KDE 3.3 About to Finish: Public Release Candidate 2

We have already received a lot of feedback to our KDE 3.3 Betas, so we can release a public Release Candidate. We want to have this tested as soon as possible, so we didn't wait for vendor binaries. So please test the sources if you have experience in this (or use Konstruct).
You can download the Release Candidate from download.kde.org.

Dot Categories: 

Comments

by testerus (not verified)

e.g. http://bugs.kde.org/show_bug.cgi?id=66411 makes sftp unusable.
And there are many more bugs that make kio slaves unusable in practice, see http://tinyurl.com/63l28

Unusable, yes. And please fix these bugs too:

trolls usually post anonymous...

Good that you gave your full name and email address. You are not familiar with irony, or?

by Nicolas Goutte (not verified)

Well for things to be fixed you need a developer with time, knowledge and a SSH access. It seems that these conditions are currently not met for FISH and sftp.

Have a nice day!

> http://bugs.kde.org/show_bug.cgi?id=66411 makes sftp unusable.
It is prefectly usable for me since I do not have to worry about encoding issues. If people that use the software and experience the same issues as you do not step up and offer to fix such problems, it will never get fixed. I personally started maintaining the sftp io-slave on a part-time basis because it was completely neglected and abondened by the original author.

> And there are many more bugs that make kio slaves unusable in practice, see
> http://tinyurl.com/63l28

Please stop making nonsense argument like this. Just because there are bugs in some io-slaves, which BTW might or might not be still valid, does not mean that they are "unusable in practise". Did you even bother to go through some of these bugs and check which ones are valid and which are not ?

by Norman Urs Baier (not verified)

Why would someone want to use umlauts in file names?

Yes, what a stupid thing to do. Especially since the whole filename encoding issue is just so perfect under Linux - who needs to know which character encoding was used for a filename? And didn't you know that if your first language employs both ASCII and non-ASCII characters, it's just so cool to drop in transliterations and bizarre combinations of ASCII characters instead of using such "archaic" characters as ü and ø; it just shows how "elite" you really are.

Reality check! Ignoring the 0.0001% who consider themselves "elite" in the way described above, most people whose first language uses non-ASCII characters are likely to find the lack of proper support for their language *unfathomable* and *inexcusable*.

And as for the Linux filesystem issues, I imagine that it's unlikely that they'll be resolved for real-world users any time soon, thanks to a combination of buck passing, indecision, and the lead time on distros taking any decent solution up after people finally get their 'rses in gear.

by Jeld The Dark Elf (not verified)

I do not quite understand your argument. I am a russian speaker, and I have files named in both russian and english letters. Ever since I have given up on the encoding war and switched to utf-8 I don't have any problems with file naming.

Yes, but switching to UTF-8 is a big problem in itself. Moreover, last time I checked, there wasn't any widespread mechanism in place in Linux (which is a pretty major platform for KDE, after all) to force users to actually use UTF-8 when naming files - to the kernel and most filesystems, filenames are just byte strings and users can choose different locales and blast in character combinations which are illegal in UTF-8 but which look alright in each user's own chosen character encoding.

It's a nasty situation, but significant projects like KDE should be putting pressure on the Linux people (amongst others) to fix the insanity.

by Oliver Stieber (not verified)

This seems to be a big problem in KDE or maybe mimetypes in general (well big as soon as you start to dig).

If you look at the X clipboard most things will put text as a mime type of
text/plain;encoding, where encoding is something like UTF8 etc...

Although thease 'extended' mime types are used quite frequently, try telling KDE[QT] your document is type text/plain;UTF8, not just text/plain.

by Olle Håstad (not verified)

As this question sometimes pops up I’ll try to clarify this problem.

1) Normal users often want to use file names that describes the content of a file in their native language.

