sort of thing (Case 13521)
- Fixed a bug where particular users could, depending on their permissions, edit a certain set of their own permissions (Case 14805)
- Removed repetitive redundancy in phrasing on catgeory and index rebuilds screens (Case 13799)
Posted on August 18, 2005 5:33 PM in Announcements


Comments
What specifically is wrong with Apple's JavaScript implementation? Did you file a bug? It's open source, so you can even go fix the bug yourself!
Posted by: Michael Ditto
|
August 18, 2005 6:35 PM
Ditto, Mike! Can we get more details on what exactly is wrong with Safari's JavaScript implementation? It'd be cool to get this solved once and for all....
Posted by: Jeremy
|
August 18, 2005 7:05 PM
Mike, we have our own bugs to deal with. What are you doing right now? ;-)
From what I understand there is no way in Javascript in Safari to place the cursor at a particular point within a textarea, but I could be wrong.
My engineers (who are some of the best at Javascript) say it's not possible. I trust them.
Posted by: Jay Allen
|
August 18, 2005 7:05 PM
Thank you! I was having all kinds of bugginess after installing b3 and then b4. My quicklinks quit working, my posts would save but give a 500 error and no trackback, my blogroll plugin was giving me a template error.
All work now with this install.
Todd
PS:
Can anyone point me to a resource that helps get my trackbacks 'unblocked' from certain blogs? I've not been a troll to anyone and have tracked few sites (such as this blog) and gotten no permission errors.
Posted by: todd
|
August 19, 2005 8:07 AM
This explains so much with MTlink. Here I thought it was me...
Posted by: yasmín
|
August 19, 2005 9:13 AM
It's good! For some reason, I couldn't get the cron jobs working on b4. After the upgrade, it was back to normal.
Posted by: Philip
|
August 19, 2005 9:19 AM
I know that bug. I logged it quite a while ago. It is an annoying one.
I might see what I can do to fix it, but whether it actually makes it in to JavaScriptCore is a different question.
Posted by: Michael Ditto
|
August 19, 2005 9:53 AM
Jay, there is no spec'd way of placing the cursor in a textarea, period. There are methods in Mozilla and MSIE to do it, but they are both nonstandard extensions and are implemented differently.
Adding the functionality will require convincing Apple to add a nonstandard extension. I'm on it though.
Posted by: Michael Ditto
|
August 19, 2005 3:54 PM
Yesterday, I upgraded to Beta-5. The upgrade process was great. Everything is great. Thanks!
Posted by: Thomas
|
August 19, 2005 10:56 PM
..is this looking like being the last beta release (including nightlies)?
Posted by: demonsurfer
|
August 19, 2005 11:27 PM
I already said this on my blog, but just to let you know, my upgrade took all of 2 seconds after uploading the files. In fact, I'd be willing to wager it was less then 2 seconds.
Oh, and no bugs that I could come across.
I am going to do a clean out of MT from my server this weekend and do a full install, then see what I get for bugs.
Posted by: Rook
|
August 20, 2005 4:41 AM
I don't know if this is the right place to ask this question, slap me gentle if it's not, okay?
I upgraded to beta 5 this morning (flawless, as usual) and now I want to finish customizing the last few temmplates before I consider the heavy lifting done on my site. Yeah, there will be the release version but I don't expect to have major work to do.
My question is regarding the search template. I know it's at search_templates/default.tmpl. I understand that it had to be edited separately because it's a single template that works across all blogs so it couldn't be in the blog-specific "templates" area. But now with 3.2, that's no longer the case. Why didn't you bring it into the Movable Type interface now that you have system-wide settings? When I change my blog design, I have to remember to edit that template. Plus, I prefer linking my templates to files and that's not possible with the search template. Just a thought.
Posted by: Judi Sohn
|
August 20, 2005 11:00 AM
Another smooth upgrade!
Posted by: Dan Wolfgang
|
August 20, 2005 12:13 PM
Umm, because we had enough to do? Seriously, the search functionality needs a heck of a lot more love than just bringing the templates into the interface. And besides, you notice that there is no system-wide templates? When they come, they'll be done right.
We always have to balance the desire to improve something a little bit and doing it right. We did a lot in this release. To answer the question "Why didn't you do X", the answer is always going to be a mix of time, resources and priorities.
Posted by: Jay Allen
|
August 20, 2005 1:13 PM
You did the right thing by issuing Beta 5. If you have to, issue Beta 6 and 7 and...
Admittedly, this is a roman a clef, as I'm in the same boat myself, managing a significant migration and system change for a big website. We had to delay launch... and delay it again... and again... and again. It's hard when you want to "be there" but you know you aren't ready to launch. It's great to see you opening up the beta process this way, and I'd participate myself if I could stomach any more beta testing in my life right now. :-)
Good luck! Break a leg!
Posted by: kgs
|
August 20, 2005 7:24 PM
Previous betas were running pretty slowly on my test server with a BerkeleyDB for some reason, but beta-5 is running incredibly fast. I don't know if it's something you've changed or if my host's suddenly upgraded the server, but I'm very impressed!
Posted by: quack
|
August 21, 2005 11:31 AM
First of all - very smooth so far. I've tested since beta 3 and have moved the installation from a cPanel server to a Windows IIS 6 install without any issues whatsoever right up to this bet. Very clean, very smooth. Fantastic!
However, I notice the lack of these new design templates that's been mentioned. Can we try these out before the release? Are they presented in 3.2 as they would be in TypePad (i.e. graphical interface, pick 'n' choose) or is it simply a case of cutting and pasting from the Moveable Type web site?
Posted by: movielad
|
August 21, 2005 12:03 PM
We will release them with 3.2. The default templates are compatible with all of the new styles. Hang tight just for a couple days and you'll see what we have planned.
Posted by: Jay Allen
|
August 21, 2005 12:13 PM
From Apple:
So it should make it into the next release, hopefully that will be soon!
I tested the latest build of WebKit (Safari's engine) and it is in fact in there. So if you guys could disable it only for the current version of Safari and below, that will enable the functionality when it appears in the next Safari release.
Posted by: Michael Ditto
|
August 22, 2005 12:40 PM