Slashdot Mirror
Domain: nabble.com
Stories and comments across the archive that link to nabble.com.
Comments · 61
-
Re:GPLv3's Poison Pill and Open Source buyouts...It is a problem for Linux but attempts to avoid the problem have caused another issue in a major free software endeavour: duelling copyright holders.
The codebase? Java.
Back in the old days, Java was licensed with a scary EULA. This was fine for the end user because it was free as in beer. But it was trapped. OpenJDK is the cure. This is fine for end users and end developers. Now if a developer sees a bug (s)he'd like to fix, the license allows that.
Before Sun Microsystems released their code through OpenJDK, a rival GNU Classpath effort had been, and still is, producing a clean-room implementation. This prompted Sun to release their code under the same license in the hopes of cross-pollination.
The problem? Classpath and related projects have copyright assignment to FSF. OpenJDK requires copyright assignment to Sun. (Each, I think Sun does, may have copyright shared with the original author.)
So at the moment there's a something of a stand-off regarding this copyright. They're licensed under the same licensed but some code belongs to FSF and some belongs to Sun. So, effectively you either sell your soul to RMS or James Gosling!
Until lawyers of FSF and Sun can resolve this dispute, Classpath-derived code won't appear in OpenJDK and OpenJDK-derived code won't appear in Classpath. Third party repositories have been hosted for such hybrids. For the first scenario, IcedTea exists while for the second scenario BrandWeg has recently been announced. I quote the announcement for BrandWeg, which may be helpful in clarifying the issue:
This project is still very experimental, and is being conducted outside the repositories of either GNU Classpath or OpenJDK to retain the stability of these code bases and also avoid any unnecessary legal issues at this point (specifically the use of GPLv2 only code in OpenJDK which, if committed to the GNU Classpath codebase, would cause problems should Classpath want to move to GPLv3). -
Re:Still Stuck at 65500 rows in Calc?
>fixing the problem of only supporting 65,536 rows in Calc.
a) If you are using that many rows, you should be using a real database.
b) If you are willing to compile source code, there is an OOo code snippet showing how to have millions of rows in Calc. The use of that code snippet is entirely experimental. (A five second search on Google didn't find it. I did find the thread that discussed it at http://www.nabble.com/Recompile-Calc-to-support-more-than-64k-rows-and-256-columns--t725995.html
c) http://osdir.com/ml/gnome.apps.gnumeric/2002-10/msg00019.html is a link to how to change the number of rows in Gnumeric.
Amber -
Re:NetBeans??
Some people [http://pinderkent.blogsavy.com/archives/132, http://www.nabble.com/Very-slow-Help-t4482376.html, http://www.indicthreads.com/blogs/0105/eclipse_netbeans.html] find netbeans too slow. I use xcode so I dont know if thats true or not. I like xcode because its a native mac app and so it runs very fast.
-
Re:AtherosOf course, if you buy Atheros gear to solve your Linux wireless woes, you are rewarding a company for bad behavior. As many people have pointed out, Atheros seems to be deliberately unhelpful, releasing only binary drivers, and no documentation whatsoever. Atheros' bad engineering and refusal to work with the OSS community has lead to workarounds like the following commit in NetBSD: "The Atheros HAL on MIPS uses %s7 as a general purpose register, but the rest of the kernel uses it to store the value of curlwp. Sam won't recompile the HAL for us (fair enough), and we can't modify the HAL to use another register because doing so could put us in breach of the license (v. crappy). So, do a save/set/restore on %s7 in KernIntr() and in the stubs that the HAL uses to call back into the kernel.
"Please note that diffs are not public domain; they are subject to the copyright notices on the relevant files." More info here.
I suggest Ralink-based chipsets as an alternative. I have a couple in OpenBSD-based machines, and one in my girlfriend's Ubuntu laptop. I don't recall how difficult it was to get going in Linux, but it does work fine. The OpenBSD machines, of course, just worked out of the box. Ralink has been open and friendly with their documentation, and the chips are dirt cheap. I like this one in particular. This is the MiniPCI hard I use in my Soekris router. BTW, I am in no way affiliated with Ralink, except that I've purchased a few, and am happy with them. -
Re:I just noticed somethingI'm glad 2.2.8 will be in Mandriva 2008, and I don't mind that stuff is not backported. However, it is nice if you know how to help people and can tell them where to find third party builds. Ubuntu, Fedora, Suse have enough users that those are available.
