The KDE Project today announced the immediate availability of KDE 3.5.3, a maintenance release for the latest generation of the most advanced and powerful free desktop for GNU/Linux and other UNIXes. Unusually for a maintenance release, new features were implemented due to the long release cycle of the eagerly-awaited KDE 4. Stability and speed were also improved, along with increasingly complete translations in 65 languages. The complete release announcement is available from the KDE website. Significant enhancements include an improved startup time, over 800 minor issues fixed thanks to code checking by Coverity and small new features in Akregator, KMail and KAlarm. Finally, new translations have been added for Vietnamese and Kazakh. Packages are available for Archlinux, Kubuntu, Fedora and SUSE Linux or compile with Konstruct.
Comments
Thanks!
I just love it how each release has a faster startup speed :)
I think it was a good idea letting some new features come in.
Forrest
In a few releases, the start up time will approach and cross 0. By kde 4.1, KDE will be on before you turn the computer on.
Jokes aside. Just updated to KDE 3.5.3 - WHOA! This thing IS starting much faster!
Its not that much faster, it just shows the desktop earlier - before restoring session. Like Gnome does, btw. Windows is even smarter: they show the desktop before it is actually usable...
No, in fact it even is really faster.
Are you the Lubos who did (some/all/most) of the startup improvements? If so then I guess you should know...
Seems faster to me, though I didn't get a stop watch out. Thanks btw :)
And speaking of which:
THANK YOU, Lubos. I've been keeping a remote eye on your work on startup speed and fontconfig improvement, and I think you deserve more praise and thanks than you get. :)
Actually I think a big portion of the fontconfig praise and thanks belongs to Dirk and Coolo :).
I hope you are joking. Windows behavior is a stupid lie. I prefer a splash screen suggesting you get a pillow, but when it is gone things work.
Windows is stupid for doing that--lots of people bang on the mouse waiting for it to become usable.
Hahaha! ROTFL! Thanks for making me awake this morning!
& of course thanks to the KDE team for this release ;-)
:D
Heise reports that lower saxony (germany) has set up 12000 KDE Desktops in their government.
http://www.heise.de/newsticker/meldung/73711
whoa, 12.000??? that's huge...
Germans using KDE...who woulda thought.
If we look on this map then we see that there are just a handful americans using the other desktop. Most of them are from europe as well. And soon we get them converted too.
http://live.gnome.org/GnomeWorldWide
Unfortunately, I'm pretty sure that the KMail "Quick Filing Feature" was removed a few days before 3.5.3 was tagged. I believe the plan is to have it back and operational for the 3.5.4 release.
Where is KMail fulltext search?
edit > "find message..." or press the "s" key
The index seems to be still missing, searching for keywords in the message body is very slow.
Cool. That gives packagers a few hours to prep for the "uber-stable" Kubuntu. ;)
(Just kidding, it's just amusing timing.)
To KDE developers! Good job guys!
Unbelievable !
KDE and Gnome actually launched an update on the very same day (even if Gnome still have not announced it), as Gnome 2.14.2 was available on the 31st too !!
Was this concerted ? Is this a first ?
Anyway, that's joy for me for both projects :))
KDE4 and windows vista will come out the same day. Was this concerted ?
Lol, who cares. It comes out when it's ready.
Does the VPL stuff in Quanta already do anything useful?
Or is this still in development?
This is no trolling.
I'm using Quanta for quite some time now for normal editing purposes and
it's working mostly OK.
I just tried out VPL for the first time and found it so extremely buggy that I couldnt do anything useful with it.
What I did:
1) I started Quanta
2) I pressed F9 for Split View (VPL+Code)
3) I could not click in the VPL part (Cursor did not blink or focus)
so I clicked in the Code between the body tags
4) Now I called the table assistant and created a 2x2 table
5) All the cells were very small. I could not click or enter text in a
cell in VPL. So I clicked between the TD tags in the code and entered text.
In some cells the text did appear in VPL, in some others not. At least not
at first. When I entered text in further cells, the missing cells suddely
appeared. Now I double-clicked a cell in VPL and the text inside was highlighted. Now I tried to enter a new text inside the cell directly in VPL. The current text disappeared but I couldnt enter new text.
6) I created a second tab to repeat my steps. Now Quanta crashed.
I dont think I did anything special. Arent this just the absolute basics of
a HTML editor? Did I do anything completely wrong? Or is VPL not at a
stage where it's usable yet? Or did I get the purpose of this VPL thing
completely wrong?
Why don't you report this issue at KDE's bugzilla: bugs.kde.org?
It will surely be more useful than complaning here.
Typical standard replies are not useful. Did you read my post properly at all?
I did not complain. I was just asking if this should be in a usable state at all.