2) Calling all characters that carry dots or circles “umlauts” somewhat transforms the question to a esthetical one, which it in no way is. I will use Swedish as the example:

The Swedish alphabet consists of 28 characters; in addition to the once shared with English we have 3 additional vowels: å, ä and ö. These are as related to a or o as i are to u in English. Å is written as an A with a circle on top while U is written as to I:s connected by an arc at the bottom. I do not know how to further emphasise this so I conclude by again stating that they are separate letters pronounced totally different from a and o.

If these letters can’t be used the language loses 3 out of 9 vowels. The effect is the same as if you would remove two vowels from written English (2/5). For a demonstration try replacing all u:s and o:s with a:s in any English text. Some words will change their meaning while others will just turn into gibberish.

I have a colleague whose last name is “Ås” meaning ridge as in “mountain ridge”. In this situation his name would be transcribed as As meaning “carcass”, in Swedish used as a derogatory term. This is funny the first 3 times but I guess he appreciates systems where he is able to use his real name.

Normal users expect to be able to use any computer system in their native language including naming files, especially if the user interface is translated, which is exemplary done in the case of Swedish.

Best regards
/Olle.

Well, your explanation made sense to me, probably because I've been acquainted with the facts you provide for some time. Indeed, I imagine that they are well and willingly understood by everyone except the Commodore 64 brigade.

by Norman Urs Baier (not verified)

Pågen, Åkerman and so forth show very effectively how important the å is in names, they just write a for their international products. This is the strategie I know also from my girlfriend. Theoretically it would be possible to write ao and oe instead of å and ö. Which probably is also the way these letters were constructed some years ago, any calligrapher around?
Personnally, I insist on being able to use special caracters (for simplicitys sake called umlauts before) in normal text, but we are talking about filenames here. I have big difficulties to show any comprehension to people wanting spaces and special caracters in their filenames. This searching for trouble.

by Olle Håstad (not verified)

>Pågen, Åkerman and so forth show very effectively how important the å is in names, they just write a for their international products.

It is possible to just use A instead of Å but this transforms a name with a meaning as in your examples to just names (without meaning). Another example of this is the Swedish company “Skånska cementgjuteriet” (meaning Scanian cement foundry) changing their name to the internationally more useful name “Skanska” (without any meaning). But this in not very useful when the most important thing you want to preserve is the meaning of the word, as in a filename. Otherwise all files could be known only by the inode number.

>This is the strategie I know also from my girlfriend. Theoretically it would be possible to write ao and oe instead of å and ö. Which probably is also the way these letters were constructed some years ago, any calligrapher around?

If you by “some years” refer to the 14th century you are correct. :)
It is possible to use aa, ae and oe to transcribe Swedish text. This is however neither used nor easily understood by normal people. Remember that it has to be practical for the average user.

>Personnally, I insist on being able to use special caracters (for simplicitys sake called umlauts before) in normal text, but we are talking about filenames here. I have big difficulties to show any comprehension to people wanting spaces and special caracters in their filenames. This searching for trouble.

I would not call it special characters, as they are not more special than, say g is, to a Swede. When talking about special characters you adopt a nationality. To a Swede W could be considered a special character as it is only used in foreign words and names.

When saving a document in MS word, the save dialogue box will default to a name consisting of the first words from the text. This suggested name often describes the document fairly well and most users use this as the filename with minor changes (and, yes I know this is to ask for trouble when moving the file between different OS:s, but not for that user on his own system). As a grad student I used to be responsible for the computers in my research department, mainly consisting of windows computers (95 to 2000) with some Macs (Os 8 to X) and linux servers. The working language for everybody was English. Most people however preferred to give files names that made the files easily found and identified. A longer name consisting of several words in the users native language fulfils those demands. Over the years the interchange of files went progressively smoother, not as a result of me constantly preaching the gospel of “English file-names without spaces”, but as a result of improved interoperability of the computer systems.

The point of this is:
It is much easier to make the computers adapt a naming scheme that is comprehensible for normal people than to constantly re-educate people to the ever changing capabilities of computers. I know of the problems in doing so, but it has to be done sooner or later.