See eg http://www.nabble.com/%22bug-report%22-tf4373448.
h tml, it clearly lists:
Gramps version: 2.3.6-1 LANG: fr_FR.UTF-8OS: LinuxDistribution: 2.6.17-14mdv".My search a while ago to find a more recent package turned up empty, so source installation it is for these people if the bug is solved in a more recent version. Our info on the wiki: http://www.gramps-project.org/wiki/index.php?titl
e =Installation#Upgrading_to_the_latest_version gives the latest info we got from a user, but I don't think it is correct, and although we have a lot of users on Mandriva we tell to upgrade, none apparently get back to us on where to find alternative packages, or can write up an upgrade process (removing old version, installing source, correct path). I feel save recommending Ubuntu/Fedora/Suse: download install, and yes, I know it is the community that provides those packages, but that is just it, isn't it?Also note the install path:
i ldroot/usr/share/gramps/GrampsDb/_GrampsBSDDB.pySorry you take this bad, but as an application developer, not a distro maker, I want my users to use the latest version instead of reporting old bugs to me. It's free after all, and having always the option of using the latest version is actually the one big benefit that linux offers. The requirement of constant upgrading the OS to get new versions or do source install otherwise, just doesn't work for the normal PC user.
-
But where is the Linux IO Scheduler?
Screw the CPU scheduler at this point. The kernel folks are missing the obvious and utter brokenness of the IO scheduling. These bugs have been outstanding about a year now!! And it's not just AMD64 anymore either. Quoth the kernel bug report:
"Now, as far as this bug being AMD64 only. We develop a portable data analysis
tool and we run it on Intel Core Mobile systems (Sony UX series, Panasonic
Toughbook series) and see this bug or one almost exactly like it on those
platforms as well.
"
http://bugzilla.kernel.org/show_bug.cgi?id=7372
http://bugzilla.kernel.org/show_bug.cgi?id=8636
http://www.nabble.com/IO-activity-brings-my-deskto p-to-its-knees-(2.6.22.1-ck1)-t4192136.html
http://forums.gentoo.org/viewtopic-t-482731-start- 500.html
At first, deadline IO was touted as an answer, but that doesn't completely fix things.
Some say Native Command Queueing is broken. One person claims deadline + NCQ disabled helps.
Some say the kernel's vfs_cache_pressure settings help, while others refute it (compare kernel bug report versus page 21 of the gentoo forum thread). But no one understands what's really broken in the kernel.
Can we please get Ingo working on IO scheduling? PLEASE? -
Re:Glad to see...
How much does a shuttle launch cost: $400-$500 million
How much does a shuttle weigh: 165,000 lbs
Which comes out to anywhere from $2,424.25 / lb to $3,030.30 / lb.. Assuming the light saber weighs one pound... and averaging the costs... sending the saber will cost $2,727.28. -
Not much opposition other than DiBona
"What I find interesting is not just that this has received a lot of criticism and opposition, but that one of the key opponents is Chris DiBona, open source programs manager for Google, Inc. "
From reading the OSI discussion list, DiBona is virtually the ONLY person raising criticism and opposition. This summary is widely innacurate. -
Re:Two good reasons to stay far away
Things are getting better with the latest release of Flash 9, which has a rewritten VM which includes JIT compiling. Unfortunately, code has to be ported to the new language (AS3) in order to take advantage of the speed. But they cleaned up a lot of annoying things in the language at the same time, and made everything much more standardized and consistent. It really feels like a first-class programming language now, that can compete with C# or Java.
Isn't AS3 ECMAScript 4?. The tamarin project now has preliminary codegen for AMD64, Arm and PPC but the VM is still far from being ready for cross platform deployment. This is why there's no 64bit flash and why the inkscape project just decided to use spidermonkey instead of tamarin.
-
Re:Better than innovationLinux NTFS write support is further along than you realise, unfortunately the developers are dragging their heals a bit. To quote Anton Altaparmakov of the Linux NTFS project:
New version. Written from scratch. Does full B+tree addition operations.
I created tens of thousands of files today and not one corruption. (-:
But before you get excited: You will have to wait till next summer to see
the code. Sorry. My hope is to give the world full read-write, open
source, kernel NTFS driver on both OSX and Linux by the time the next
major Mac OSX release is released (it should be in the next OSX major
release).
Source -
Re:nVidia
Exa is a replacement for XAA, the old X Acceleration Architecture. nvidia's binary drivers do not use XAA. They cooked up their own method for accelerating their drivers, independent of what the X developers were using. Their method is superior to XAA, and it remains to be seen whether or not it's superior to Exa. If so, don't expect them to change. If not, it'll likely be a while before it's implemented. This is proprietary software, remember. It takes a lot longer.
The OSS nv driver in 7.0 does *not* have Exa support. I've tried the currently-available experimental patch, and it crashes X within a few seconds of startup on my hardware. It also breaks XAA on non-AGP cards. Note that an exa-enabled nv may or may not accelerate 2D operations as well or as fast as the proprietary nvidia driver does. Personally, as long as it's better, I'll be happy, as nvidia is terribly unstable with composite enabled.
I love how incorrect information gets marked up as "informative". *sigh*