There are so many bugs that I'm *really* wondering if this is just me or if anyone is using this already successfully. It's not useful at this point to file thousands of bugs IMO.
VPL is pretty much unmaintained in 3.5.x, what I do is just crash fixing. :-( But as I heard it is usable for some, so we did't remove it.
Ah! Thanks for answering. That explains alot. Well, I will try VPL again with Quanta 4. I'm still looking for sth. like Dreamweaver for KDE. Nvu isnt much usable either at this moment unfortunately because of important missing features and bugs. In the mean time I will stay with Dreamweaver under QEMU.
What is VPL?
a sort of wysiwyg mode for quanta (webpage builder/html editor and more).
> Over 800 minor issues fixed thanks to Coverity
Wow, an automatic bug-detector...
More a "potential bug detector". By analyzing source code, you can figure out whether all the variables are initialized, if pointers are non-NULL, etc. Coverity has advanced tools (like ESC/Java, but for C and C++) for doing the analysis and a fairly nice user interface for examining the results of the analysis. Basically, we got hundreds of reports like "Pointer p may be NULL in code p->foo() in file foo.cpp on line 121." So there's your bug (crash) detector; Dirk did most of the fixing, if I remember correctly.
Shouldn't this be done via debugging tools?
Test suites are of great importance to improve code quality.
Simple debugging is inadequate and inefficient. Coverity tools perform what is known as "static analysis" - they actually "reason" on the code and look for ways that small sections of code can be accessed (for example, if using a debugger only, you may never hit the 'magic' combination of tests and system state to try a portion of an if-test or rarely-called subroutine. Static analysis doesn't have that problem, it can walk through almost any possible code scenario without actually running the code. It can be time-consuming and tricky, doing all that backtracking and multi-path analysis.)
It's probably still best to perform static and dynamic analysis as well as conventional debugging - and all of it behind a strong lead-in based on quality development practices like Rational Unified Process, strong Systems Engineering (in the "Blanchard" style, not some ad-hoc method), code reviews, formalized build processes, etc.
If you're thinking "I can do it all with a debugger", you haven't been paying attention to the software engineering profession and developments over the past many years. There are better ways to get better code, and a debugger alone just doesn't qualify as "quality" anymore. Coverity (and no, I'm not an employee, own stock in them, benefit in any way from promoting them, etc) is a very good supporting member of the cast of development characters.
I just took a look at
http://scan.coverity.com/
and if i have to believe the statistics, it seems kde is bug hunting like mad!
Kde: 817 bugs resolved, 2 verified, 35 uninspected of unverified, lines of code: 4,062,450, bugs: 0.009 bugs / 1000 lines of code.
Gnome: 253 bugs resolved, 35 verified, 634 uninspected of unverified, lines of code: 2,227,424, bugs: 0.300 bugs / 1000 lines of code.
keep it going guys ;)
Keep up the good work, the bootup is ALOT quicker n1! :)
And for the new KDE it's time to update to Konqueror-Pro:
http://www.kde-apps.org/content/show.php?content=39419&vote=good&tan=190...
Note that parent has crafted an URL that tricks you into voting for whatever this so-called "Pro" thing is if you follow it.
Perhaps he is working for Diebold?
wooeee. That was my first real laugh today! I never trust electronic polling unless there is a print/paper check before submitting the result.
Keep calm, I guess the tan part will make it valid for only one vote.
http://www.kde-apps.org/content/show.php?content=39419
Is the link to actually SEE this Konqueror-Pro. whatever it is. And I don't force you to vote, either.
Sorry, just copy and pasted the URL.
Seems to be a problem of kde-apps.org that the voting is encoded in the URL.
And I don't think the voting actually happend because the TAN number schould be invalid.
It has no extra usability, it just adds redundant search fields and crams the exact same features into less space.
There is nothing redundant in this Konqueror profiles. I'm using it for more than a week now and it's better than the default profile and much better than the simple-browser profile (both included in KDE 3.5.x).
I've had a strange issue since I upgraded to 3.5.X versions of KDE : before, I could select a file in Konqueror, copy it then paste it in the same directory. Konqueror would prompt me at this point for a new name. Now, it just tells me that the file exists and that it can't copy a file on itself. I have to copy the file to another directory, change its name and move it back.
Is this a decision to change the behaviour or is this a bug that hopefully will have been fixed in 3.5.3?
Or maybe my setup is juste strangely broken.....
Best regards,
I've been having the same problem, but since I have a weak memory it made me doubt if that even worked before. I'm installing 3.5.3 right now, so I'll know if it's still there in a while, and will report it if it still is.
Seems fixed in 3.5.3. Cool!
why don't you just rename the file, in stead of trying to copy-paste the file over itself?