I often use fish or sftp to access DNA-sequence files saved from a dedicated Mac Os9 sequencing computer on a Win2000-server through an smb-mount on my linux-server to my kde laptop, which I work on through a cygwin X-server on my windows XP workstation. Well, it is not optimal but this is the way it is in many universities and the KDE end of it usually works like a dream.

It is however very difficult for me to make the rest of the department adjust to my “safer and superior” file-naming scheme, when their intuitive and simple method of just describing the file (meaning, not name) in Swedish works perfectly for them on their computers.

Best regards
/Olle.

by Norman Urs Baier (not verified)

Your reply was diplomatically well written in a way that I just agree, especially as we are both considering the KIO Slave behaviour a bug.
Unfortunately we did not get any further what the solution concerns. How much coding effort would it take to eliminate that bug?

by Waldo Bastian (not verified)

bugs.kde.org has approx. 7000 open bug reports, we can't fix those before KDE 3.3

We do like to fix anything that got broken by KDE 3.3 though. If there are things that used to work in KDE 3.2.x but that no longer works in the KDE 3.3 beta / RC, please post it here.

by James (not verified)

what is the reason of this rush release?

by Amadeu (not verified)

Why do you think the release is being rushed?

by More-OSX_KDE_In... (not verified)

It's SuSE vs. Ximian at the Novell Coral!

by Waldo Bastian (not verified)

If we can release a KDE 3.3.0 today that is better than KDE 3.2.3, why should we wait with it till tomorrow?

by James Richard Tyrer (not verified)

I think that it depends -- depends on how long till the next release.

The GIMP released 2.0.0 which was rather buggy but they have already released 5 (count them FIVE) bug fix releases. This is the release early and often approach. Perhaps 2.0.0 should have been called a release candidate.

OTOH, if we intend to be waiting like 6 months till 3.3.1 then we should try to get more of the bugs fixed before it is released. This approach is more like commercial software.

It also depends on how stable our customers expect an X.y.0 release to be. If they expect stable, then we could still release 3.3.0 today, but we should call it an RC.

--
JRT

by Anonymous (not verified)

KDE 3.3.1 will be rather released 6 weeks, not months, after KDE 3.3.

by James Richard Tyrer (not verified)

We can only hope that in 6 weeks that at least the serious problem in the window manager can at least be turned off and the the desktop icon spacing problem is fixed.

--
JRT

you know i'm getting pretty sick of your pissing and moaning JRT. yes your finding bugs and your a help to the KDE QA team but god damn you can be an arrogant sod. Perhaps instead of proclaiming tiny bugs in the mailing lists as "SHOWSTOPPERS" you should get a clue, chill out and start being more polite to the developers that make linux useable.

by James Richard Tyrer (not verified)

I don't piss and I don't moan. I try to be direct and tell the truth as I see it. For some reason, developers take this as a personal insult.

In my not at all humble opinion, it is the developers that need to get a clue. They need to somehow see things from the users perspective. I am sorry that I don't have humble opinions -- that is the personality that I was born with and I am probably too old to change it.

As I see it, a significant number of developers have been from somewhat to very rude to me and this has resulted in a somewhat negative attitude on my part. Perhaps this is because I forgot to kowtow to the developers. I would be happy to treat the developers as equals, but I get the impression that they are not willing to concede to such a large upgrade to my status.

I realize that there is a large culture gap between an engineer and the hackers, but I am willing to try to meet them half way. Perhaps if you could explain what you expect in the way of being "polite" it would help. I was trained as an engineer. I learned that culture and part of that is being direct and objective. If others see this as being impolite then it is actually their problem.

I think that the best example is that I took the time to explain to someone exactly what was wrong with his application and what needed to be changed to fix it and he took it as a personal insult. This is the culture gap -- another engineer would have thanked me for the help (after a while he figured out that it *was* help).

Anyway, since the 3.1.0 release we have had various problems with some of the releases ranging from things that wouldn't compile to serious regressions. To a professional engineer, this is totally unacceptable. I will say it straight out to all the developers: this is NOT the way to do things. They really need to get a clue and figure out how to avoid this in the future. If we expect for governments and major corporations to adopt out software for their desktops, we MUST have better quality control. Why is it so hard for the developers to understand this -- it is obvious to me.

There is no excuse for the problem with the vertical icon spacing and the cute little comment in the code is indicative of the problem with the attitude of some of the developers. This should have been a show stopper because it is (1) a serious bug and (2) a regression. If you think that it is a tiny bug, it is your attitude that needs adjusting. Many users have many icons on their desktop. When they upgrade and open their desktop for the first time, their icons will be sprayed all over the place. There is nothing that upsets users more than this happening.

As I said elsewhere, the broken "feature" in KWin that you can't even turn off is more than just a major bug. It indicates that the development model needs some work -- this should have never been committed to the main branch in CVS before it was designed and tested.

In the case of the 3.3.0 release there is a simple solution to the problem, call it 3.3.0RC1 and then in 6 weeks fix all of the serious bugs and regressions and release 3.3.0. But that is just a quick fix, we need to develop better methods. If we don't, the new releases will get worse rather than better -- like that other wellknown desktop software did after the 3.10 release.

I am here to help. If, as it appears, the developers don't want help, then I am willing to just give up on KDE if that is what they really want.

--
JRT

by Ben Damm (not verified)

I am not a KDE developer, but I can understand why software developers find you irritating. The programmers do deserve your respect, they are the ones putting their time and effort into development; it is no small amount of work.

If you want bugs fixed, you need to do more than find them; you must convince someone with the power to fix it that it should be fixed. Getting rude is not the answer. Proclaiming yourself to be a peer to the programmers in this project, when the amount of work you put in is a fraction of the work that they put in, is not the way to do it. How about doing something useful such as establishing priorities?

by Nicolas Goutte (not verified)

It was decided to release KDE 3.3 before aKademy.

So perhaps the decision was good, perhaps it was bad but it is not the time anymore to discuss about it (that was at last in May 2004 to do.)

There is still KDE 3.3.1 (and later) and this is not the end of the world.

Have a nice day!

by Anonymous (not verified)

> It was decided to release KDE 3.3 before aKademy.

If applicable, not just for the sake of it. But at the moment there doesn't seem to be a reason not to reach this goal.

by Anonymous (not verified)

> There is still KDE 3.3.1 (and later) and this is not the end of the world.

Exactly, there is always an "But you can't release with this bug!" outcry before every major release. But what shall the developers do if no patch to fix it is available (after days or even weeks). It has to be released eventually.

by Derek Kite (not verified)

You misunderstand the release process. We are not looking for perfect software, otherwise there wouldn't ever be a release. 3.3 has fixes and improvements function wise that you and others may enjoy or need. So development is stopped for a bit, or slowed down, everything is stabilized to a reasonable degree, and it is released.

Now you and others can use it, find flaws or shortcomings, fix them and forward patches.

Then 6 or so months from now another release will contain your bug fixes.

Derek

by Carlo (not verified)

>bugs.kde.org has approx. 7000 open bug reports, we can't fix those before KDE 3.3
Right. But the broken kio stuff is rather annoying. The amount of bugs and the average response time doesn't really encourage to report new (in fact old) bugs. I don't want to insult anyone - of course - but I'm holding bug reports back, because I see not much sense in it, to report them... Btw.: We had a Gentoo user, who reported that holding down when running konqueror is a fast way to start lots of konsoles. Did he report the bug/is this fixed in 3.3?

by Waldo Bastian (not verified)

<20 konsole windows later> Nope, it still does that.

by Jaco (not verified)

Only 20? Lucky. I had a couple more when I tested it over here.

Well, it is to be expected since on my quite busy box it took a while to appear, so I tried again :)

Jaco

by Hans-Peter Schadler (not verified)

40 konsoles later...
Oh damn man. havent known that. bad thing *g*

by Allan S. (not verified)

I thought it didn't work here, so I tried a few times.. Got 208 konsoles a few minutes later. Arghh!
Die die die.

by Waldo Bastian (not verified)

The problem is that keyboard shortcuts are subject to the keyboard repeat rate.
Should not be too hard to fix, but that might break applications that depend on such behavior, so it is not something we are going to fix for KDE 3.3 any more.

The problem also exists in KDE 3.2.x, it is not new in KDE 3.3

by Dik Takken (not verified)

Wow, I had to reset my machine after pressing F4 for three seconds. It became completely unresponsive because of the heavy swapping. After 15 minutes, I pressed the reset button...

by MaxAuthority (not verified)

hmm, couldn't reproduce it here on my gentoo box.

opens just one xterm, and this gets the focus immediately, so the next F4 key events go to xterm and does not open another one.

but maybe that's just when using xterm's or Ion as the window manager.

by Dik Takken (not verified)

xterm loads faster than the keyboard repeat rate. Konsole takes a few seconds.

by Peter (not verified)

And what's the problem with that?

If I press ctrl+alt+backspace on someone else's machine, my X session resets.

If you don't like the default behaviour, change it. (in my opinion, it's better to have default keybindings than to have to assign them by yourself)

by Anonymous (not verified)

Why don't you look up it yourself at http://bugs.kde.org if he reported it?

by Carlo (not verified)

Did you ever search for "F4"? Even in conjunction with another keyword it is hard to estimate, if your result set doesn't include the bug or if it does not exist; And the bug is an ugly one, but not so important that it's worth a dupe. Also we have better things to do, than to forward _all_ bugs Gentoo users find.

by Anonymous (not verified)

> Did you ever search for "F4"?

I wouldn't do that, "Konqueror open terminal" seems to more appropriate.

> Also we have better things to do, than to forward _all_ bugs Gentoo users find.

And the KDE developers have better things to do than to scan the mailing lists of all distributions for descriptions of possible KDE bugs. If you don't care to report it then don't moan that it stays unfixed!

by Carlo (not verified)

>If you don't care to report it then don't moan that it stays unfixed!
I don't moan. And definitely not about unreported bugs. I'm aware that fixing bugs is not (always) simple and that some bugs/wishes need major changes to the relevant source code (and time). It's just that the perception from the outside is a steadily growing bug number and more features (which isn't bad), but a lot of remaining odd bugs from one version to the next.

by Derek Kite (not verified)

>perception from the outside...

And the perception from the inside is a continuous flow of bugs and desired features. KDE is noteworthy for accepting contributions from newcomers. If you are interested in a stable KDE, patches are welcome.

Derek

by Sean Lynch (not verified)

http://bugs.kde.org/show_bug.cgi?id=60625

I've had this bug reported for about a year now. The same "repeated keyboard shortcuts" happens with just about all shortcuts, including control+w to close a tab in konqueror.

Sometimes you accidently hold it down for a half second longer than you meaned to, closing all your opened tabs, or at least the good majority.

by TDH (not verified)

> The amount of bugs and the average response time doesn't really encourage to
> report new (in fact old) bugs

While reporting old bugs doesn't make sense, I've had rather good luck with reporting new ones. When I've kept pursuing them until the problem can be reproduced, I've never had to wait more than a release cycle for a fix. The process is a bit like raising an infant: if you expect him/her to start talking tomorrow, you'll be disappointed, but with some patience and investment on your part the progress over time is remarkable.

by Carlo (not verified)

>While reporting old bugs doesn't make sense
Depends on the definition. A bug in 3.1 still to be found in 3.2 or 3.3, but never reported, is an old bug to me. You don't want to get it fixed?

by superstoned (not verified)

this is clearly my experience too.